JP4800723B2 - Elevator group management system and control method thereof - Google Patents

Description

  The present invention relates to an elevator group management system and a control method thereof, and more particularly, to an elevator group management system and a control method therefor in which performance evaluation as a system and elevator assignment control for hall calls are improved.

  The elevator group management system provides a user with an efficient operation service by handling a plurality of elevator cars as one group. Specifically, multiple elevator cars (usually 3 to 8) are managed as a group, and when a new hall call occurs on a certain floor, one optimal car is selected from this group. Select and assign a hall call to the car.

  The current group management system is based on allocation control using an evaluation function based on the predicted waiting time. For example, when a new hall call occurs, the car that has the smallest waiting time for the hall call that each car is responsible for, the car that has the smallest maximum waiting time, or the car that has the smallest average waiting time, Assign the hall call. The allocation control based on the predicted waiting time is a basic method adopted in the group management control of each elevator maker, but has the following two problems.

  1) The optimal car allocation for a hall call that has already been generated, and the influence of the hall call that will be generated in the future is not necessarily fully considered.

  2) Since the estimated waiting time is used as an index and assigned to the car, the positional relationship of each car is not considered.

  In order to solve the problem of the allocation method based on the predicted waiting time, various control methods have been proposed so far. The basic idea can be summarized in the control to move each elevator car at regular intervals. If the elevator cars are not evenly arranged, that is, if the time interval between two cars is long, and if a new hall call occurs on the floor between the cars, the call will have a longer waiting time. Probability is high. It has long been known that long waiting can be suppressed if each car can be arranged at regular intervals in time, and there are the following techniques as a conventional control method for the purpose of arranging equal intervals in time. .

  Patent Document 1 “Allocation Evaluation Control Incorporating Time Equal Interval State as an Index” predicts the placement of each car at the previous time point and predicts the time interval of each car at that time point. . An assignment limit evaluation value is calculated from the predicted car interval, and assignment is controlled so that the car is not assigned to some floor areas. As a result, the aim is to make the intervals of the cars approach the same time interval.

  Further, in Non-Patent Document 1 “Arrival Prediction Time Calculation Method Using Dynamic Programming”, a dynamic call method is used to determine the floor of a car call caused by an unserviced landing call in order to accurately obtain the predicted waiting time. An estimation method is disclosed.

Japanese Patent Publication No. 7-72059 (Overall) 2003 IEEJ Electronics, Information and Systems Division Conference, GS18-3, “Elevator Group Management Arrival Prediction Time Calculation Method Using Dynamic Programming” p1099-1102 (Overall)

  In the above-mentioned Patent Document 1, it is important to predict the future arrangement of each car and determine the assignment from the time interval of each car at that time. However, this method merely evaluates the spatial arrangement at a certain point in time in the future (there may be a plurality of cases), and lacks information for analyzing the contents of the evaluation. Specifically, when analyzing the cause of the allocation evaluation, information on the spatial arrangement of the car at a certain point of time must be individually analyzed, which makes it difficult to analyze from a macro viewpoint. In addition, since the position is predicted for each floor, there is a problem that the prediction accuracy is not good. In particular, in the case of a building where the height between floors is long, the prediction accuracy affects the control performance.

  Non-Patent Document 1 aims to improve the estimation method of the estimated arrival time and optimize the prediction waiting time described above, and is not related to the prediction of the arrangement of each car.

  In general, performance evaluation as a system in an elevator group management system includes a short average waiting time and a low occurrence probability of long waiting calls. However, it is often difficult to make a detailed evaluation, such as what causes the difference in control performance under various traffic demand changes in the building.

  One of the objects of the present invention is to provide a system that supports evaluation of control performance as a system in an elevator group management system.

  Another object of the present invention is to analyze the cause of allocation evaluation from a more macro viewpoint in allocation evaluation for controlling each car to be arranged at equal intervals in time in order to solve the above-mentioned problems of the prior art. It is to provide a system that can.

  It is still another object of the present invention to provide a system that can analyze more detailed information about the arrangement of each car and can utilize the information at the time of assignment.

  In a preferred embodiment of the present invention, a prediction trajectory creating means for creating a predicted trajectory indicating movement of a predicted position of each elevator on a time axis in a predetermined period from the present time to the near future is provided. Yes.

  In a preferred embodiment of the present invention, a predicted trajectory creating means for creating a predicted trajectory indicating movement of a predicted position of each elevator on a time axis in a predetermined period from the present time to the near future, and for each elevator The present invention is characterized in that a predicted trajectory display means for displaying the predicted trajectory is provided.

  Furthermore, in a preferred embodiment of the present invention, a predicted trajectory creating means for creating a predicted trajectory indicating a predicted position of each elevator with respect to a time axis for a predetermined period in the future, and an evaluation value relating to a positional relationship between the predicted trajectories of each elevator And an assigning means for assigning an elevator to the generated hall call based on the evaluation value.

  According to the elevator group management system according to the preferred embodiment of the present invention, it is possible to appropriately evaluate the elevator group management control from the predicted trajectory of each elevator on the time axis. In addition, it is possible to obtain technical support for evaluation and improvement, such as giving suggestions for necessary improvements. For example, if the prediction error is large, it can be determined from the inclination of the prediction trajectory whether it is due to an error in estimating the traffic demand.

  Other objects and features of the present invention will become apparent from the following description of embodiments.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  First, the concept of control of the elevator group management system in the present invention will be described with reference to FIG. 1A, 1B, and 1C are control image diagrams of an elevator group management system according to the present invention. First, FIG. 1A will be described. FIG. 1A shows a situation in which a car responding to the hall call is assigned immediately after a new hall call is generated. FIG. 1A shows an elevator operation diagram, where the horizontal axis represents time and the vertical axis represents the floor position on the building. The time axis represents the future time starting from the current point. That is, this figure represents a predicted operation diagram of the elevator in the future. The elevator is composed of two cars, Unit 1 and Unit 2. As shown in FIG. 1 (a), Unit 1 is currently near the third floor and is moving upward. Unit 2 is near the 5th floor and is moving upward. The predicted trajectory of each car is represented by a line on the figure. It can be seen that the two predicted trajectories are close to each other and are in a dangling state. In this situation, consider a case where a new hall call is generated upward on the 8th floor.

  FIG. 1B shows the predicted trajectory of each car when a newly generated hall call is temporarily assigned to the first car. Looking at the predicted trajectory of Unit 1, it stops 8 floors above to serve a new hall call. As a result, the predicted trajectories of the subsequent No. 1 and No. 2 units are wider than the previous state of FIG. When the interval of the predicted trajectory is evaluated at the positional relationship evaluation time tref after a predetermined time has elapsed from the present time, it can be clearly seen that the interval is increased in FIG. 1B compared to FIG. A typical example of the positional relationship between the cars is an interval. Hereinafter, the positional relationship evaluation time tref is simply referred to as an interval evaluation time tref.

  FIG. 1C shows a predicted trajectory of each car when a newly generated hall call is temporarily assigned to the second car. Looking at the predicted trajectory of Unit 2, it stops 8 floors upward to serve a new hall call. As a result, the prediction trajectories of the subsequent No. 1 and No. 2 units are shorter than the previous state shown in FIG.

  Comparing the intervals between the cages at the interval evaluation time tref with respect to FIG. 1 (b) when provisionally assigned to Unit 1 and FIG. 1 (c) when provisionally assigned to Unit 2, the one assigned to Unit 1 is better. It can be seen that the interval is approaching. Therefore, it can be evaluated that it is better to assign to Unit 1 in order to approach the equally spaced state. Such a series of evaluation methods is an outline of group management control according to an embodiment of the present invention. As a result of such control, an appropriate interval can always be maintained, and unnecessary long waits can be reduced. For this purpose, it is necessary to obtain the predicted trajectory of each elevator for a predetermined time from that point. FIG. 1 illustrates the length of the horizontal axis according to the necessity, and the length of the predetermined time is set to a time longer than the average one-round time of the elevator at that time.

  FIG. 2 is a control block diagram of the entire elevator group management system according to one embodiment of the present invention. The operation of the N elevator cars 32A, 32B, 32C,... Is controlled by the elevator car control devices 31A, 31B, 31C,..., And the group management control unit 1 supervises each of these car control devices. Control.

  The group management control unit 1 performs the following processing. First, information on the hall call buttons (41A, 41B) on each floor and information on each of the N elevator apparatus 31A, 31B, 31C,... Are stored in the input information storage unit 1. Here, when a new hall call is generated, the waiting time evaluation value calculation unit 3 uses the information in the input information storage unit 1 to calculate a predicted waiting time for each hall call including the already generated hall call. The waiting time evaluation value W is calculated based on this. Further, as described with reference to FIG. 1, the interval evaluation value calculation unit 4 predicts the future positional relationship of each elevator car, and calculates the interval evaluation value E based on this. The weighting factor setting unit 8 sets a weighting factor WT corresponding to the situation at that time. The feature of this embodiment is the method of setting the weighting coefficient, and details thereof will be described later. The comprehensive evaluation value calculation unit 6 calculates a comprehensive evaluation value Φ obtained by weighting and adding the waiting time evaluation value and the interval evaluation value from the waiting time evaluation value, the interval evaluation value, and the weighting coefficient. The overall evaluation value Φ is expressed by the following formula, for example.

Φ = W + WT · E ………… (2)
This total evaluation value is calculated for the case where each car is provisionally assigned to a new hall call, and the assigned elevator determination unit 7 assigns the elevator car having the best evaluation in terms of waiting time and equal distance between the cars. Determined as a basket.

  Here, the main points of the weighting coefficient setting method, which is a feature of the present embodiment, will be described. The setting of the weighting factor according to the present embodiment is mainly constituted by two methods. The first method is a method in which the current traffic flow is determined, a group management control simulation is repeatedly executed based on the traffic flow, and the most appropriate weight coefficient value is obtained by searching. The second method is a method of predicting the number of hall calls and obtaining the setting range of the weighting coefficient and the initial value of the setting based on this. In particular, the latter (second method) is the point of this embodiment.

  Each will be specifically described below. First, in the first method, the traffic flow determination unit 20 determines the current traffic flow from the information in the input information storage unit 2, and the weight coefficient optimum solution search unit determines the most suitable weight coefficient for the traffic flow. Search at 21. Here, the search for the optimum solution of the weight coefficient is performed by repeatedly executing the simulation of the group management control under the traffic flow conditions at that time. The simulation of the group management control is executed in the simulation unit 22. The search for the optimum solution for the weight coefficient may be executed online or offline (for example, at night). When executed offline, the main traffic flow of the building (hereinafter referred to as traffic flow mode) is extracted in advance, and the group management control simulation for this traffic flow mode is executed offline. .

  Next, in the second method, the hall call occurrence number calculation unit 10 calculates the hall call occurrence number or an amount related to the hall call occurrence number based on the input information of the input information storage unit 2. The initial value and range (upper limit value and lower limit value) of the weighting factor are calculated by the weighting factor initial value calculating unit 12 and the weighting factor range calculating unit 11 based on the number of hall calls generated.

  For the calculated initial value of the weighting factor, there are two types of flows: a flow sent to the weighting factor setting unit 8 and a flow sent to the weighting factor optimal solution search unit 21. The initial value sent to the weight coefficient optimum solution search unit 21 is used as an initial value for search (initial value when searching for the first traffic flow). The initial value sent to the weighting factor setting unit 8 is the first time when the search for the optimum solution of the weighting factor has not converged with respect to the traffic flow occurring at that time, or when the traffic flow occurring at that time is the first time. This is set as a weighting factor that is actually used in the case of traffic flow.

  FIG. 3 is a control processing flowchart of the elevator group management system according to the embodiment of the present invention shown in FIG. Hereinafter, the flow will be described with reference to FIG.

  First, it is determined whether or not a weight coefficient optimum solution search is executed (ST001). This is a process in the case of performing a weighting factor search in an offline simulation. For example, it is determined whether or not the weighting factor optimum solution search is to be executed based on the load state of a calculation processing device such as a microcomputer or a personal computer and time information such as daytime or nighttime. If it is determined to be executed, the optimum solution of the weighting factor is searched for the previously extracted traffic flow mode (ST002). The search method at this time includes a search method for all possible values of the weighting factor (actually all values within a certain range), a branch and bound method, a hill descending method, a search using a neural network, A method using a search by a genetic algorithm is used. After the execution of the weight coefficient optimum solution search (this includes the case where the optimum solution is not found and is interrupted in the middle) or when the search is not executed, input information is input from the input information storage unit (2 in FIG. 1). (ST003). After obtaining the input information, it is checked whether or not a car assignment process has occurred (ST004). If no assignment process has occurred, the process returns to ST001. If an assignment process has occurred, the process proceeds to the next process. .

  When the car assignment process has occurred, the temporary assignment that temporarily assigns each elevator car to the assigned hall call (usually newly generated hall call) and calculates the evaluation value in that case Car cage processing (ST005) is executed. This executes a process of changing the Ka from 1 to N (the number of group management elevators is N) in order from the temporarily assigned car No. Ka.

  First, a predicted trajectory is calculated for the temporarily assigned car No. Ka (ST006). The predicted trajectory of the temporarily allocated car corresponds to the predicted trajectory of the first car in FIG. Next, a predicted trajectory is calculated for a car K other than the temporarily assigned car (each car with K ≠ Ka) (ST007). This trajectory corresponds to the predicted trajectory of Unit 2 in FIG. Then, the interval evaluation time tref is calculated (ST008) (an example of tref is shown in FIG. 1B), and the predicted interval value Bm (m = 1, 2,..., N) of each car at the point of tref. ) Is calculated (ST009). Although the calculation method of the prediction interval value will be described later, the prediction interval can be calculated from the temporal distance or the spatial distance between the positions based on the predicted position of each car by tref. If the predicted interval value of each car is calculated, the interval evaluation value is calculated based on this (ST010). This is an interval evaluation value when the temporarily assigned car is Ka No. machine, and is expressed as E (Ka). E (Ka = 1), E (Ka = 2)... Are calculated by repeating the temporary allocation car loop processing.

  Once the interval evaluation value is calculated, the waiting time evaluation value W (Ka) when the temporary assigned car is Ka No. is then calculated (ST011). The waiting time evaluation value is determined by, for example, a method of setting the waiting time for the call when the assigned hall call is assigned to the generated hall call, or the maximum waiting time among hall calls held by the Ka No. There is a way to set. Furthermore, a method for setting the average waiting time for hall calls held by all units including the Ka unit, and a method for setting a square sum of the waiting times for hall calls held by all units including the Ka unit. Etc.

  After the interval evaluation value E (Ka) and the waiting time evaluation value W (Ka) are calculated, the weighting coefficient WT is calculated (ST012). The method for calculating the weighting factor is outlined in FIG. 2, and the details will be described later.

  Now, based on the interval evaluation value, the waiting time evaluation value, and the weighting factor, a comprehensive evaluation value serving as an index for determining allocation is calculated (ST013). The overall evaluation value is expressed as follows.

Φ (Ka) = FT (W (Ka), E (Ka), WT) (3)
Here, FT represents a function. More specifically, it is expressed by a linear sum expression such as the following expression.

Φ (Ka) = W (Ka) + WT · E (Ka) ………… (4)
The series of processes from ST005 to ST013 described above are executed until the temporary allocation car loop is completed (until the temporary allocation process is performed on all cars) (ST014). If not, the temporary allocation car is updated to the next car (ST015), and the process from ST006 is executed on the updated temporary allocation car Ka. When the temporary assigned car loop is completed, the overall evaluation values Φ (Ka = 1), Φ (Ka = 2),..., Φ (Ka = N) are compared, and the assigned elevator is assigned to the car having the best evaluation value. Determine (ST016). After the assignment is determined, the process returns to the first process ST001, and the above process is repeated to execute the process.

  Hereinafter, the weighting coefficient setting method according to the present invention will be described with reference to FIG.

  FIG. 4 is a graph illustrating a weighting factor setting method according to an embodiment of the present invention. The horizontal axis represents the number of hall calls generated per elevator lap, and the vertical axis represents a weighting coefficient. Here, the number of hall calls per elevator lap is the number of halls generated during one lap (for example, from the lowermost floor in the upward direction to the lowermost floor in the downward direction) for each group-managed elevator. Represents the average number of calls. The number of hall calls per lap is large when it is busy, and it is small when it is quiet.

