JPH01231778A - Device and method for controlling elevator and device and method for inputting sensibility of control device - Google Patents

Abstract

PURPOSE:To easily execute the control method desired by a customer by converting a sensibility target input into a control target value using a variable function and replacing same to a control method by means of a previously set rule, etc., considering the weighting between each items of the control target. CONSTITUTION:A sensibility target is inputted and set from the input terminal 2-1 of a sensibility target setting means 2 based on the conditions 2-2 of a building specification, an elevator specification, and using information, using Japanese and a laser chart, etc., which enable easy setting even by a nonexpert. This set sensibility target is converted into a control target value by the target converting means 3 of a group controlling control device body 1 using a variable function obtained by a questionnaire, etc., and for those which are mutually related for each item, a control method is obtained based on a previously set rule, etc., considering weighting, to execute a group controlling control by a control executing means 4. Thereby, a desired control method can be executed without need for troubling an expert.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an elevator control device and a control method, as well as a sensory input device and method for the control device, and is particularly applicable to elevators subject to group management control. The present invention relates to a device and method suitable for realizing control corresponding to human sensibilities.

[Prior Art] Conventionally, in an elevator group management control device, the elevator
1 with the aim of improving operating efficiency and improving passenger service.
The hall calls that occur are monitored online, and the service status of all hall calls is taken into account to determine the optimal elevator.
A method has been adopted to reduce the average waiting time by allocating calls to Recently, when selecting an elevator from among a plurality of elevators to which a hall call is to be assigned, a variable parameter is added to the evaluation function for evaluating each elevator, and the variable parameter is added to the evaluation function for evaluating each elevator. Parameter values are changed and the obtained results are used to learn parameter values that satisfy preset target values, and call assignment control is executed using the parameters according to the operating status of the elevator. The method of
It is proposed in No.

In this system, two control target values, waiting time and energy saving, can be set using switches and level commands from the building management system, and the system evaluates stop calls in the vicinity of generated calls, and determines whether the elevator waits for many stop calls. This system introduces a stoppage and evaluation index for preferentially allocating generated hall calls to -. By appropriately changing the weighting coefficient of the stop call evaluation index, it is possible to obtain a plurality of weights that optimize the waiting time, and conversely, by increasing the weighting coefficient, an energy saving effect can be obtained.

On the other hand, Special Publications No. 1982-70 and Special Publication No. 1982-71 also contain considerations for waiting time and energy saving, and on the other hand,
Special Publication No. 58-56709 describes an evaluation index in which the full forecast is added, and Special Publication No. 62-47787 describes an evaluation index in which at least one of the probability of incorrect prediction and the probability of fullness is added as an evaluation index. .

[Problem to be solved by the invention]

Among the above-mentioned conventional technologies, those that have two control objectives of waiting time and energy saving can alleviate passenger dissatisfaction by shortening the average waiting time.
However, it still had some problems. Specifically, long waiting times occur on certain floors at the same time every time, calls are assigned to elevators far away even though there is a waiting elevator nearby, and the number of wagons etc. Because the elevator was assigned to a crowded elevator even though it was transporting a large item, the elevator had to call again after the elevator departed.

Various complaints arose, and those were the building owners.

This was communicated to administrators.

In addition, no consideration is given to control target values other than waiting time, energy saving, waiting time and probability of fullness (probability of forecast failure), and although there are many other control targets, these controls are not applied to buildings. Owner.

It is difficult for a manager or the like to instruct them to do so, and it is not possible to control multiple objectives related to various combinations.

Once the building has been delivered, in order to respond to various complaints, it is necessary to ask the building owner, manager, etc.
The user requested modification from the Elevator-15N manufacturer, and the Elevator-IO manufacturer had to make changes and additions to the program and rewrite the ROM. The reality is that this requires a lot of manpower and time.
Furthermore, it is not possible to satisfy all the diverse demands of users, such as passengers getting on and off the building, building owners, managers, etc., and it is not possible to satisfy all of the diverse demands of users, such as passengers getting on and off the building, building owners, managers, etc. It also had the problem of being difficult to present.

In this way, the resources available in each building.