In the graph of FIG. 4, lines representing the input / output characteristics of the three functions are drawn. The line F01 is a function for determining an appropriate initial value WT ₀ of the weighting factor, and the line F02 is an appropriate upper limit for the weighting factor. A function that determines the _{upper limit} of the value WT, and a line F03 is a function that determines an appropriate lower limit WT _lower limit of the weighting coefficient. A function (line F01) for determining an appropriate initial value of the weighting coefficient is used by the weighting coefficient initial value calculation unit 12 of FIG. 2, and a function for determining an appropriate upper limit value of the weighting coefficient (line F02) and an appropriate lower limit value are set. The function to be determined (line F03) is used in the weighting coefficient range calculation unit 11 in FIG.

  The major features of the three functions that determine these weighting factors are as follows.

1) The appropriate initial value, upper limit value, and lower limit value of the weighting factor can be obtained immediately according to the number of hall calls.
2) The continuous change in the number of hall calls per lap, which is an input variable, causes the weighting coefficient, which is an output, to be continuously determined.
3) The input variable is a scalar value (one variable), and 4) The input variable can take a continuous value as a real number.

  That is, the weight value is calculated based on a function whose output value changes continuously with respect to a change in input. In other words, the weight value is calculated based on a function whose input is a real number and whose value changes continuously. The effect of this feature will be described later with reference to FIG. 8 in comparison with other setting methods (FIGS. 6 and 7).

As shown in the graph of FIG. 4, for example, when the number of hall calls per lap is NA, the initial value of the weighting coefficient is WT ₀ = F ₀ (NA), and the upper limit value and the lower limit value are respectively WT _{upper limit} = F _{upper limit} (NA), WT _{lower limit} = F _{lower limit} (NA). Regardless of how the NA changes due to a change in traffic demand, WT ₀ , the WT _{upper limit} , and the WT _{lower limit} can be obtained immediately. This is a major feature. Each function has a value on the vertical axis that is always zero for a value smaller than the horizontal axis from a value on the horizontal axis that intersects the zero value on the vertical axis.

  In the graph of FIG. 4, the horizontal axis shows an example of the number of hall calls generated per lap. However, the present invention is not limited to this, and a value based on the number of hall calls generated (for example, the number of hall calls generated in a predetermined time) ) Can achieve the same effect. Furthermore, the same effect can be obtained not only by the number of hall calls generated but also by a scalar value index related to traffic demand. For example, the same effect can be obtained with a value based on the number of users, a value based on the total number of hall calls and car calls, an average waiting time, and the like.

  The reason why an appropriate weighting factor can be obtained by a function as shown in FIG. 4 by using the number of hall calls generated per elevator lap as an input will be described below. The interval evaluation value (already described with reference to FIG. 2) is an index for evaluating the time interval between the cars, and the time interval between the cars corresponds to the maximum waiting time for a hall call that will occur in the future. Therefore, the importance of the interval evaluation value has a strong relationship with the number of hall calls to be generated in the future. For example, as the number of hall calls to be generated in the future increases, it is better to make the intervals as equal as possible in time, and it is better to make the interval evaluation value act stronger. Here, the number of hall calls to be generated in the future is considered to have a strong correlation with the number of hall calls generated per round at that time or at a later time. Therefore, there is a relationship between the number of hall calls generated per lap and an appropriate weighting factor value. By expressing this as a function as shown in FIG. 4, the number of hall calls generated per elevator lap is appropriate. Weighting factors (in practice, appropriate initial values and ranges) can be determined.

  FIG. 5 is an image diagram of an optimum solution search method for weighting factors according to an embodiment of the present invention. In the figure, the horizontal axis represents the value of the weighting coefficient, and the vertical axis represents the average waiting time as a result of executing the group management control simulation. A characteristic of the average waiting time when the group management control of the present invention shown in FIG. 2 is simulated for each value of the weighting coefficient WT on the horizontal axis is represented by a curve 901. In the case of searching for the optimum solution of the weighting factor by repeating the simulation, or in the case of searching for the optimum solution of the weighting factor by the simulation shown in FIG. 2 (portions 20, 21, and 22 in FIG. 2), any characteristic curve 901 is selected. The location is the initial value (starting point). For this reason, it takes a certain amount of time to reach the optimal solution. In particular, for traffic flows that appear for the first time, the performance (average waiting time) that appears initially may vary greatly depending on where the initial value (starting point) is determined. Also, during the search transition period, the number of simulations is small (not enough), so there is a possibility that the search will be performed in a bad direction. Therefore, in this embodiment, by using the continuous function shown in FIG. 4, an appropriate initial value 902 of the weighting factor can be immediately determined by the number of hall calls generated at that time, and similarly an appropriate upper limit value 903 is obtained. And the lower limit value 904 can be determined immediately. These functions are realized by functional elements indicated by reference numerals 10, 11, 12, and 8 in FIG. Furthermore, after determining an appropriate initial value and range, an optimal solution is searched for within the range based on the initial value, so that the optimal solution can be obtained quickly and stably.

Note that the initial value WT ₀ , the upper limit value WT _{upper limit} , and the lower limit value WT _{lower limit} are determined according to the number of hall calls, so these values are also variably adjusted according to the traffic demand state at that time. It is like that.

  The reason why the method of setting the weighting coefficient by the function of the number of hall calls as shown in FIG. 4 is superior will be described below in comparison with the method of inputting the traffic flow and setting it by repeating the simulation. First, as the previous stage, an outline of the traffic flow will be described with reference to FIG. 6, and an example of a weighting coefficient setting method in the case where the input is handled with the traffic flow will be described with reference to FIG.

  FIG. 6 is an example of a weighting method for traffic flow in a 6-floor building. The table on the left side of FIG. 6 represents an OD (Origin-Destination) matrix representing traffic flow. In this OD matrix, the column direction (horizontal direction) is represented by the ride floor (Origin = meaning of the departure place) and the row direction (vertical direction) is represented by the descending floor (Destination = meaning of the destination). It represents the number of people (the number of people who moved in a predetermined time) corresponding to the row / column elements. For example, the number of people on the 2nd floor and the 5th floor is 3 people from the figure. Moreover, since it is a 6th floor building, it becomes a 6x6 matrix. The traffic flow is a total representing the number of people moving on each floor, and can be represented by such an OD matrix (the OD matrix is used in traffic analysis for roads).

  The movement number of each element is represented by a variable tr1, tr2,..., Which is the second table (OD matrix) from the left in FIG. 6, which is a 30-dimensional vector (tr1, tr2, tr3,... tr30). If such a traffic flow is handled as it is and the weighting coefficient is obtained, a 30-dimensional function such as the following equation must be obtained.

WT = F (tr1, tr2, tr3,..., Tr30) (5)
This is a very complex function and it is practically impossible to analyze. Therefore, instead of handling the 30-dimensional vector space as it is, consider dividing it into several main subspaces. In this way, the object to be handled becomes a finite number of subspaces, which makes it easy to handle. This subspace corresponds to the traffic flow mode. Hereinafter, this will be described with reference to FIG.

  FIG. 7 is an explanatory diagram of an example of a weighting factor setting method when the input is handled by a traffic flow. In order to simplify the explanation, a building on the second floor is taken as an example here. In the case of a two-story building, the OD matrix is simply a 2 × 2 matrix as shown in FIG. 7A, and the traffic flow is a two-dimensional vector of (tr1, tr2) as shown in FIG. 7B. expressed. Here, tr1 and tr2 are represented as a two-dimensional graph with the horizontal axis and the vertical axis, respectively, as shown in FIG. The points on the graph in FIG. 7C represent the traffic flow. For example, as in work, tr1 is large (moving up from the first floor to the second floor is large) and tr2 is small (from the second floor). A traffic flow with a small movement on the first floor is represented as a dot in the figure. As shown in FIG. 7C, even a traffic flow of a simple two-story building becomes a two-dimensional plane of (tr1, tr2), and if this is to be handled by a function of a two-dimensional variable such as equation (6), You have to deal with complex functions.

WT = F (tr1, tr2) (6)
Therefore, representative traffic flows are grouped as one lump on the two-dimensional planes tr1 and tr2 in FIG. FIG. 7D shows an example, in which the two-dimensional plane tr1 and tr2 is divided into four regions. Then, representative traffic flow vectors V1, V2, V3, and V4 are defined for each of the four regions. V1, V2, V3, and V4 are traffic flows representing the whole, and correspond to the above-described traffic flow mode. Note that the area is not divided in one way, but various ways can be considered according to the traffic demand characteristics of the building.

  By expressing in the traffic flow mode, as shown in FIG. 7E, a weighting factor suitable for the traffic flow mode may be determined. This optimal solution can be obtained by repeatedly executing the simulation of the group management control for the traffic flow mode while changing the weighting factor. For example, in the case of FIG. 7E, for the traffic flow mode V3, the optimum weighting factor that can lower the average waiting time by repeatedly executing the simulation while changing the weighting factor is WT = 5.6. A value can be defined. Similarly, when an optimum weighting coefficient is determined by simulation for each traffic flow mode, a correspondence table between the traffic flow mode and the weighting coefficient as shown in FIG. 7F is obtained.

  To summarize the above, when input is handled by traffic flow, it is complicated to handle multi-dimensional vectors as they are, so it is represented by main traffic flow modes, and simulation is repeated for each mode. A method of setting weights will be taken.

  FIG. 8 is a comparison diagram of one embodiment of the present invention and a conventional weighting factor setting method, and is compared and organized by item. In this table, for example, the setting method disclosed in Japanese Patent Laid-Open No. 1-2226676 is compared with the setting method of this embodiment for five items. Hereinafter, the first item is compared in order. First, with respect to input variables, the conventional technique uses a traffic flow mode in which a traffic flow vector or its main component is extracted. On the other hand, in this embodiment, the number of hall calls generated is used. The nature of each input variable is a multi-dimensional vector, for example, a vector such as tr1, tr2, tr3,... In the prior art, but in the present embodiment, it is one variable, in other words, a scalar value. Yes. In addition, the method of determining the weighting coefficient, which is the output value, is set by a continuous function as shown in FIG. 4 in the present embodiment, whereas the prior art repeatedly selects a group management control by performing a search. is doing. Therefore, each feature has a feature that a certain amount of time is required until the value is selected in the prior art, whereas the method of this embodiment can determine the value instantaneously. From this feature, the method of the present embodiment can immediately set an appropriate weighting factor (more precisely, an initial value of an appropriate weighting factor) for various changes in traffic flow compared to the prior art, There is an effect that the control performance can be exhibited stably.

  FIG. 9 is a processing flow diagram summarizing the weighting factor setting method according to one embodiment of the present invention described so far. This series of processing includes hall call occurrence number calculation unit 10, weight coefficient range calculation unit 11, weight coefficient initial value calculation unit 12, traffic flow determination unit 20, weight coefficient optimal solution search unit 21, simulation unit 22, It is executed in the weight coefficient setting unit 8. Hereinafter, it demonstrates in order. First, input information is input (ST101), and the number NA of hall calls generated per lap is calculated based on this input information (ST102). This value can be calculated as follows, for example. If the total number of hall calls generated in a predetermined time set to be longer than the average one lap time of the elevator at that time is NH and the total number of direction inversions is NR, NA is obtained by the following equation.

NA = NH / {(NR / 2) +1} (7)
Incidentally, the denominator {(NR / 2) +1} in the equation (7) corresponds to the total number of laps in a predetermined time of all elevators.

When the number of hall calls generated per round NA is calculated, based on this, the initial value WT ₀ of the weighting coefficient is calculated by the function WT ₀ = F ₀ (NA) (ST103). Then, the traffic flow mode occurring at that time is determined (ST104), and it is determined whether or not the optimum solution search for the weighting factor has already been executed for the traffic flow mode (ST105). If the optimum solution search has been performed and a weight coefficient value that produces a better result than the initial value is searched, the weight coefficient is set to that value (the optimum solution at that time) (ST106). If the optimal solution has not been executed, or the weighting factor value that produces a better result than the initial value has not yet been searched, the weighting factor value is set to the initial value WT ₀ (ST107).

Since the weighting factor is set by the processing flow as shown in FIG. 9, even if it is the first traffic flow in the building, the appropriate weighting factor value is immediately set by the initial value WT ₀ determined by the function. Can do. Even in the initial operation of the building, an appropriate weight coefficient value can be immediately set by the initial value WT ₀ determined by the function. Even when the optimum solution search has not converged and an appropriate solution has not been found, an appropriate weighting factor value can be set by the initial value WT ₀ determined by the function.

FIG. 10 is a processing flow diagram summarizing an optimum solution search method for weighting factors according to an embodiment of the present invention. This series of processing is performed in the hall call occurrence number calculation unit 10, the weight coefficient range calculation unit 11, the weight coefficient initial value calculation unit 12, the traffic flow determination unit 20, the weight coefficient optimum solution search unit 21, and the simulation unit 22 in FIG. Made. Hereinafter, it demonstrates in order. First, the traffic flow mode data is input (ST201), and it is determined whether or not the optimum solution search process is executed for the traffic flow mode (ST202). If it has been executed, the initial value of the search is set to the optimum value at the previous search (ST208). If it is not running, it calculates a hall call occurrence of NA per one turn in the traffic flow mode (ST 203), based on the value, to calculate the initial value WT ₀ weight coefficient (ST 204), the upper limit value The WT _{upper limit} is calculated (ST205), and the lower limit WT _{lower limit} is calculated (ST206). Then, the initial value of the search is set to WT ₀ .

A simulation of group management control for the determined traffic flow mode is executed for the set initial value, and the optimum solution search is executed by changing the initial value and repeating the simulation (ST209). After execution of this optimum solution search (or during the search), it is determined whether the obtained optimum weighting factor is within the range between the upper limit value WT _{upper limit} and the lower limit value WT _{lower limit} (ST210). If it is within the range, the value of the weighting coefficient is the optimal solution. If not within the range, the value of the weighting factor is set to the upper limit value WT _{upper limit} or lower limit value WT _{lower limit} or the optimal solution at the previous search (ST211).

  Since the search for the optimum solution of the weighting coefficient is executed by the processing flow as shown in FIG. 10, the initial value of the search can be set to an appropriate value by the function of the number of hall calls generated even for the first traffic flow. As a result, the optimum solution search can be performed more efficiently. In addition, since the search range is set to an appropriate range by the function of the number of hall calls, the optimal solution search can be performed more efficiently without searching for an inappropriate area even in the initial stage or transient state of the search. . These advantages have been described with reference to a search image in FIG.

  FIG. 11 shows details of the input information storage unit 2 shown in FIG. The input information storage unit 2 includes the following data storage. First, building facility data 201, group management elevator facility data 202, and group management elevator state data 203 at the present time. Next, state data statistics 204 of the group management elevator, state data 205 of each hall of the building at the present time, traffic flow data 206 of the building, and time information data 207 are shown. The building facility data 201 stores data such as the number of floors of the building, the height of each floor, and the floors to be serviced by the group management. The group management elevator equipment data 202 stores data such as the number of group management elevators, the rated speed of each elevator, the car capacity, the door opening / closing speed, and the standard door opening time. The current state data 203 of the group management elevator includes car position, direction information, speed information, car load information, assigned hall call information, car call information, stop floor information, hall call of each hall call. There are data such as duration information and one round time of each car. The group management elevator status data statistics 204 stores data such as the number of hall calls generated, the number of car calls generated, the number of users, the average hall call duration, the number of direction inversions, the average load, and the average one lap time. Has been. In the hall state data 205 at the present time, data such as information on hall call buttons 41A and 41B and information on cameras 51A and 51B of waiting customers in the hall are accumulated. In the building traffic flow data 206, building traffic flow data as shown by the OD matrix in FIG. 6 is accumulated. The time information data 207 stores clock information, calendar information such as year, month, day, day of the week, holidays, and days with special events. The input information storage unit 2 stores all the data described above. The input information storage unit 2 is not necessarily integrated with these data, and the above data may be stored in a distributed manner. In this case, what is virtually integrated with these data may be considered as the input information storage unit 2.

  So far, the details of the method for setting the weighting factor have been described. Next, details of a method of calculating the car interval evaluation value will be described with reference to FIG. The point in the method of calculating the car interval evaluation value is the time setting method for evaluating the interval.

  FIG. 12 shows a detailed configuration of the car interval evaluation value calculation unit 4 shown in FIG. First, a predicted trajectory calculating unit 401 that calculates a predicted trajectory of each future car is provided, and a predicted car interval calculating unit 402 that calculates a car interval after a predetermined time (interval evaluation time tref described later) based on the predicted trajectory. It has. Next, a car interval evaluation value calculation unit 403 that calculates a car interval evaluation value based on the prediction interval, and a farthest call that obtains the farthest call in time (call including the hall call and the car call) by searching. A search unit 404 is provided. Further, it is constituted by an interval evaluation time setting unit 405 for obtaining an interval evaluation time tref based on the estimated arrival time for the farthest call, that is, the maximum estimated arrival time.

  Hereinafter, the function of each element of the interval evaluation value calculation unit 4 of FIG. 12 will be described with reference to FIG. 1 described above. In FIG. 1 (b), the predicted trajectory calculation unit 401 in FIG. 12 calculates the predicted trajectories (the trajectory of the black line from the present time to the future direction in the figure) of Unit 1 and Unit 2 in FIG. The interval evaluation time setting unit 405 sets the interval evaluation time tref of b). The predicted position at the interval evaluation time from the predicted trajectory of each car, that is, the car positions of the first car and the second car drawn on the interval evaluation time in FIG. 1B are obtained, and the time of each car is determined from this predicted position. The predicted cage interval calculation unit 402 in FIG. 12 obtains the target interval or the spatial interval. The car interval evaluation value calculation unit 403 calculates an evaluation value for evaluating the equidistant property from the value of the car interval. For example, when FIG. 1B and FIG. 1C are compared, FIG. 1B is in a state where the equi-interval property is higher, and the car interval evaluation value calculation unit 403 digitizes this.

  In calculating the car interval evaluation value described above, one of the important elements is a method of setting a time for estimating the car interval. This setting method is also a feature of the present embodiment, and the method for setting the interval evaluation time will be described below with reference to FIGS.

  FIG. 13 is a processing flowchart for setting the interval evaluation time tref. This series of processing is executed by the farthest call search unit 404 and the interval evaluation time setting unit 405 of FIG. First, a car loop process for sequentially searching each car is executed (ST501). Here, the car to be searched is No. K (K = 1, 2,..., N). The initial value of K is 1. First, all hall calls assigned to the No. K machine are searched, and the maximum predicted arrival time ART_H (K) is selected (ST502). Next, all car calls handled by the K-th car are searched, and the maximum estimated arrival time ART_C (K) is selected (ST503). Thereafter, ART_H (K) and ART_C (K) are compared, and the larger one is obtained as the arrival prediction time ART_MAX (K) of the farthest call of the K-th unit (ST504). This is expressed by the following equation.

ART_MAX (K) = MAX {ART_H (K), ART_C (K)} (8)
It is determined whether or not the processing up to this point has been performed for all elevators based on whether or not K is equal to N (ST505). If not equal, the value of K is incremented by one (ST506). The process returns to the first process (ST502). If the estimated arrival time of the farthest call is calculated for all elevators, the maximum value of the estimated arrival time of the farthest call for all elevators is set as the interval evaluation time tref (ST507). This is expressed by the following equation.

tref = MAX {ART_MAX (K)} (9)
In this way, for the hall call and car call that all elevators have at that time, the call with the maximum estimated arrival time (the farthest call for all units) is selected and the estimated arrival time for that call. Is the interval evaluation time. The effect of determining the interval evaluation time by such a method will be described with reference to FIG.

  Note that the estimated arrival time of the farthest call for all the units changes due to a change in time (for example, a change of about 20 seconds) and a value changes, so the value changes every time allocation processing is performed. As a result, the value of the interval evaluation time changes with every allocation process. This indicates that the status of the call that is occurring (the status of the number of calls that are being generated, etc.) changes from time to time, and accordingly, the interval evaluation time is adjusted accordingly.

  FIG. 14 illustrates a concept for setting the estimated interval time. In the graph of FIG. 14A, the horizontal axis represents the time axis with the origin as the current time, and the vertical axis represents the floor position. In the graph of FIG. 14A, the two trajectories represented by solid lines represent the predicted trajectories of Unit 1 and Unit 2, respectively. F10 represents the predicted trajectory of Unit 1 and F11 represents the predicted trajectory of Unit 2. The point at which the interval is evaluated for the two predicted trajectories is a point for setting the interval evaluation time. This interval evaluation time has the following properties.

  First, when the interval evaluation time is set in a region close in time to the current time, there is a problem that the influence of a call (hall call or car call) that is already handled at a later time cannot be considered. This influence is particularly significant when overtaking occurs in the middle. For example, assuming that the interval evaluation time is set at the time indicated by reference numeral F12 in the graph of FIG. On the other hand, it is determined that the subsequent unit No. 2 is approaching and is in a dumped state. For this reason, if it evaluates by the cage | basket | car space | interval of this time (time of F12), the control which advances 1st machine (suppresses allocation) and delays 2nd machine (promotes allocation) will be good. However, looking at the predicted trajectory, Unit 2 overtakes Unit 1 after that, and if Unit 2 is delayed, it will work in the direction of prolonging the dango state. As described above, when the interval evaluation time is set in a region close to the current time, the influence of not considering the calls already handled by each elevator is increased.

  Next, considering the case where the interval evaluation time is set in a region far in time from the current time, this is a future event, so there is a high possibility that a new hall call or car call will occur after that, and the predicted trajectory is It can change a lot. For example, in FIG. 14A, if it is assumed that the interval evaluation time is set at the time indicated by reference numeral F13, there is a high possibility that a new hall call or car call will be generated by this time. The car spacing at is highly uncertain and can change significantly.

  FIG. 14B is a graph showing the characteristics of the interval evaluation time described above. In the graph of FIG. 14B, the horizontal axis represents the interval evaluation time, and the vertical axis represents the prediction accuracy when the car interval is estimated at the corresponding interval evaluation time. In the region where the interval evaluation time is close to zero (corresponding to the present time), the prediction accuracy of the car interval is small, and the prediction accuracy increases as the value of the interval evaluation time increases from there. And it becomes the maximum at a certain value, and after that, prediction accuracy falls, so that a value becomes large. The place where the prediction accuracy is maximized is considered to be in the vicinity of the arrival prediction time of the farthest call for all the units already described. The reason is that all the hall calls and car calls that have already occurred are included by the estimated arrival time for the farthest call, and all of them can be considered. Since there is no call that has already occurred before this time, there is no reliable information and the prediction accuracy is simply reduced.

  Therefore, by setting the estimated interval time at or near the estimated arrival time of the farthest call for all the cars, it is possible to perform cage interval evaluation with high prediction accuracy. As a result, it is possible to execute the assignment that approaches the equally spaced state more reliably, and to suppress the occurrence of a long waiting time.

  FIG. 15 is a functional block diagram of a second embodiment of the car interval evaluation value calculation unit different from FIG. In FIG. 15, the same elements as those shown in FIG. 12 are assigned the same reference numerals, and the description thereof is omitted. FIG. 15 differs from FIG. 12 in that the interval evaluation time is set according to the traffic flow mode generated at that time. Specifically, the traffic flow mode determination unit 406 determines the traffic flow mode occurring at that time as a representative traffic flow vector among the traffic flow vectors generated in the past in the building. Then, the interval evaluation time suitable for the traffic flow mode is referred to from the interval evaluation time database 407 for the traffic flow mode, and the value is set. Here, the interval evaluation time database 407 for the traffic flow mode is a database in which the traffic flow modes of the buildings extracted in advance and the interval evaluation times corresponding thereto are arranged in a table format. By using this, if the traffic flow mode is determined, the corresponding interval evaluation time can be set by referring to the table.

  Since there is a relationship between the traffic flow and the occurrence of the call, an appropriate interval evaluation time can be determined even if the traffic flow mode is used instead of the predicted arrival time of the farthest call, and the same effect can be expected.

  In this case, the time interval for setting the interval evaluation time is substantially equal to the time constant of the traffic flow change.

  FIG. 16 is a functional block diagram of a third embodiment of a car interval evaluation value calculation unit different from FIG. In FIG. 16, the same elements as those shown in FIG. 12 are assigned the same reference numerals, and the description thereof is omitted. FIG. 16 differs from FIG. 12 in that the interval evaluation time is determined based on the average one round time at that time. Specifically, the average one-round time calculation unit 408 obtains the average one-round time T for all elevators at that time based on the input information (input from the input information storage unit 2 in FIG. 1). Based on this average one round time T, the interval evaluation time setting unit 405 determines the interval evaluation time tref by the following equation.

tref = F (T) ………… (10)
Here, F (T) represents a function of T. The expression (10) is expressed as follows, for example.

tref = α · T (11)
Here, α represents a constant.

  As with traffic flow, the average round trip time is related to the occurrence of the call, so even if the average round trip time is used instead of the estimated arrival time of the farthest call, an appropriate interval evaluation time can be determined. Similar effects can be expected.

  Similar to the above-described method for setting the interval evaluation time, the method for creating the predicted trajectory is important for determining the interval evaluation value. The creation of the predicted trajectory is performed in the predicted trajectory calculation unit 401 in FIG. 12, FIG. 15, or FIG. 16, or the predicted route creation unit 411 in FIG.

  FIG. 17 is a functional block diagram of a fourth embodiment for a car interval evaluation value calculation unit 4 in place of FIG. Here, the expression “prediction route” is used, but this prediction route means the same as the prediction trajectory described so far. Details of FIG. 17 will be described later. Hereinafter, a method of creating a predicted trajectory that is one of the points of the present embodiment will be described with reference to FIG.

  FIG. 18 shows a process flow diagram of the entire predicted trajectory creation method. Hereinafter, the flow will be described. First, a variable K representing the elevator car name is set to 1 (FA01). Next, it is determined whether or not the K machine is a group management target (FA08). Since the elevator car separated from the group management for reasons such as exclusive operation is operated independently of the remaining elevators managed by the group, such processing is excluded from the target for creating the predicted trajectory. Next, it is determined whether the K machine is directional (FA02). Here, the determination as to whether or not the K machine is directional is equivalent to determining whether or not the K machine is in charge of a hall call or a car call. Therefore, if Unit K is in charge of hall call or car call (when Unit K is in a direction), it proceeds to the process of creating a multiple lap predicted arrival time table (FA03), and if neither call is in charge (K03) If the car is in a non-directional direction, the process proceeds to a predicted trajectory table creation process in the non-directional direction (FA05).

  In the multiple-round arrival prediction time table creation processing (FA03), an arrival prediction time table for a plurality of rounds, for example, three or more rounds, is created. The multi-round arrival prediction time table is represented by a variable called tar_table (i, j, c, K) below. Here, i is the floor, j is the direction, c is the number of laps, and K is the name of the car. Details of creation of the predicted arrival time table for a plurality of rounds will be described later with reference to FIG. When a predicted arrival time table for multiple rounds is created, a predicted trajectory table for a directional direction is created based on this table (FA04). Here, the predicted trajectory table in the directional direction is represented by two variables ir (t, K) and jr (t, K). ir (t, K) represents the car position of car K after t seconds from the current time, and jr (t, K) represents the car direction of car K after t seconds from the current time. Details of the prediction trajectory table creation in the directional direction will be described later with reference to FIGS. 25 and 26.

  In process FA02, when the K machine is non-directional, a predicted trajectory table for non-directional is created (FA05). Also in this case, the prediction trajectory table is represented by two variables ir (t, K) and jr (t, K). Details of creation of the predicted trajectory table in the non-direction will be described later with reference to FIG.

  When a prediction trajectory table is generated for directional or non-directional with respect to the K machine, one K is added (FA06), and for the new K machine, the process returns to the process FA08 and the above process is repeated. This is executed for all group management target machines (FA07).

  The major features of creating a predicted trajectory according to this embodiment are two points: 1) creating a multiple-round arrival prediction time table, and 2) creating a predicted trajectory by distinguishing between directional and non-directional directions. is there. For example, in the case of 1), the creation of the predicted arrival time differs depending on the number of laps (details will be described later). In the case of 2), as shown in FIG. 30, a directional car trajectory (trajectory FJ03 in FIG. 30) and a non-directional car trajectory (trajectory FJ02 in FIG. 30) are created in different shapes (details are given below). Will be described later). As a result, it is possible to create a more accurate predicted trajectory in consideration of the state of each unit and the traffic demand at that time.

  FIG. 19 is a process flow diagram for creating a predicted arrival time table for multiple laps according to an embodiment of the present invention. As described above, the arrival prediction time table is represented by the variable tar_table (i, j, c, K). The details are shown in FIG. 21, which will be described later, in which an example of the creation of a plurality of rounds of predicted arrival time table is shown. The first round of arrival prediction time table FG02, the second round of arrival prediction time table FG03, and the third round of arrival prediction time table FG04 is represented. According to the flowchart of FIG. 19, such an estimated arrival time table is created.

First, initial values are set for the predicted arrival time tar, the variable i representing the floor position of the K-th car, and the variable j representing the direction of the K-th car (FB01). Specifically, tar is set to zero, i is set to the current car position of car K, and j is set to the current car direction of car K. Next, a variable c representing the number of laps is set to 1 (FB02). This indicates that it is created from the estimated arrival time table for the first round. Next, a variable n representing the number of scans when scanning each floor in order is reset to zero. This n is added one by one (FB05), and is rotated in a loop until n exceeds 2 (n _max −1) (FB06). Here, n _max represents the total floor element through which the K machine passes. In the formula, it is expressed as follows.

n _max = Top floor of the service zone of Unit K-Bottom floor +1 (12)
The meaning represented by the value of 2 (n _max −1) will be described using the leftmost diagram FG01 in FIG. 22 as an example. In the leftmost diagram of FIG. 22, the row direction represents the floor, the column direction represents the upward direction, and the downward direction, and the display method is represented as a ring (wheel) that goes around the elevator in one turn. In this figure, there are 6 floors, and the upward direction of the 6th floor, which is the top floor, and the downward direction of the 1st floor, which is the bottom floor, are substantially meaningless and are omitted. As a result, the effective rank is 6 × 6-2 = 34. This agrees with the value when n _max = 6 in 2 (n _max −1). In other words, n represents the scanning floor when scanning one floor at a time, assuming the leftmost diagram in FIG. 22 for the process of one round of the elevator, and 2 (n _max −1) is per one lap. Represents the total number of floors scanned.

For each n, an arrival time table tar_table (i, j, c, K) for a plurality of rounds is calculated (FB04). This is executed by an arrival time table calculation routine described later (described later with reference to FIG. 20). This processing is performed 2 (n _max −1) times as described above (FB06), and after that, calculation of the next round is started, so one variable c representing the number of rounds is added. The In this way, the arrival prediction time table tar_table (i, j, c = 1, K) for c = 1 (first round), the arrival prediction time table tar_table (i, j, c) for c = 2 (second round). c = 2, K) is calculated. Further, an arrival prediction time table tar_table (i, j, c = 3, K) of c = 3 (third lap) is calculated, and processing is performed until the arrival arrival time of the final arrival floor in each lap exceeds t _max. It repeats (FB08).

  Through the series of processes described above, it is possible to create a predicted arrival time table of a plurality of rounds as shown in FIG. Details of the predicted arrival time table calculation routine (FB04 in FIG. 19) will be described below with reference to FIG.

  FIG. 20 shows a process flow diagram of the predicted arrival time table calculation routine (FB04 in FIG. 19). First, I will explain the general processing flow in words. 1) The variable of the estimated arrival time is tar. 2) Set the next moving floor (upward if the floor is -1, if down, the floor is +1). 3) It is determined whether the number of laps is the first lap or after the second lap. 4) In the first round, if there is a call stop on the target floor, the stop time is added to tar, and if there is no call stop, the stop probability is added to tar. 5) In the second and subsequent rounds, the stop probability is added to tar. 6) Add the travel time to the next moving floor to tar. 7) The value tar_table (i, j, c, K) in the predicted arrival time table for the target floor is set to tar.

  The rough flow of processing will be described in detail with reference to FIG. Here, the floor to be scanned with respect to the No. K machine is i, and the direction is j. First, in each scanning process of the target K machine (scanning the floor in a ring shape), it is determined whether the car direction j is upward (FC01). In the upward direction, the floor obtained by subtracting 1 from the variable i representing the car position at the time of scanning becomes the next moving floor i2 (FC02). In the downward direction, the floor obtained by adding 1 to i becomes i2 (FC03). The next moving floor i2 determines the direction reversal floor (for example, the top floor or the bottom floor) of the K machine (FC04). If i2 is the direction reversal floor, the direction j2 on the next moving floor is set to the opposite direction of j (FC05). Thus, it is one point of the present invention that the direction reversal floor is set for each unit. If i2 is not the direction inversion floor, i2 is set in the same direction as j (FC14).

  It is determined whether the floor / direction (i, j) being scanned is the service floor of the K-th unit (FC06). If the floor / direction (i, j) is not the service target floor, there is no call stop, so the processing related to the following stop is skipped and the processing jumps to the processing FC11.

  If the floor / direction (i, j) is a service target floor, it is next determined whether or not the number of laps c for creating the predicted arrival time table is two or more (FC07). If c is after the second lap, it is considered that there is no call stop during that lap, and the expected stop time calculated based on the stop probability is used for all service target floors (FC10). This corresponds to the predicted stop time predicted by the probability. This is one point of this embodiment. Specifically, the expected stop time value in the (i, j) floor / direction is added to the variable tar representing the estimated arrival time.

  If c is the first lap, determine whether there is a stop due to hall call or car call for the floor / direction (i, j) being scanned (FC08), and if there is a call stop, that floor Is added to tar (FC09). If there is no call stop, the expected stop time value in (i, j) floor / direction is added to tar (FC10). Even in the first round, one of the points of the present invention is that the expected stop time value is taken into consideration for the service target floor where there is no call stop. In addition, the stop time and the expected stop time value of each floor / direction are updated each time corresponding to the change in traffic flow (FC15). For example, the expected stop time for each floor / upward increases when the attendance is busy, and the expected stop time for each floor / downward increases during the first half of lunch.

  When the processing related to reflecting the call stop is completed, the travel time tmv (i2, j2) required for moving from (i, j) floor / direction to (i2, j2) floor / direction is added to tar. (FC11). As a result, the stop time and the travel time are added, and the predicted arrival time in the (i2, j2) floor / direction of the travel destination is calculated. i is updated to i2 and j is updated to j2 (FC12), and the value tar_table (i, j, c, K) of the new (i, j) floor / direction predicted arrival time table is set to tar (FC13). ). The above processing is the processing of the predicted arrival time table calculation routine, and this is recursively executed by changing i, j, c, K in the loop processing of FIG. 19, so that the predicted arrival time table tar_table (i, j , C, K) is completed.

  The characteristics of the estimated arrival time table for multiple rounds are summarized as follows. 1) A direction reversal floor is set for each unit. 2) For the estimated arrival time table for the first lap, the stop time for a call that has occurred (hall call, car call) and the expected stop time value for a call that has not occurred are used together. 3) For the second and subsequent laps, the expected stop time value is used assuming that all calls have not occurred. 4) The above stop time and expected stop time value have set values for each floor and direction. 5) The above stop time and expected stop time value change (update) according to the traffic flow. Since the predicted trajectory in the position example of the present invention is created based on the predicted arrival time table of multiple laps having such characteristics, it is accurate and predictable according to the characteristics of each unit, call stop status, traffic flow status A highly accurate predicted trajectory can be created. As a result, it is possible to accurately evaluate the car interval or the like based on the predicted trajectory and reduce the waiting time.

  Hereinafter, a specific example of the arrival prediction time table for a plurality of rounds will be described with reference to FIGS. 21 and 22. First, FIG. 21 shows a table of (a) stop time table, (b) stop probability table, and (c) expected stop time value for each floor and direction. First, the stop time table of FIG. 21A represents stop times for each floor and direction. In this example, the same stop time (8 seconds) is used for all floors and directions, but different stop times may be set for each floor and direction.

  The stop probability table in FIG. 21B represents the stop probability for each floor and direction. For example, the probability of stopping in the third floor upward direction is 0.6, which indicates that the probability of stopping by a call when the elevator makes one round in a ring shape is 0.6. In the example in the figure, the probability of stopping in the upward direction is different from that of the downward direction. This indicates that the traffic demand at that time tends to stop in the upward direction. Thus, the stop probability reflects the traffic demand (or traffic flow) at that time, and the value of each floor / direction of the stop probability changes according to the change of the traffic demand.

  The expected stop time value table in FIG. 21C represents the expected stop time value for each floor and direction. This expected stop time value is calculated by multiplying the stop time by the stop probability. In this embodiment, this value is used as the predicted stop time (expected value of stop time) for the floor where no hall call or car call is generated.

  FIG. 22 shows a specific example in which a predicted arrival time table for a plurality of rounds is calculated using the stop time table, stop probability table, and expected stop time value table shown in FIG. First, the leftmost diagram (FG01) in FIG. 22 shows the situation of the first unit at the present time. Unit 1 (FG05) is located on the second floor (FG05), the third floor is the upward hall call (FG06), the fifth floor is the car call (FG07), and the fifth floor is the downward hall call (FG08). ) Therefore, Unit 1 stops in the third floor upward direction, the fifth floor upward direction, and the fifth floor direction.

  The arrival prediction time table for the first, second, and third laps for the first car is represented as a table indicated by reference numerals FG02, FG03, and FG04 in FIG. First, since the estimated arrival time table (FG02) for the first round is currently located in the second floor upward direction, it starts from the third floor upward direction (FG09), goes around in a ring shape and ends in the second floor upward direction It is shaped (FG10). The predicted arrival time in the third floor upward direction is 2 seconds because of movement only, and the predicted arrival time in the fourth floor upward direction is stopped at the third floor, so that 8 + 2 = 10 seconds is added to become 12 seconds. The estimated arrival time in the upward direction on the fifth floor is 18.8 seconds by adding 4.8 + 2 = 6.8 seconds depending on the expected stop time value. The estimated arrival time in the 6th floor downward is 28.8 seconds, because 8 + 2 = 10 seconds is added because there is a stop in the 5th floor upward. The estimated arrival time in the 5th floor direction is 32.4 seconds by adding 1.6 + 2 = 3.6 seconds depending on the expected stop time value. Thereafter, by repeating the same, the arrival prediction time table for the first round can be calculated. This series of processing is the same as the flowchart of FIG. The second arrival prediction time table (FG03) also starts from the third-floor upward direction (FG11), and goes round in a ring shape and ends in the second-floor upward direction (FG12). In this second round, the estimated arrival time is calculated using all expected stop time values. For example, the predicted arrival time on the fourth floor upward in the second round is calculated as 73.6 seconds by adding 4.8 + 2 = 6.8 seconds to the predicted arrival time 66.8 seconds on the third floor upward. The Similarly, the estimated arrival time table (FG04) for the third round is also calculated as shown in the figure.

  Next, details of the predicted trajectory table creation in the directional direction described in FIG. 18 (processing FA04 in FIG. 18) will be described. This creation process is represented by the process flow of FIGS. 25 and 26, which will be described later. Before explaining the details of this flow, first, an overall rough process flow will be described with reference to FIGS. 23 and 24. explain.

  First, FIG. 23 is an example of a predicted trajectory table that is finally created. As shown in the figure, the predicted position of each unit for each time (FH01 column), direction for each unit (unit No. 1 is the field of code FH02, unit 2 is the column of code FH03), direction Data (in case of FH05, Unit 1) is stored. When this data is traced on the time axis, a predicted trajectory for each unit can be created.

  The predicted trajectory table in FIG. 23 is created from the data of the predicted arrival time table for multiple rounds in FIG. Specifically, the predicted position at each time can be calculated by interpolation from the predicted arrival time of each adjacent floor and direction. FIG. 24 shows an image of this calculation method.

  In FIG. 24, the diagram on the right represents a graph in which the horizontal axis represents time and the vertical axis represents the floor position. A line FF01 on the graph represents each floor / direction ix of the K-unit and its predicted arrival time tx by a point (tx, ix) on a two-dimensional coordinate, and each point is connected by a line segment. Yes. This line FF01 becomes the original shape of the predicted trajectory. The figure on the left shows the current position / direction (FF02) of K-unit, and it can be seen that it is on the second floor. Therefore, the line FF01 in the right diagram also has a point FF02 plotted in the second floor upward direction at time zero (current time).

  In the line FF01, a point FF03 is a point representing the position in the third floor upward direction and the predicted arrival time, and a point FF04 represents a point representing the position in the next fourth floor upward direction and the predicted arrival time. Assuming that the coordinates of the point FF03 are (tA, iA) and the coordinates of the point FF04 are (tB, iB), an arbitrary point (t, ir (t, K)) on the line segment connecting these two points. Can be expressed as:

ir (t, K) = {(iB−iA) / (tB−tA)} (t−tA) + iA (13)
That is, if t is determined under the condition of tA ≦ t ≦ tB, the predicted position ir (t, K) of the car at the time t can be obtained. Further, the predicted car direction jr (t, K) can also be obtained from the slope of the section. Here, the two points are defined as the 3rd floor upward direction, the 4th floor upward position, and the estimated arrival time, but if this is shifted one by one, for all the times t, the corresponding car The predicted position ir (t, K) and the direction jr (t, K) can be obtained. This concept is the basis for creating the predicted trajectory table of FIG. 23 from the data of the predicted arrival time table of multiple rounds as shown in FIG. To summarize, from the data of two adjacent floors / directions and the predicted arrival time in the predicted arrival time table of multiple rounds as shown in FIG. Calculate ir (t, K). The floor and direction are shifted one by one, and t, ir (t, K), and jr (t, K) are obtained over a plurality of circuit areas. The result is a predicted trajectory table as shown in FIG.

  In the foregoing, the general idea of creating a prediction trajectory table has been described. Hereinafter, a specific processing flow will be described with reference to FIGS. FIG. 25 and FIG. 26 show the flow of creating the predicted trajectory table divided into two diagrams.

First, for FIG. 25, initial values are set first (FD01). Here, the time variable parameter t in the prediction trajectory table is set to zero, and the variable parameter c representing the number of laps is set to zero. In addition, the flag variable z indicating whether or not the calculation of the initial floor (requires special processing to add the time origin) is set to zero, and the variable parameter i representing the position of the scanning floor is set to the current car position of the car K The variable parameter j indicating the direction of the scanning floor is set to the current direction. Next, a variable n representing the number of scanned floors is set to zero. The meaning of n is the same as n used in FIG. 19, and when the value of n reaches 2 (n _max −1), it means that the floor for one round has been scanned. This addition process of n is executed in process FD24 in FIG. 26, and the determination of whether or not one round has been reached (whether n = 2 (n _max −1) has been reached) is executed in processes FD13 and FD25. . Next, it is determined whether the flag variable z (is zero (FD03). If the flag variable z is zero, the first floor is processed, and only in this case, a different process is performed.

  When the flag variable z is zero, the variable iA is set to the value of i (current scanning floor), the variable jA is set to the value of j (current scanning floor direction) (FD04), and tA is set to zero (FD05). ). If the flag variable z is not zero, the variable iA is set to i and the variable jA is set to j (FD06), and tA is set as shown in the following equation based on the arrival prediction time table for multiple rounds (FD07). .

tA = tar_table (iA, jA, c, K) (14)
That is, tA is set to the estimated arrival time of the c-th lap of the floor (iA, jA) with respect to K-unit. These variables iA and tA correspond to iA and tA shown in FIG. 24 (corresponding to point FF03 in FIG. 24). JA represents the direction with respect to iA. That is, the time t and the predicted arrival time ir (t, K) between the two points (tA, iA) and (tB, iB) are calculated, and the starting point of these two points is determined.

Next, it is determined whether or not the current scanning floor direction j is upward (FD08). If it is upward, the variable iB is decremented by 1 (FD09), and if downward, iB is increased by 1 (FD10). . Here, the variable iB corresponds to iB shown in FIG. 24 (point FF04 in FIG. 24). It represents the position of the end point of the two points. It is determined whether iB is the direction reversal floor of K machine (FD11). If it is the direction reversal floor, variable jB is set in the opposite direction to j (FD12). Otherwise, variable jB is the same as j. The direction is set (FD27). Further, it is determined whether or not the variable n representing the number of scanned floors has reached 2 (n _max −1) (FD13). As described above, this determination is a determination as to whether or not scanning of the number of floors for one round is being performed. If the number of scanned floors is less than one round, tB is set to tar_table (iB, jB, c, K) (FD14). If the number of scanned floors is one round, tB is set to tar_table (iB, jB, c + 1, K) (FD15). For the latter, since the number of scanned floors is one round, refer to the value from the arrival prediction time table tar_table (iB, jB, c + 1, K) of the next round obtained by adding 1 to tB and c. Means. The subsequent processing continues to (a) of FIG.

  The processes after (a) in FIG. 26 will be described. Since (tA, iA) and (tB, iB) have been set, the value of ir (t, K) is calculated by equation (13) (FD16). This is the same as obtaining the value of the point (t, ir (t, K)) on the line segment (tA, iA)-(tB, iB) in FIG. . Next, the direction jr (t, K) is determined in the same direction as jA (FD17).

When the calculation of ir (t, K) and jr (t, K) is completed for t, t is updated by adding Δt to t (FD18), and ir (t, K), jr (t, K). Here, when t exceeds _tmax , the process of creating the predicted trajectory table is terminated (FD19). t _max corresponds to the time width of the predicted trajectory and is set to a predetermined value in advance. The specific value of t _max is desirably after the interval evaluation time tref described with reference to FIG. 12, and is therefore preferably longer than the estimated arrival time of the farthest call. In a case where the control is performed with a long-term predicted trajectory to some extent, it is desirable that it is at least one lap time of the elevator (for example, 60 seconds or more).

When t is tB or more (FD20), i is set to iB and j is set to jB (FD22). Considering this in FIG. 24, this corresponds to starting t from tA and proceeding by adding Δt, and when tB is exceeded, processing for moving to the next section is performed. At this time, it is determined whether or not the flag variable z is zero (FD22). If it is zero, z is set to 1 (FD23). Since this process has passed through the section starting from time zero, it is meaningful to indicate this by the flag variable z. The fact that the section has changed means that the scanned floor on the estimated arrival time table has moved to the next floor, and n is incremented by 1 (FD24). Further, it is determined whether or not n exceeds 2 (n _max −1), and if it exceeds, +1 is added to the variable c representing the number of laps. This means that the next lap estimated arrival time table is entered.

  Finally, returning to FIG. 18, the details of the predicted trajectory creation process (FA05 in FIG. 18) in the non-direction (when there is no handling hall call or car call) shown in FIG. 18 will be described. FIG. 27 shows a flowchart of the predicted trajectory creation process in the non-direction.

Hereinafter, the process of FIG. 27 will be described. First, as an initial setting, a variable parameter t representing time is set to zero (FE01). Next, a variable ir (t, K) representing the car position of the car K after time t, a variable jr (t, k) representing the car direction, ir (t, K) as the current car position, jr (t, t, k) K) is set to no direction (FE02). Δt is added to the value of t (FE03), and the processes FE02 and FE03 are repeated until t exceeds _tmax (FE04). The non-directional electronics carry out such processing on the assumption that the standby at the current position is continued.

  As described above, by the series of processes shown in FIGS. 18 to 20 and FIGS. 25 to 27, first, data of a plurality of rounds of arrival prediction time table (an example in FIG. 22) is created, and further a predicted trajectory table (an example in FIG. 23). Data can be created. As shown in FIG. 23, the data of the predicted trajectory table represents the position and direction of each elevator that is ahead of each time from the present time, and this corresponds to the predicted trajectory. The creation method shown in the present embodiment is a characteristic creation method in accordance with the actual situation in order to increase the prediction accuracy, and therefore the prediction trajectory also shows a characteristic form. Hereinafter, the characteristics of the predicted trajectory that can be created according to the present embodiment will be described with reference to FIGS.

  FIG. 28 is an example of a predicted trajectory created by the predicted trajectory creation method shown in the present embodiment. Here, in the right figure FI01, the respective predicted trajectories for the two cars of the first car and the second car are shown. The predicted trajectory of the first car is represented by a solid line FI02, and the predicted trajectory of the second car is represented by a broken line FI03. In addition, the current state of the first car, that is, the position and direction of the car, the hall call, and the handling status of the car call are shown in the leftmost figure with the reference numeral FI04. The current state of the second car is indicated with the reference numeral FI05. It is represented by the attached left figure.

  First, looking at the predicted trajectory FI02 of Unit 1, in the first lap, that is, the trajectory from the second floor upward direction to the first floor upward direction, the stop by the hall call and the car call is a change in the inclination angle. Appears. However, the second and subsequent rounds are represented as straight lines FI106 and FI107, and are represented only by the expected stop time value based on the stop probability. As a result, as shown in the figure, with respect to the trajectory of the first round, the trajectories after the second round have the same shape on each circumference. Moreover, since the stop probability is reflected, the upward inclination (FI06) and the downward inclination (FI07) with respect to the trajectories after the second round are different. The slope of the trajectory reflects the probability of stopping, and the value of the stopping probability is set corresponding to the change in traffic flow, and the slope changes according to the traffic flow. If the stop probability is not reflected, the inclination of the predicted trajectory is determined only by the elevator speed, so that the inclinations of the upper and lower trajectories are the same or symmetrical. In the predicted trajectory of FIG. 28, the data shown in FIG. 21 is used as the data of the stop time, stop probability, and expected stop time value.

  In addition, each predicted trajectory indicates a trajectory whose direction is reversed between the uppermost floor and the lowermost floor for each circumference (after the second week). One of the reasons for creating a trajectory that reverses direction on the end floor (top floor or bottom floor) in this way is that the evaluation based on the predicted trajectory of each unit is aimed at a time-equally spaced state, This is particularly effective during congestion. In the case of congestion, it is expected that many hall calls and car calls will be generated, and as a result, the actual operation trajectory also becomes a trajectory whose direction is reversed at the end floor as shown in FIG. Therefore, the predicted trajectory is created so as to be a trajectory whose direction is reversed at the end floor.

  FIG. 29 is an explanatory diagram of a predicted trajectory that waits with an emphasis on a quiet time. When it is quiet, it is not always necessary to have a predicted trajectory that reverses at the end floor as shown in FIG. 28. As shown in FIG. 29, the floor or time at which the call service ends is predicted, and after that, it waits at that position. A predicted trajectory may be created. That is, the predicted trajectory creation means creates a predicted trajectory of an elevator that does not have the assigned hall call and car call so as to be parallel to the time axis. In this case, the final car call floor is estimated from the traffic demand at that time, and after that, a predicted trajectory waiting at that position is created. For example, in the case of FIG. 29, the first car (predicted trajectory FQ01) predicts that the second floor down direction will be the final car call, and thereafter, the trajectory (FQ02) stands by in the second floor down direction. Similarly, the second car (predicted trajectory FQ03) is a trajectory (FQ04) waiting in the upward direction on the fifth floor, predicting that the upward direction on the fifth floor will be the last car call.

  Returning to FIG. 28, when looking at the predicted trajectory FI02 of Unit 1 and the predicted trajectory FI03 of Unit 2 in FIG. 28, although the positions are currently spaced at almost equal intervals, It is predicted that it will approach in the future depending on the situation. By drawing the predicted trajectory in this way, it is possible to predict in advance the situation that will occur in the previous time by creating the predicted trajectory of each unit, and by reflecting this information in the hall call allocation, The trajectory can be appropriately controlled. For example, the future of two units can be evaluated by evaluating the interval between the trajectory of the predicted trajectory FI02 of Unit 1 and the predicted trajectory FI03 of Unit 2 (quantitative evaluation is possible by the area sandwiched between the trajectories) before and after the allocation of each unit. The degree of deviation can be evaluated. Further, by observing the trend (temporal change) between the two predicted trajectories, it is possible to evaluate the temporal change in the trajectory interval (for example, approaching the dango operation state).

  As described above, the predicted trajectory according to the present embodiment takes into account the probable stop for a call that has occurred and the probabilistic stop for a call that has not occurred. Since this is reflected, a prediction trajectory with high prediction accuracy can be created.

  FIG. 30 shows a second example of the predicted trajectory by the creation method shown in this embodiment. The prediction trajectory of Unit 1 is represented by a solid line FJ02, and the prediction trajectory of Unit 2 is represented by a broken line FJ03. In addition, the current situation of Unit 1-car position / direction, hall call, and car call handling status are shown in the figure with the symbol FJ04, and the situation of the current Unit 2 with the symbol FJ05. It is represented by

  As can be seen from the code FJ04 in FIG. 30, Unit 1 does not have a call service at present and is in a standby state on the fourth floor. Accordingly, the predicted trajectory FJ02 of the first car is a trajectory that continues to be in a standby state on the fourth floor. The trajectory of this non-directional car is first skipped to the predicted trajectory table creation process (FA05 in FIG. 18) when the non-directional car is in the non-direction by the overall process flow of the predicted trajectory creation shown in FIG. Detailed data is created by the processing flow of the non-directional unit shown.

  FIG. 31 shows a third example of the predicted trajectory by the creation method shown in this embodiment. This example shows an example of a predicted trajectory when the stop probability is different for each floor and direction. The left side of FIG. 31 (FK04) shows the current car position and direction, and the first car (FK05) and the second car (FK06) are in the positions and directions shown in the figure. In this figure (FK04), the calls handled by each unit are omitted.

  The figure on the right side of FIG. 31 (FK01) represents the predicted trajectories of the two elevators. The predicted trajectory of the first car is represented by a solid line FK02, and the predicted trajectory of the second car is represented by a broken line FK03. The difference between the predicted trajectory of FIG. 31 and the predicted trajectory of FIG. 28 is that the predicted trajectory of FIG. 31 has a different probability of stop for each floor / direction, and therefore the inclination of the predicted trajectory differs for each floor / direction. Usually, the trend of users on each floor and direction is different, so the stop probability is also different. Therefore, it can be said that the shape of the predicted trajectory in FIG. 31 reflecting the respective stop probabilities for each floor and direction is the most accurate trajectory. Furthermore, since the probability of stopping at each floor and direction changes as the traffic flow changes, prediction trajectories of various shapes are created reflecting the change in the traffic flow.

  Furthermore, in the predicted trajectory based on the stop probability for each floor / direction as shown in FIG. 31, the shape of the predicted trajectory is accurately determined for each floor / direction, so that it is more accurate when evaluating the predicted trajectory interval of each unit. High evaluation is possible. Therefore, by using such a predicted trajectory based on the stop probability for each floor and direction, the accuracy of interval evaluation is improved, more appropriate allocation is possible in controlling the interval, and long waiting time can be reduced. it can.

  FIG. 32 shows a fourth example of the predicted trajectory by the creation method shown in this embodiment. The difference between the predicted trajectory in FIG. 32 and the predicted trajectory in FIG. 28 is that the probabilistic stop element determined by the call stop and stop probability on the predicted trajectory is shaped. In the predicted trajectory FL02 of FIG. 32, the call stop portion is represented by the horizontal part element (FL03) of the fifth-floor / downward trajectory of the first round, and the stochastic stop portion is 4 in the first round. It is represented by the element (FL04) in the horizontal part of the trajectory in the upper and lower directions. An element (FL05) represented by an oblique line segment represents a downward movement state.

  The advantage of the predicted trajectory shown in FIG. 32 is that it is possible to create a predicted trajectory that reflects the actual shape with higher accuracy by dividing the three elements of stop by call, probabilistic stop, and movement in detail. is there. The actual movement of the elevator is not a trajectory represented by an oblique line on the time axis as shown in FIG. 28, but has a configuration including a horizontal portion because there is a stop. Therefore, the predicted trajectory in FIG. 32 is a trajectory that reflects the actual situation. In addition, the predicted trajectory of FIG. 32 also has an advantage that the three elements of call stop, probabilistic stop, and movement can be easily distinguished visually. With such a shape, it is possible to understand at a glance where the call stop occurs, what the situation of the stop probability of each floor / direction is, and how the call stop and stop probability influence the trend of the trajectory. For example, if the predicted trajectory of Unit 1 and Unit 2 indicates that it will be dango driving in the future, whether it is caused by a bias in the floor / direction of the probability of stopping or whether it is caused by how to handle call stopping I can understand.

  For the predicted trajectory as shown in FIG. 32, the data creation of the predicted trajectory table may be divided into stop, probabilistic stop, and movement. For example, in the process of preparing the arrival time table for multiple laps, it passes through the process of adding the call stop time or the expected stop time and adding the travel time, but if each is stored separately in the data Good.

  FIG. 33 shows a fifth example of the predicted trajectory by the predicted trajectory creation method shown in the present embodiment. The difference between the predicted trajectory in FIG. 33 and the predicted trajectory in FIG. 28 is that the predicted trajectory in FIG. 33 is represented by a rough trajectory for each of the upward and downward directions. Specifically, the predicted trajectory in FIG. 33 can be created by connecting the points on the uppermost floor and the points on the lowermost floor of the predicted trajectory in FIG. This corresponds to a predicted trajectory based on data obtained by integrating call stop, probabilistic stop, and travel time of each floor / direction by direction.

  The predicted trajectory in FIG. 33 has a rough shape, but the necessary data may be the predicted arrival times of the uppermost floor and the lowermost floor in each lap, so that the number of necessary data can be greatly reduced. Therefore, when using a microcomputer that is inexpensive but has low processing capability, it can be said that the application of such a simplified predicted locus is effective.

  In addition, every time hall call allocation is performed, when the predicted trajectory data (predicted trajectory table) is recorded and left as a history of allocation control contents, for example, when trying to record a history of data for one week, The amount of recording becomes enormous. In such a case, if the data is stored in the form of a predicted locus as shown in FIG. 33, the data recording amount can be greatly reduced. It is also conceivable that the predicted trajectory actually used in the control has a detailed shape as shown in FIG. 28 or 31 and that the predicted trajectory in FIG. 33 is used only for data storage purposes. In this case, the predicted trajectory data for storage can be compressed without reducing the accuracy of the predicted trajectory in the control. The data stored and recorded in this manner is effective because it can be analyzed based on the data of what kind of trajectory is predicted when the allocation evaluation is checked later. An embodiment configuration for recording the predicted trajectory data will be described later with reference to FIG.

  FIG. 34 shows a sixth example of a predicted trajectory by the predicted trajectory creation method shown in the present embodiment. The prediction trajectory of FIG. 34 differs from the prediction trajectory of FIG. 28 in that the prediction trajectory of FIG. 34 is a prediction trajectory corresponding to a high-rise zone (a zone of a plurality of non-stop floors in the middle).

  FIG. 34A shows an example in which a predicted trajectory in a high-rise zone is created according to the same method as FIG. The diagram on the right side of FIG. 34 (a) shows the current car position / direction, and the area of the hatched portion FN02 (for the 8th floor from the 2nd floor to the 10th floor) is a high-rise zone. The predicted trajectory for the elevator in the right diagram in FIG. 34A is the predicted trajectory FN01 in the left diagram in FIG. In the high-rise zone, since call stop does not occur, the stop probability is zero, and the inclination of the predicted trajectory is steep and constant. The predicted trajectory shown in FIG. 34 (a) is an example when all floors are displayed.

  On the other hand, FIG. 34 (b) shows a case where the high-rise zone portions are collectively represented on one floor. Specifically, as shown in the right diagram of FIG. 34 (b), the second floor to the tenth floor are represented by one floor FN04. The predicted trajectory in this case is as shown in the trajectory FN03 in the left diagram of FIG. As shown in the figure, even if the high-rise zone is represented by one floor, the predicted trajectory can be accurately represented by matching the passing time of the zone. It can be confirmed that the passage times of the zones match from the fact that the passage times of the high zone in FIG. 34 (a) and the passage times of the high zone in FIG. 34 (b) match.

  The advantage of the predicted trajectory as shown in FIG. 34B is that redundant data can be removed first. For example, in the example shown in the figure, since no calls are generated from the second floor to the tenth floor, the trajectory information between them is redundant. The important areas are the floors on the upper and lower sides, and the locus can be evaluated by closing up the important locus portion by the way of representing the locus as shown in FIG. Further, since data from the second floor to the tenth floor can be deleted, there is an effect of compressing the data.

  FIG. 35 shows a seventh example of the predicted trajectory by the creation method shown in this embodiment. The difference between the predicted trajectory in FIG. 35 and the predicted trajectory in FIG. 28 is that the predicted trajectory in FIG. 35 is a predicted trajectory when the service target floor zones of the elevators are uneven. Specifically, the service target floor zone of Unit 1 is all floors from the B1 floor to the 14th floor as shown in the leftmost figure FP03, whereas the service target floor zone of Unit 2 is from the left As shown in the second figure FP04, the range is from the second floor to the tenth floor. Thus, even if the service target floor zones of each unit are not uniform, a predicted trajectory reflecting the service target floor zone of each unit is created as shown in the prediction trajectory on the right. Specifically, the predicted trajectory of the first car is a trajectory FP01, and the predicted trajectory of the second car is a trajectory FP02.

  The reason why such a predicted trajectory can be created is that the direction inversion floor is identified for each unit in the flowchart of the predicted arrival time table calculation routine shown in FIG. In the example of FIG. 35, the direction reversal floor of Unit 1 is the B1 floor and the 14th floor, and the direction reversal floor of Unit 2 is uneven by identifying the direction reversal floor for each unit, such as the 2nd floor and the 10th floor. A predicted trajectory corresponding to a floor can be created. As a result, it is possible to create a more accurate predicted trajectory according to the characteristics of each unit.

  The details of the predicted trajectory creation method (implemented by the predicted trajectory calculation unit in FIG. 12) and an example of creating a predicted trajectory have been described above. The prediction trajectory creation method according to the present embodiment is a creation method that is more realistic (reflects the actual situation) in order to increase the prediction accuracy, and as a result, has a variety of accuracy as shown in FIGS. A high predicted trajectory can be created. Therefore, the evaluation accuracy such as the evaluation of the interval between the predicted trajectories is good, the allocation of the car interval can be controlled more appropriately (control considering the uniform interval), the car interval is kept appropriate, and the long wait occurs. Can be suppressed.

  36 to 38 show details of the prediction interval value calculation process. Hereinafter, the calculation method of the prediction interval value will be described based on the flowchart of FIG. First, the phase time value tp of each car at the interval evaluation time tref is calculated using the predicted route of each car described above for the creation method (ST801 in FIG. 36). Here, the interval evaluation time tref is set by the setting method already described. Details of the calculation of the phase time value of each car will be described with reference to FIGS.

  FIG. 37 shows how the prediction interval value is calculated from the prediction route. Here, the group management by three elevators is shown, and the diagram on the left side of FIG. 37 represents the position and direction of the elevator car at the present time in a ring expression. 37, the first machine 610 is moving upward between the seventh and eighth floors, and the second machine 611 is moving upward between the fifth and sixth floors. The third machine 612 is moving downward from the third floor to the second floor. The diagram on the right side of FIG. 37 represents the predicted route of each car on the graph with the horizontal axis representing time and the vertical axis representing position. The origin of the time axis represents the current time. The current car position 600 of the first car, the car position 601 of the second car, the car position 602 of the third car, respectively, are shown, and the predicted route of each car from that position is shown (prediction of the first car) Route 603, Unit 2 predicted route 604, Unit 3 predicted route 605). The car position at the interval evaluation time tref (606 in FIG. 37) can be predicted from the predicted route of each car. For example, the predicted position / direction of Unit 1 at the interval evaluation time tref is 7th floor upward (607 in FIG. 37), and the predicted position / direction of Unit 2 is 4th floor upward (609 in FIG. 37). The position / direction is the 6th floor downward direction (608 in FIG. 37). The predicted interval between the cars can be obtained from the predicted position / direction of each car.

  FIG. 38 shows a process for obtaining the prediction interval. The diagram on the left side of FIG. 38 shows the predicted positions and directions of the three cars at the interval evaluation time tref shown in FIG. In the diagram on the right side of FIG. 38, the horizontal axis represents the phase time value, and the vertical axis represents the position. Here, the phase time value means a time value normalized by the time of one round (same as the period) (meaning a time value having the same meaning as the phase). ). This topological time value is obtained on the basis of an average one-round locus (703 in the right diagram of FIG. 38) of the elevator car with respect to the current traffic flow. By mapping the predicted position of each car shown in the left diagram of FIG. 38 on the average one-round locus, it can be converted into a topological time value. For example, the temporal phase value of the first machine 705 can be obtained as tp (k = 1) from the average one-round locus 703 in FIG. 38, and similarly, the temporal phase value of the second machine 704 is tp ( k = 2) The temporal phase value of the third machine 706 can be obtained as tp (k = 3). Thus, the reason for converting from the predicted position to the temporal phase value is to obtain the interval of each car as a temporal interval value in units of time.

  A temporal phase value is obtained from the predicted position of each car (ST801 in FIG. 36), and then each car is ranked in order of the magnitude of the topological time value (ST802 in FIG. 36). For example, in the case of FIG. 38, the magnitude of the topological time value of each car is as follows.

tp (k = 2) <tp (k = 1) <tp (k = 3) (15)
Therefore, when the label variable representing the rank is represented by m, the car with m = 1 is No. 2, the car with m = 2 is No. 1, and the car with m = 3 is No. 3.

  A predicted interval value Bm between the cars is obtained in accordance with the order of the order m of the topological time values (ST803). For example, in the case of FIG. 38, the prediction interval value between the cars of m = 1 and m = 2 is Bm = 1 (corresponding to the interval value of the section 707 in FIG. 38), and between the cars of m = 2 and m = 3 The prediction interval value Bm = 2 (corresponding to the interval value in the section 708 in FIG. 38). Similarly, the prediction interval value between the cars of m = 3 and m = 1 is Bm = 3 (corresponding to the interval value of the section 709 in FIG. 38). Expressed by mathematical formulas, each prediction interval value is expressed as follows.

Bm = 1 = tp (k = 1) −tp (k = 2) (16)
Bm = 2 = tp (k = 3) -tp (k = 1... (17)
Bm = 3 = {T-tp (k = 3)} + tp (k = 2) (18)
Here, T in Expression (18) represents the period of the average one-round locus.

  As described above, since the predicted interval of each car is obtained by the topological time value based on the average one-round locus for the current traffic flow, a more appropriate time interval corresponding to the traffic flow at that time is obtained. be able to. For example, since many hall calls going downward occur at the start of lunch, the average one-round trajectory has a gentle slope in the line segment on the lower side, and the topological time value per floor is higher than that on the upper side. Become longer. Therefore, for example, when two cars are separated by two floors in terms of distance, the interval value is evaluated to be different between the case where it is upward and the case where it is downward. The downward direction has a higher possibility of stopping, so it is evaluated that the same second floor is further away. Thus, the time interval can be evaluated appropriately according to the traffic flow.

  FIG. 39 is a comparison of the trajectory on the time axis of the elevator car in the case of three vehicle group management between the result before the control according to this embodiment and the result after the control according to this embodiment. FIG. 39A shows a trajectory on the time axis of the elevator car before the control according to this embodiment. Looking at this trajectory, it can be seen that the three trajectories overlap in some places and inefficient dango driving is occurring.

  On the other hand, FIG. 39B shows a trajectory on the time axis of the elevator car after the control according to the present embodiment. It can be seen that the trajectories of the three cars maintain the same phase as in the three-phase alternating current, that is, the time-equal state. In this way, since the time equidistant state can be maintained, the elevator car can arrive immediately regardless of the floor / direction of the hall call, and the occurrence of a long waiting time can be suppressed.

  Here, returning to FIG. 17, another configuration example for the car interval evaluation value calculation unit 4 illustrated in FIG. 2 will be described. The configuration of FIG. 17 is based on the deviation between the route between the target route and the predicted route (predicted trajectory) with reference to the target route (target trajectory) that is an ideal route (trajectory) in a time equidistant state. It has become a method to evaluate. Use of this target route is a feature of the configuration shown in FIG.

  Although details of the target route will be described later, in the configuration of FIG. 17, the target route becomes a detailed reference for equalization in time, so that each car can be equally spaced over a wide time domain. As a result, the above effects can be obtained with the configuration of FIG.

  First, based on the input information (input from the input information storage unit 2 in FIG. 2), an ideal time equidistant route is created for each car in the target route creation unit 410, and a predicted route creation unit is further created. At 411, a predicted route for each car is created. The method of creating the predicted route is the same as the method already described, and the creation of the target route will be described later with reference to FIGS. The route deviation calculation unit 414 calculates the deviation between the target route and the predicted route for each car. This deviation can be calculated, for example, by the area of the difference between the two routes. The car interval evaluation value calculation unit 415 calculates a car interval evaluation value for each car (a car interval evaluation value when each car is temporarily assigned to a hall call) based on the calculated inter-route deviation. Based on this car interval evaluation value, a hall call is assigned to the car with the best evaluation value. The average one-round time calculation unit 412 calculates the average one-round time based on the input information, and the adjustment reference time setting unit 413 determines the value of the adjustment reference time based on the average one-round time T. The set adjustment reference time is used when creating the target route. Details of this will also be described later with reference to FIGS.

  Hereinafter, details of the allocation evaluation control (one of evaluation functions for future calls) based on the target route shown in FIG. 17 will be described. As already described with reference to FIG. 17, the allocation evaluation control based on the target route is based on three elements: the target route creation unit 410, the predicted route creation unit 411, and the inter-route deviation calculation unit 412.

  First, the operation image (control principle) of the target route control will be described with reference to FIGS. FIG. 40 is a diagram illustrating an example of a control image of target route control. In the figure, the diagram on the left side is a diagram conceptually showing the cross section (vertical direction) of the hoistway in the building and the state of the elevator car moving in the shaft. In the figure on the right side, the horizontal axis (A01) represents the time axis, and the vertical axis (A02) represents the floor axis of the building (the axis of the vertical position of the building), and each elevator on the time axis. It represents the trajectory of the car's operation (generally called an operation diagram). In the figure, the situation of management of two elevator groups is shown as an example. From the figure on the left, Unit 1 (the car described as 1) reverses on the 1st floor and runs upward, and Unit 2 (the car described as 2) descends from the 2nd floor. Yes. When this situation is seen in the operation diagram on the right side, the first unit (A03) and the second unit (A04) are moved down in the left direction from the current axis (A02). You can see how it is located on the floor. That is, in the right travel diagram, the locus of each elevator car on the left side from the present time represents the actual locus. For example, the actual trajectory of Unit 1 is the trajectory of A031, and the actual trajectory of Unit 2 is the trajectory of A041.

  The point of a present Example exists in the locus | trajectory drawn on the time axis of the future direction on the right side from the present time in a service diagram. This represents the 'target trajectory' that each car should pass in the future. Hereinafter, this target locus is referred to as a “target route”. The feature of the assignment control by the target route is that the operation (more precisely, the assignment) of each elevator car is controlled to follow the target route. Specifically, the target route of each car is A032 for the first car, and A042 for the second car. The introduction of this target route, that is, the trajectory that is the target (or reference) that each unit should pass on the time axis, is a unique feature of the present invention that is not present in the conventional group management control.

  FIG. 41 is a diagram showing a state in which the allocation of elevator cars to hall calls is determined according to the target route. FIG. 41 is basically the same diagram as FIG. 40, the left side shows the state of the elevator on the vertical cross section of the hoistway, and the right side shows the operation diagram. First, suppose that a new hall call occurs in the upward direction on the third floor. See the diagram on the left side of the figure. In response to this hall call, the group management control assigns an appropriate number of either the first car (B03) or the second car (B04). Here, focus on the movement of Unit 1 (B03). The target route of Unit 1 is the locus of B032. The predicted route of Unit 1 (predicted trajectory from the present time to the next time point, hereinafter referred to as “predicted route”) is the B033 route (predicted route 1) when a new hall call is passed without being assigned. Therefore, when a new hall call is assigned, the route is B034 (predicted route 2). Here, in the group management control of the present embodiment, the movement of each car is to move according to the target route. Therefore, the closer to the target route is the predicted route 1 of B033, that is, the route that passes the hall call without being assigned, and the hall call is not assigned to the first car. As a result, the actual trajectory of Unit 1 operates so as to follow the target route.

  According to the control by the target route, the target route is drawn so that each elevator car becomes a trajectory in a time equidistant state in the future. As a result, the actual car trajectory follows the target route, and as a result, it is possible to control each car so as to keep the trajectories at regular intervals in a stable manner over the long term. For example, in the case of FIG. 41, the actual trajectories of Unit 1 (B03) and Unit 2 (B041) up to the present time: The trajectory of Unit 1 (B031) and the trajectory of Unit 2 (B041) are approaching and dangling. You can see that it is in a state. Here, when a new hall call generated in the upward direction on the third floor is assigned to the second car, the distance between the first car (B03) and the second car (B04) is still close, and the dango operation continues. However, if Unit 1 and Unit 2 are separated and controlled along the target route set so that the trajectory of each car is equally spaced in time, it will not be assigned to Unit 1 (B03), and the target route As you can see, the time is approaching the same time interval.

  The features of the control principle of the elevator group management system according to this embodiment will be summarized below based on FIG. 40 and FIG.

  1) As shown in FIG. 40, a target locus and target route on the time axis are set for each car.

  2) As shown in FIG. 41, the target route is compared with the predicted route so that the trajectory of each car follows the target route, and the allocation of hall calls to the car that is closer to the target is determined.

  3) As a result, each car operates to follow the target route.

  4) Here, the target route is basically set so that the trajectories of each car are equally spaced in time, so that each car is stable in the long term and is in a time equally spaced state. Controlled.

  FIG. 42 shows an overview of the target route creation process. This figure shows a target creation process using an adjustment area (described later). The graph D01 is a graph in which the current time is the origin (D03) of the time axis, the horizontal axis is time, and the vertical axis is the position of the floor of the building. Although a graph is not drawn in the figure, for example, a target route as shown in FIG. 43 is drawn therein.

  FIG. 43 is an example of target route shapes before and after adjustment according to an embodiment of the present invention. As the target route, a route that is in a temporally equidistant state at a predetermined time in the future is created, and this predetermined time corresponds to the adjustment reference time axis D04. The target route is an area between the current time axis D03 and the adjustment reference time axis D04 (referred to as an adjustment area), and is represented by a route representing a transient state until the time interval becomes equal. It is expressed as a route that is in a time equidistant state from the reference time axis D04 onward.

  Such a goal creation process includes the following four processes.

1) Draw a current predicted route (ST701 in FIG. 42),
2) Calculate the current phase time value of each car on the adjustment reference time axis (ST702),
3) Based on the current phase time value, calculate the adjustment amount of each car so as to be equally spaced in time (ST703),
4) The grid of the predicted route in the adjustment area is adjusted according to the adjustment amount, and this becomes the target route (ST704).

  FIG. 44 is a diagram showing an example of the configuration of the target route creation unit. The configuration of the target route creation unit shown in the figure is mainly composed of the following four elements.

1) Target route update determination unit 103A,
2) Current phase time value calculation unit 103B,
3) Adjustment amount calculation unit 103C for the phase time value of each car, and 4) Route creation unit 103D after adjustment.

  First, as an explanation of the control image, the operation of the above four elements will be described. The target route update unit 103A determines whether or not to update the current target route. When it is determined that the target route is updated, the current phase time value calculation unit 103B in the next stage uses the index of the interval state of each car route for the predicted route of each elevator car at that time as an index called a phase time value. evaluate. Here, the reason for using the concept of “phase” is, for example, when a three-phase AC waveform of a sine wave is considered in electric circuit theory, the state in which the waveform of each phase is equalized is the phase of each phase. This is based on the fact that the phase is equal to 2π / 3 (rad). In other words, if the route of each car is regarded as a waveform, and the “index like phase” is used for the waveform, the interval state for each route can be easily evaluated. This 'index like phase' corresponds to an index called a phase time value used in this embodiment. The phase time value will be described later. After the current phase time value calculation unit 103B calculates the phase time value at that time, the phase time value adjustment amount of each car for equalizing the phase time value is set as the phase time value adjustment amount of each car. Calculation is performed in the calculation unit 103C. Based on the adjustment amount calculated above, the adjusted route creation unit 103D adjusts the time phase value of the predicted route 103B of each original car. The route obtained as a result of the adjustment becomes the target route for each car.

  The operation for the schematic control configuration described above will be described using the operation image of FIG. FIG. 43 is an operation image diagram of the target route creation process executed by the target route creation unit shown in FIG. Here, first, an operation image of control based on the general control content described above will be described. First, the diagram of FIG. 43A (target route shape before adjustment) corresponds to the predicted route of each car at the present time as a base for creating the target route. Here, three elevator group management systems are considered. In FIG. 43 (A), the first car C010, the second car C020, and the third car C030 are descending on the 8th floor and descending on the 3rd floor on the current axis C050. It is in the middle state. The predicted routes (predicted trajectories) of the three cars after the present time are a trajectory C011 where the first car is a solid line C011, a trajectory C021 where the first car is a one-dot chain line, and a trajectory C031 where the third car is a dotted line. The predicted route creation method will be described in detail in the description of the predicted route creation unit. These trajectories are clearly close to each other, and it can be seen that they are close to the dango driving state. Returning to the control configuration of the target route creation unit in FIG. 44, first, when the target route update determination unit 103A determines that the target route is updated, the current phase time calculation unit 103B performs each of the operations shown in FIG. With respect to the predicted routes C011, C021, and C031 of the car, these are regarded as a kind of waveform, and the respective phase time values are calculated. This phase time value is calculated at the intersection point where the predicted route of each car crosses the adjustment reference time axis C040 in the graph of FIG. Next, based on this phase time value, an adjustment amount for each predicted route to be in an equally spaced state is calculated by the phase amount value adjustment amount calculation unit 103C of each car. This adjustment amount is represented as three black dots on the adjustment reference time axis C040 in FIG. For example, in the case of the first car, the point C01A reflects the adjustment amount, and the predicted route C011 of the first car is adjusted by the next process so as to pass this point C01A. Similarly, the predicted route C021 of Unit 2 is adjusted by the following processing so as to pass through the point C02A, and the predicted route C032 of Unit 3 is adjusted through the point C03A. The adjusted route creation unit 103D shown in FIG. 44 performs this adjustment process. Here, the predicted route is adjusted based on the adjustment amount, and a new target route is created. The result is the locus shown in FIG. FIG. 43B is a diagram showing a new target route created based on the predicted route shown in FIG. For each of the three cars C010, C020, and C030, the target route of Unit 1 C010 is a solid locus C011N, the target route of Unit 2 C020 is a one-dot chain line C021N, and the target route of Unit 3 C030 is a dotted line The locus is C031N. The feature of the trajectory of the target route is that the route of each car is drawn so as to lead to a time-spaced state as shown in FIG. Specifically, in FIG. 5B, the target routes of the three cars are in an equally spaced state with respect to time at a time after the adjustment reference axis C040. In the time between the axis C050 representing the current time and the adjustment reference time axis C040 (the time area indicated as the adjustment area in FIG. 42), a trajectory is drawn so as to lead each car to such a time-equally spaced state. ing. Based on the predicted route shown in FIG. 43A, each route is adjusted so that each route passes through the points obtained from the adjustment amount, that is, points C01A, C02A, and C03A of the adjustment reference axis. Thereby, such a target route as shown in FIG. 43B can be created.

  Hereinafter, detailed elements in the target route creation unit illustrated in FIG. 44 will be described. The current phase time value calculation unit 103B includes an initial state route creation unit 103B1, an adjustment reference axis setting unit 103B2, a phase time value calculation unit 103B3 of each car on the adjustment reference axis, and a sorting unit 103B4 in order of phase time values. The initial state route creation unit 103B1 creates a predicted route for each car at that time and sets this as the initial route. This route in the initial state corresponds to the target route shape before adjustment shown in FIG. The adjustment reference axis setting unit 103B2 sets the adjustment reference time axis. The phase time portion calculator 103B3 of each car on the adjustment reference time axis calculates the phase time value of each car on the adjustment reference time axis. Specifically, a phase time value is calculated for the intersection of the predicted route of each car and the adjustment reference time axis (corresponding to the predicted position of each car on the adjustment reference time axis). After the phase time value of each car is calculated by the phase time value calculation unit 103B3 of each car on the adjustment reference time axis, the phase time value for each car is sorted in the order of the phase time value by the sorting unit 103B4 in order of the phase time value. To do. Hereinafter, this order is referred to as a phase order. For example, taking the state of three cars in the target route shape before adjustment (corresponding to the predicted route) in FIG. 43A as an example, the intersection of the adjustment reference axis C040 and the predicted routes C011, C021, and C031 of each car Therefore, the order of the phase time values of each car is the phase order of No. 3, No. 2, No. 1 from the smallest. The sorting unit 103B4 in order of phase time values obtains such a phase order using a sorting algorithm such as a direct selection method or bubble sort.

  Based on the calculated phase time value of each car and its phase order, the adjustment amount calculation unit 103C for the phase time value of each car calculates the interval of each car as the phase time value and makes this value an equal interval. Is compared with a reference value to become, and an adjustment amount of the phase time value of each car represented as the difference is calculated. The idea here is to find the interval (evaluated by the phase time value) of each car from the predicted route, compare this with the reference value for equal intervals, and use that difference as the adjustment amount to be adjusted from now on. become. Hereinafter, the processing content of the adjustment amount calculation unit 103C for the phase time value of each car will be described with reference to FIG. 43 (A) as an example. As described above, in FIG. 43A, the phase order of the phase time values on the adjustment reference time axis C040 of the predicted routes C011, C021, and C031 of each car is the order of Unit 3, Unit 2, and Unit 1. Yes. Assuming that the one-round time of the predicted route is T (all three cars have the same one-round time), the phase time value tp (k) of the k-th unit is tp (3) = 0.09T for the third unit and tp ( 2) = 0.17T, Unit 1 becomes tp (1) = 0.77T. When the distance between each car is calculated in order of phase, the distance between Unit 2 and Unit 3 is tp (2) -tp (3) = 0.08T, and the distance between Unit 1 and Unit 2 is tp (1) -tp (2) = The interval between 0.6T, No. 3 and No. 1 is tp (3) -tp (1) + T = 0.32T. Thus, by quantifying the interval of each car based on the phase time value, the interval of each car can be quantitatively evaluated. For example, it can be seen from the above results that the distance between Unit 2 and Unit 3 is very close. In the phase time value, one round time is set as T. Therefore, in the case of management of N groups, the interval of each car in the target time equal interval state can be expressed by T / N. In the example of FIG. 43 (A), the target car interval is T / 3 = 0.33T due to the group of three cars. The difference between the target interval and the current interval of each car is the interval to be adjusted. For example, between the second and third units, + 0.25T (= 0.33T-0.08T) is the interval value to be adjusted. Also, -0.27T (= 0.33T-0.6T) should be adjusted between Units 1 and 2, and + 0.01T (= 0.33T-0.32T) should be adjusted between Units 3 and 1. The interval value. In the above, a positive sign represents an increase in the interval, and the current interval needs to be increased with respect to the target. On the other hand, the negative sign represents a decrease in the interval, and the current interval needs to be reduced with respect to the target. Based on the interval value to be adjusted, the adjustment amount of the phase time value for each car is calculated. This can be determined by the following algorithm. For example, as group management of three units, it is assumed that the A machine, the B machine, and the C machine are arranged in order of phase. In order to generalize, the name of the machine is written here in alphabet. From the above, 0 ≦ tp (A) ≦ tp (B) ≦ tp (C) <T holds. Here, the adjustment amount of the phase time value for each car is represented by Δtp (k) (k represents that the car is the k-th car). First, in order for the interval between the cars after adjustment to satisfy the target interval T / 3, the following equations must be satisfied.

(Tp (B) + Δtp (B)) − (tp (A) + Δtp (A)) = T / 3 (19)
(Tp (C) + Δtp (C)) − (tp (B) + Δtp (B)) = T / 3 (20)
(Tp (A) + Δtp (A)) − (tp (C) + Δtp (C)) + T = T / 3 (21)
For example, in the equation (19), the adjusted phase time value is expressed as tp (B) + Δtp (B) with respect to the current phase time value tp (B). Therefore, the equation (19) represents that the difference between the adjusted phase time value of the machine B and the adjusted phase time value of the machine A, that is, the interval satisfies T / 3. Here, since the above three equations are not independent of each other, Δtp (A), Δtp (B), and Δtp (C) cannot be solved with only these three equations. Therefore, as another condition, a condition is added in which the center of gravity on the arrangement viewed from the phase time value of each current car matches the center of gravity on the arrangement viewed from the phase time value of each adjusted car. This condition is as follows.

(Tp (A) + tp (B) + tp (C)) / 3 = {(tp (A) + Δtp (A)) + (tp (B) + Δtp (B)) + (tp (C) + Δtp (C))} / 3 (22)
When formula (22) is arranged, formula (23) is obtained.

Δtp (A) + Δtp (B) + Δtp (C) = 0 (23)
Solving the equations (19), (20), (21), and (23) with respect to Δtp (A), Δtp (B), and Δtp (C), the following equation is obtained.

Δtp (A) = (− 2/3) tp (A) + (1/3) tp (B) + (1/3) tp (C) + (− 1/3) T (24) )
Δtp (B) = (1/3) tp (A) + (− 2/3) tp (B) + (1/3) tp (C) (25)
Δtp (C) = (1/3) tp (A) + (1/3) tp (B) + (− 2/3) tp (C) + (1/3) T (26)
In summary, with respect to the three cars A to C whose phase time values are 0 ≦ tp (A) ≦ tp (B) ≦ tp (C) <T, the three cars are arranged at equal intervals after adjustment. The adjustment amounts Δtp (A), Δtp (B), Δtp (C) satisfying the condition that the center of gravity of the change does not change can be obtained. These adjustment amounts Δtp (A), Δtp (B), and Δtp (C) can be obtained by equations (24), (25), and (26), respectively. For example, taking FIG. 43 (A) as an example, Units A, B, and C become Units 3, 2, and 1, respectively. Therefore, tp (A) = tp (3) = 0.09T, tp (B) = tp (2) = 0.17T, and tp (C) = tp (1) = 0.77T. Therefore, the adjustment amount for each car is Δtp (A) = Δtp (3) = − 0.081T, Δtp (B) = Δtp (2) = 0.177T from the equations (24) to (26). , Δtp (C) = − 0.096T. As confirmation, each adjusted phase time value is obtained. tp (A) + Δtp (A) = tp (3) + Δtp (3) = 0.010T, tp (B) + Δtp (B) = tp (2) + Δtp (2) = 0. 343T, tp (C) + Δtp (C) = tp (1) + Δtp (1) = 0.777T. That is, the intervals of the respective cars are all 0.33T, and the conditions of equal intervals are satisfied.

  Next, details of processing for creating an adjusted route by the adjusted route creation unit 103D using the adjustment amount obtained by the adjustment amount calculation unit 103C for the phase time value of each car will be described. In the adjusted route creation unit, first, the grid adjustment amount calculation unit 103D1 on the route of each car calculates the adjustment amount of the grid on the target route (corresponding to the predicted route) before the adjustment of each car. The grid is defined as the direction reversal point of the target route in the adjustment area. By adjusting the position of the grid in the horizontal direction, the phase time value of the target route can be adjusted. The adjustment amount of each grid is determined by a method of assigning the adjustment amount of the car as a total amount and sequentially assigning a value exceeding the limiter value set for the grid from the grid closest to the current time. Here, the limiter value of the adjustment amount of each grid is set by the grid limiter value setting unit 103D2. In the adjusted grid position calculation unit 103D3, the adjusted grid position gp_N (k, i) from the adjustment amount Δgtp (k, i) for each grid and the position gp (k, i) of the grid before adjustment. Calculate For example, when k = 2 and the number of grids is 3 (i = 1, 2, 3), the calculation formula of each grid is as follows.

gp_N (k = 2, i = 1) = gp (k = 2, i = 1) + Δgtp (k = 2, i = 1) (27)
gp_N (k = 2, i = 2) = gp (k = 2, i = 2) + Δgtp (k = 2, i = 1)
+ Δgtp (k = 2, i = 2) (28)
gp_N (k = 2, i = 3) = gp (k = 2, i = 3) + Δgtp (k = 2, i = 1) + Δgtp (k = 2, i = 2) + Δgtp (k = 2, i = 3) ……………… (29)
Since the adjustment amount of the grid is carried over to the succeeding grid, the position of the final grid is adjusted by the total amount of the phase time value adjustment amount for the car. As described above, a new target route can be created by connecting the adjusted grid positions to each other. The target route data calculation unit 103D4 calculates and updates this new target route data.

  The newly updated target route (adjusted target route) passes through the adjusted target point set as the adjustment amount of the phase time value. Since the route of each car is adjusted so as to pass through the adjusted target point, the result of combining the three cars is as shown in FIG. 43 (B), and the three targets after the adjustment reference time axis C040. It can be seen that the routes C011N, C021N, and C031N are in a time equidistant state. Naturally, each of the routes C011N, C021N, and C031N passes through the adjusted target points C01A, C02A, and C03A. It can also be seen that the target route in the adjustment area adjusted by the grid plays a role of a transitional guide for achieving a time equidistant state after the adjustment reference time axis. The above is the details of the target route creation process.

  FIG. 45 shows a second embodiment different from FIG. 2 of the control block of the entire elevator group management system according to the present invention. 45, the same elements as those in FIG. 2 are assigned the same reference numerals, and description thereof is omitted here. The difference from FIG. 1 is that the weighting factor is directly determined by the number of hall calls. Specifically, the hall call occurrence number calculation unit 10 obtains the hall call occurrence number, and based on this, the weight coefficient calculation unit 5 directly obtains the weight coefficient from the hall call occurrence number.

  FIG. 46 shows an example of a function for obtaining a weighting factor used in the weighting factor calculation unit 5. In FIG. 46, the horizontal axis represents the number of hall calls generated per round, and the vertical axis represents the weighting coefficient. Here, the number of hall calls per elevator lap is the number of halls generated during one lap (for example, from the lowermost floor in the upward direction to the lowermost floor in the downward direction) for each group-managed elevator. Represents the average number of calls.

  In FIG. 46, the function for determining the weighting coefficient is represented by a curve F04. The features of this function include the following four points.

1) An appropriate weighting factor value can be obtained immediately according to the number of hall calls.
2) The continuous change in the number of hall calls per lap, which is an input variable, causes the weighting coefficient, which is an output, to be continuously determined.
3) The input variable is a scalar value (one variable), and 4) The input variable can take a continuous value as a real number.

  From the graph of FIG. 46, for example, when the number of hall calls per round is NA, the value of the weighting coefficient is WT = F (NA). Regardless of how the NA changes due to changes in traffic demand, the WT can be obtained immediately. This is a major feature. In each function, the weighting coefficient value is always zero for a value equal to or less than the value on the horizontal axis that intersects the zero value on the vertical axis.

  The reason why an appropriate weighting factor can be obtained by using the number of hall calls generated per elevator lap as input is that the importance of the car interval evaluation value is stronger than the number of hall calls that will be generated in the future. Because of the relationship. For example, as the number of hall calls to be generated in the future increases, it is better to make the intervals equally as much as possible, and it is better to make the car interval evaluation value work more strongly. Here, the number of hall calls to be generated in the future is considered to have a strong correlation with the number of hall calls generated per round at that time or at a later time. Therefore, there is a relationship between the number of hall calls generated per lap and an appropriate weight coefficient value. By expressing this as a function as shown in FIG. 10, the number of hall calls generated per elevator lap is appropriate. Can be determined.

  In the graph of FIG. 46, the horizontal axis shows an example of the number of hall calls generated per lap. However, the present invention is not limited to this, and a value based on the number of hall calls generated, for example, the number of hall calls generated in a predetermined time, etc. But you can. Furthermore, it is not limited to the number of hall calls, but may be a scalar value index related to traffic demand. For example, it may be a value based on the total number of users or the number of hall calls and the number of car calls.

  As described above, the configuration of FIG. 45 has an effect that an appropriate weighting factor can be quickly set with a simpler configuration than the configuration of FIG. As a result, stable performance can be maintained even if a microcomputer, an arithmetic processor, or the like, which is inexpensive but has a low calculation processing amount, is used.

  FIG. 47 shows a third embodiment of the elevator group management system according to the present invention, which is different from those shown in FIGS. 47, the same elements as those in FIG. 2 are denoted by the same reference numerals, and description thereof is omitted. A difference from FIG. 2 resides in that an allocation evaluation information display processing unit H01 is newly provided. This allocation evaluation information display processing unit H01 records internal information such as an evaluation value for determining the allocation and intermediate information for the hall call allocation processing, and displays the contents as necessary. It is characterized by being transferred to the outside or recorded on a recording medium. The purpose is to analyze the cause of the allocation for each allocation process. For example, when a long wait occurs, in what situation the long wait occurred. To elucidate the cause. In particular, since the elevator group management system according to the present invention can visually indicate the intention of the group management control side based on the predicted trajectory and the target route, such display processing means is effective.

  Hereinafter, the configuration of the allocation evaluation information display processing unit H01 will be described. The allocation evaluation information display processing unit H01 includes a determination unit H02, a recording unit H03, a drawing processing unit H04, a recording medium H05, and an output unit H06. The determination unit H02 is a part that determines whether to record evaluation value information, and has a mechanism for recording allocation evaluation information in the recording unit H03 when a recording permission signal is output from the determination unit H02. . This determination includes, for example, a specific time zone, a specific traffic flow, occurrence of an average waiting time exceeding a predetermined value, generation of an average hall call duration exceeding a predetermined value, and the like. The recording unit H03 records various data related to the allocation evaluation calculated by the group management control unit 1. For example, the waiting time evaluation value calculated by the waiting time evaluation value 3, the output of the car interval evaluation value calculation unit 4, the weighting factor set by the weighting factor setting unit 8, the output of the comprehensive evaluation value calculation unit 6, and the assigned elevator determination unit The allocation decision machine name which is the output of 7 is recorded. A detailed function of the car interval evaluation value calculation unit 4 is shown in FIG. 48 and will be described in detail later. In the drawing processing unit H04, drawing data creation processing for drawing information on the allocation recorded in the recording unit H03 on the screen is performed. In the output unit H06, drawing data is displayed on the screen. Data recorded by the recording unit H03 is stored in the recording medium H05. As the recording medium H05, a floppy disk, a memory card, a USB memory, a hard disk, or the like is used. The drawing data created by the drawing processing unit H04 is transferred to the outside via the communication network H07, and the drawing data can be remotely displayed on the screen and stored.

  FIG. 48 is a detailed functional block diagram of the interval evaluation value calculation unit 4, showing the flow of data information recorded by the recording unit H03. The configuration of FIG. 48 is based on FIG. 12, and the same elements as those of FIG. 12 are denoted by the same reference numerals, and description thereof is omitted here. The data sent to the recording unit H03 first includes an interval evaluation time tref calculated by the interval evaluation time setting unit 405, which is output from the interval evaluation time data output unit 4Z2. Furthermore, there is predictor trajectory data created by the predicted trajectory calculation unit 401 (specifically, a predicted trajectory table in FIG. 23), which is output from the predicted trajectory data output unit 4Z3. Further, since the predicted trajectory data has a large data amount as it is, for example, if the data amount is compressed by converting into a simplified predicted trajectory as shown in FIG. 33, the data amount can be suppressed to an appropriate amount. It is the predicted trajectory data conversion unit 4Z1 that executes such conversion processing. In addition, data such as predicted car interval data (output from the output unit 4Z5) calculated by the predicted car interval calculation unit 402 and interval evaluation value data (output from the output unit 4Z6) calculated by the interval evaluation value calculation unit 403 The data is output to the recording unit H03 and recorded.

  FIG. 49 shows an example of screen output data created by the drawing processing unit H04 of FIG. 47 and displayed by the output unit H06. 49 can be roughly divided into the following four groups.

1) a group L001 for displaying and outputting information on the predicted trajectory,
2) Group L002 for displaying and outputting information related to hall calls to be allocated;
3) a group L003 for displaying and outputting evaluation value information; and 4) a group L004 for displaying detailed information of interval evaluation values.

  Details of each will be described below.

  First, the group L001 that displays and outputs information on the predicted trajectory is characterized in that the following data is collectively displayed on a graph L006 that takes time and floor position as two axes. The data displayed on the graph includes the predicted trajectory of each car (predicted trajectory L010 of No. 1 and the predicted trajectory L011 of No. 2), the interval evaluation time L012, and the car position of each car at the interval evaluation time (the car of No. 1). There is a position L015, a car position L016 of the second car. With such a display, it is possible to show in detail how the predicted trajectory of each unit is, whether the car interval of each unit is being evaluated, and what the positional relationship of the cars at that time is. it can. As a result, it is possible to visually and easily confirm how the car interval evaluation has been performed together with the car interval evaluation value described later. In particular, in the present invention, since the interval evaluation time is adaptively adjusted according to the call generation status and traffic flow status, it is important to clearly indicate which time is set as the interval evaluation time at the time of the allocation. In addition to this, the upper left diagram L005 (the first car L007, the second car L008, and the hall call L009) displaying the car status and call status at the time of allocation is displayed. These have the effect of supporting information on the predicted trajectory. For example, it is possible to understand the basis of the predicted trajectory drawn from any call generation state.

  Next, in the group L002 that displays and outputs information related to hall calls to be allocated, time information L017, generated hall call floor information L018, and generated hall call direction information L019 are displayed. With these pieces of information, the details of the hall call to be assigned can be confirmed.

  In the group L003 that displays and outputs the evaluation value information, the assigned machine name L020, the waiting time evaluation value L021, the car interval evaluation value L022, the weight coefficient value L023, and the overall evaluation value L024 are displayed. All of these pieces of information are important information that is the core of assignment determination, and by displaying them side by side, it is possible to know what factors caused the assignment to be determined. That is, the factor of the allocation decision can be roughly estimated by the information displayed in the group L003 that displays and outputs the evaluation value information. For example, the waiting time evaluation value was strong in order to avoid the occurrence of long waiting time, the car interval evaluation value was avoided from worsening, and at that time, the weighting factor value was large and the car interval evaluation value was emphasized For example, it can be a support for the allocation determining factor search. In addition, by displaying these information together with the information of the group L001 that displays and outputs information related to the predicted trajectory, it is difficult to grasp only by numbers, and the contents of the car interval evaluation value can be understood more intuitively. become.

  In the group L004 that displays and outputs the detailed information on the car interval evaluation value, the detailed information inside the processing calculated when calculating the car interval evaluation value is displayed. Specifically, information on the interval evaluation time value L025, the predicted car position L026 of each car at the interval evaluation time, and the predicted car interval L027 is displayed. Since the car interval evaluation value is information that aggregates all, the basis of the evaluation value may be difficult to understand quantitatively. In that case, more detailed analysis is possible from the above detailed information.

  The example of the screen output data generated by the drawing processing unit H04 of FIG. 47 and displayed by the output unit H06 has been described with reference to FIG. The main features of this screen output data are the prediction trajectory of each car, the interval evaluation time when predicting the car position on the prediction trajectory, and the predicted car position and direction of each car at the car interval evaluation time in one graph. It is in the point that is displayed all over. Further, as evaluation value information, the waiting time evaluation value, the car interval evaluation value, the weight coefficient value, and the comprehensive evaluation value of the allocation evaluation target car are displayed in parallel. As a result of such display, the predicted trajectory of each car for the target assignment, the time point when the car interval on the predicted trajectory is predicted, and the car position at that time can be visually specified. Therefore, it is possible to display in an easy-to-understand manner what situation (predicted situation) the car interval is evaluated. For example, for a user who has doubts about the basis of a certain allocation, what kind of prediction can be made by visually showing the predicted trajectory, the prediction time (interval evaluation time), and the predicted car position It can be easily shown below whether such an assignment has been made. Further, since the breakdown of the evaluation values is shown, it is possible to easily know whether the allocation has been determined due to the waiting time evaluation value or the car interval evaluation value. Considering that the waiting time evaluation value is an evaluation value of an actual call that has already occurred and the cage interval evaluation value is an evaluation value for a future call that has not yet occurred, either the actual call or the future call may cause an allocation. It can be said that this information represents what has been done. That is, the basic intention for the assignment can be grasped.

  FIG. 50 shows an example of screen output data different from FIG. 50, the same elements as those in FIG. 49 are denoted by the same reference numerals, and description thereof is omitted here. 50 differs from FIG. 49 in that the predicted trajectory before hall call assignment shown by a solid line (predicted trajectory L010 of the first car) and the predicted trajectory after hall call assignment shown by a broken line (predicted trajectory L101 of the same first car). It is in the point that and are displayed. In the predicted trajectory L101 after the hall call assignment, a stop due to the hall call assignment appears in the locus. That is, the locus of the horizontal portion indicated by reference numeral L102. The hall call L009 at this time is generated on the 13th floor / upward.

  As shown in FIG. 50, by displaying the predicted trajectories before and after hall call allocation in an overlapping manner, it is possible to more specifically display how the car interval changes depending on the allocation. In the example of FIG. 50, the prediction trajectory L011 of the second car indicated by the alternate long and short dash line spreads to the predicted trajectory L101 after the assignment of the first car. In particular, since the car intervals before and after the assignment can be compared, the effect of the assignment can be displayed quantitatively and easily. This is effective for analyzing allocation factors and understanding the basis of allocation.

  FIG. 51 shows a third example of screen output data different from FIG. 51, the same elements as those in FIG. 49 are denoted by the same reference numerals, and description thereof is omitted here. 51 differs from FIG. 49 in that the actual operation trajectory of each unit before the allocation is displayed side by side with the predicted trajectory of each unit. Specifically, in FIG. 51, the actual operation trajectory of the first car is represented by a trajectory L201 indicated by a solid line, and the actual operation trajectory of the second car is represented by a one-dot chain line L202. The allocation is performed at the axis L203, and the left side of the axis is the actual operation locus and the right side is the predicted locus. In the actual operation trajectory, the car stops are all actually stopped, and the reference L204 shown in the middle of the actual operation trajectory L201 of Unit 1 represents the result that Unit 1 stopped on the third floor in the upward direction. ing. That is, there is no stochastic stop. This is the apparent difference between the actual trajectory and the predicted trajectory.

  As shown in FIG. 51, for each unit, the predicted trajectory and the actual operation trajectory before the allocation execution time are displayed together (connected), so that the transition of each car interval before the allocation execution and the future You can compare the predictions of As a result, for example, it was a dango driving state until then, but it seems that it is acting to avoid dango driving by the predicted trajectory, and conversely, until that time it was in an equally spaced state in time, It can also be seen from the predicted trajectory that it is approaching dango driving in the future. Therefore, in order to avoid this, it is possible to confirm a state in which assignment that is intentionally separated is carried out preventively.

  From the above, as shown in FIG. 51, the predicted trajectory and the actual operation trajectory are displayed together (connected), so that the transition of the car interval viewed on the time axis can be shown in an easy-to-understand manner. In addition, it is possible to easily understand the situation of the change in the car interval and the past history, and it is possible to more easily analyze the factor of allocation.

FIG. 52 shows an example of the processing flow of the predicted trajectory display method. In this example, a predicted trajectory is displayed for each temporary assigned car. Hereinafter, the flow of processing will be specifically described.
First, necessary data is read (L301), and then a display format is selected (L302). Examples of selection of the display format include selection of display items and display time range setting. Next, the hall call to be displayed is selected (L303). Next, a variable K representing a temporary allocation number is first set to K = 1 (L304). Thereafter, the display data (A) of the predicted trajectory of No. K when the No. K is temporarily assigned to the hall call selected above is created (L305). Further, display data (B) of the predicted trajectory of each unit other than unit K is created (L306). Next, display data (C) of interval evaluation time is created (L307). After the above processing is completed, the screen combining the above (A), (B), and (C) is stored as display or data as the image data of the predicted trajectory of each unit when the temporary allocation is set to unit K. (L308).

  Next, 1 is added to K (L309), and the above process is repeated for the next unit, and is repeated until it is performed for all units (L310). By executing such processing, the display screen data described with reference to FIGS. 49 to 51 can be created.

  FIG. 53 shows an example of a processing flow of a predicted trajectory display method different from that in FIG. In this example, the predicted trajectory for the unit to which the call is assigned is displayed. Hereinafter, the flow of processing will be specifically described.

  First, necessary data is read (L401), and then a display format is selected (L402). Examples of selection of the display format include selection of display items and display time range setting. Next, the hall call H to be displayed is searched from a database in which hall calls generated in the past are recorded (L403). Next, for the retrieved hall call H, the machine K that has actually been assigned is searched from the past database (L404). Thereafter, display data (A) of the predicted trajectory of the K machine when the hall call selected above is assigned to the K car is created (L405). Furthermore, the display data (B) of the predicted trajectory of each unit other than the unit K is created (L406). Next, display data (C) of the interval evaluation time is created (L407). After the above processing is completed, a screen that combines the above (A), (B), and (C) is displayed or displayed as the image data of the predicted trajectory of the assigned unit (K) and the other units for the hall call H. (L408). By executing such processing, display screen data as shown in FIGS. 49 to 51 can be created.

The control image figure of the elevator group management system by this invention. The control functional block diagram of the elevator group management system by one Example of this invention. The control processing flowchart of the elevator group management system by one Example of this invention. The graph explaining the weighting coefficient setting method by one Example of this invention. The image figure of the optimal solution search method of the weighting coefficient by one Example of this invention. An example of traffic flow in a 6th floor building and an explanatory view of the concept of weighting. An explanatory view of an example of a weighting factor setting method when an input is handled by a traffic flow. The comparison figure of one Example of this invention and the conventional weighting coefficient setting method. The processing flowchart of the weighting coefficient setting method by one Example of this invention. The processing flowchart of the optimal solution search method of the weighting coefficient by one Example of this invention. FIG. 3 is an example diagram showing data in an input information storage unit 2 shown in FIG. 2. The detailed block diagram of the cage | basket | car interval evaluation value calculating part 4 shown in FIG. The processing flowchart which sets the space | interval evaluation time tref. Explanatory drawing with respect to the setting of the space | interval estimated time by one Example of this invention. The functional block diagram of the 2nd Example of the cage space | interval evaluation value calculating part replaced with FIG. FIG. 13 is a functional block diagram of a third embodiment of a car interval evaluation value calculation unit instead of FIG. 12. FIG. 13 is a functional block diagram of a fourth embodiment of a car interval evaluation value calculation unit instead of FIG. 12. The whole process flow figure of the prediction locus | trajectory creation method by one Example of this invention. FIG. 5 is a process flow diagram for creating an estimated arrival time table for multiple laps according to an embodiment of the present invention. FIG. 20 is a process flow diagram of an estimated arrival time table calculation routine (FB04 in FIG. 19). Explanatory drawing of the stop time to each floor, a stop probability, and a stop time expected value table. The specific example figure of the arrival prediction time table | surface of multiple rounds using the numerical value shown in FIG. An example figure of the prediction locus table finally created by one example of the present invention. The prediction position calculation image figure used as the basis of the prediction locus | trajectory by one Example of this invention. The processing flowchart (the 1) of the prediction locus | trajectory table preparation by one Example of this invention. FIG. 9 is a process flow diagram (part 2) for creating a predicted trajectory table according to an embodiment of the present invention. The flowchart of the prediction locus | trajectory creation process at the time of no direction. An example figure of a prediction locus created with a prediction locus creation method by one example of the present invention. Explanatory drawing of the predicted trajectory that waits with emphasis during off-peak hours. The 2nd example of the prediction locus created with the prediction locus creation method by one example of the present invention. The 3rd example of the prediction locus created with the prediction locus creation method by one example of the present invention. The 4th example of the prediction locus created with the prediction locus creation method by one example of the present invention. The 5th example of the prediction locus created with the prediction locus creation method by one example of the present invention. The 6th example of the prediction locus created with the prediction locus creation method by one example of the present invention. The 7th example of the prediction locus created with the prediction locus creation method by one example of the present invention. The processing flow figure of prediction interval value calculation by one example of the present invention. Explanatory drawing which calculates a prediction interval value from a prediction route by one Example of this invention. Explanatory drawing of the process which calculates | requires a prediction interval by one Example of this invention. The comparative example figure of the service locus | trajectory of the cage | basket | car before and after one Example adoption of this invention. The image figure of an example of target route control by one example of the present invention. Explanatory drawing of a mode that a hall call is allocated to an elevator according to a target route. The outline explanatory view of the target route creation process by one example of the present invention. An example figure of the target route shape before and behind adjustment by one example of the present invention. The functional block diagram of an example of the target route preparation part by one Example of this invention. The elevator group management system control functional block diagram of 2nd Example of this invention. 46 is a graph illustrating an example of a weighting factor calculation function used in the weighting factor calculation unit 5 of FIG. The elevator group management system control functional block diagram of 3rd Example of this invention. The detailed functional block diagram of the space | interval evaluation value calculating part by one Example of this invention. An example figure of screen output data displayed by one example of the present invention. The 2nd example figure of the screen output data displayed by the other Example of this invention. The 3rd example figure of the screen output data displayed by the other Example of this invention. The processing flowchart of the prediction locus | trajectory display method by one Example of this invention. The processing flow figure of the prediction locus display method by other examples of the present invention.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Group management control part, 2 ... Input information storage part, 3 ... Wait time evaluation value calculation part, 4 ... Car space | interval evaluation value calculation part, 5 ... Weight coefficient calculation part, 6 ... Comprehensive evaluation value calculation part, 7 ... Assignment Elevator determination unit, 8 ... weight coefficient setting unit, 10 ... hall call occurrence number calculation unit, 11 ... weight coefficient range calculation unit, 12 ... weight coefficient initial value calculation unit, 20 ... traffic flow determination unit, 21 ... weight coefficient optimum solution Search unit, 22 ... simulation unit, 31A to 31C ... AC control device, 31A to 31C ... A to C elevator car, 41A, 41B ... Hall call button, H01 ... assignment evaluation information display processing unit, H02 ... determination , H03 ... recording unit, H04 ... drawing processing unit, H05 ... recording medium, H06 ... output unit (display unit).

Claims (4)

In an elevator group management system that manages multiple elevators that service multiple floors,
Determine whether each elevator is directional with a hall call or car call or no direction with no hall call or car call,
For a directional of the elevator hall calls and to floor the car call has not occurred, a stop time for a hall call calls and the car has occurred, hall call and car not occurred corresponding to traffic flow wherein based on the elevator stop probability by call, as well as create a predictive trajectory indicating a predicted floor position and direction of each of the elevator at each time within a predetermined future time from the present time,
For non-direction of the elevator, to create the predicted trajectory to continue waiting state,
The predetermined period is a period in which the elevator circulates a plurality of times,
2. The elevator group management system according to claim 2, wherein the second and subsequent rounds of the predicted trajectory are predicted trajectories based on the stop probability of the elevator caused by a hall call and a car call that have not occurred . The elevator group management system according to claim 1, wherein an evaluation value related to a positional relationship between the predicted trajectories of each elevator is calculated, and an elevator is assigned to the generated hall call based on the evaluation value. The elevator group management system according to claim 2 , wherein the positional relationship between the cars is a time interval and / or a spatial interval between the cars. In any one of claims 1 to 3, the elevator group control system characterized by comprising a predicted trajectory display means for displaying the predicted trajectory of each elevator.

Images