There are restrictions on the number of elevators, capacity, elevator speed, etc., and inputting desired values that can be achieved within these restrictions requires a lot of trial and error and experience. Furthermore, it is difficult to quantify the user's wishes without contradiction in advance. However, in the above-mentioned conventional technology, there is no idea of inputting the user's wishes, so no consideration is given to a method for determining the wishes without contradiction, and it is difficult to realize control tailored to the individuality of the elevator installed. was difficult.

The purpose of the present invention is to input user requests in an easy-to-understand format and incorporate them directly into the control system.

The purpose is to solve the above problem by immediately reflecting the problem in the control. Another object of the present invention is to provide an elevator control device and control method that allows users to easily add and delete various requested items.

It is also an object of the present invention to provide a sensibility input device and a sensibility input method for an elevator control device that can respond to various usage demands and assist the user in inputting appropriate requests in an easy-to-understand format based on sensibility. Our goal is to provide the following.

[Means to solve the problem]

The present invention has the following various features to achieve the purpose.

The first aspect relates to an elevator control device, which is characterized by having means for inputting a user's sensibility and means for executing group management control in accordance with this sensibility.

This allows users' sensibilities to be reflected in group management control.

Furthermore, if a means for converting sensitivity into a plurality of control targets is provided, application to group management control becomes easy.

Specifically, the sensory input means sets a combination of multiple sensory goals and the elevator usage environment (elevator operating conditions, traffic demand, etc.), and the goal conversion means sets the predetermined items of the sensory goals. A knowledge base is provided with a conversion function (corresponding to a fuzzy control membership function) for each environment in which the elevator is used, and this conversion function is used to convert the elevator control target value. (A HP (Ana
lyticalHierarchy Process
) using methods such as ). ) Second, regarding the elevator control device, control targets (at least, hall waiting time 9 car hours, reservation change rate, transportation capacity 2 number of passengers, long waiting time, reservation notification time, first-come-first-served car call rate, number of car passes, Number of complete passes, amount of information guidance, noise level, energy saving.

The present invention is characterized by comprising means for directly inputting two or more of the number of elevator starts and scheduled operation), and means for executing group management control using these control targets.

The third relates to the elevator control device, and the first or second
The present invention is characterized by comprising means for determining a control method in order to achieve the control target obtained in the above.

Specifically, means for selecting at least one control method candidate for achieving a control objective, means for determining predicted achievement values of each control objective for the selected control method candidates, and these predicted achievement values. The invention is characterized by comprising means for determining a control method based on. (It would be more efficient to calculate the predicted achievement value by performing a simulation.

) The fourth relates to an elevator control device.
The present invention is characterized in that a registration means (decision table) is provided that allows a driving method to be selected depending on the driving situation or time of day, and the contents of the registration means are added based on an external signal.

Fifth, regarding the elevator control method, emotional input is performed,
The present invention is characterized in that a call assignment evaluation formula and parameters for group management are calculated from this sensitivity, and group management control is executed in accordance with the evaluation formula and parameters.

The sixth aspect relates to a sensory input device for an elevator control device, and is characterized by comprising means for inputting sensory input and means for converting into a plurality of control targets.

The seventh aspect relates to a sensory input device for an elevator control device, and includes a means for inputting sensory input and converting it into a control target, a means for inferring and determining a control method, and a means for registering a control method in the elevator control device. It is characterized by being equipped. Specifically, the means for inputting sensibility and converting it into multiple control targets determines control target values by an inference function using a preset sensibility input knowledge base and environment/traffic demand database. AH weighting or priority between each item
The means for inferring and determining the control method uses the environment/traffic demand database and the control method determination knowledge base based on the control target value and the weighting or priority order between each item. The control method is determined using the following information.

The eighth aspect relates to a method for inputting the sensibility of an elevator control device, and is characterized by inputting the user's qualitative sensibilities in a predetermined guidance format.

[Effect]

Regarding the input of emotional goals (i.e., values, interests, tastes, emotional types, likes and dislikes, etc.), which are qualitative requirements for elevator operation, plain Japanese or laser charts are used to input the emotional goals of elevator control. It can be easily set even by non-experts, and the set sensory goals are converted into control target values using, for example, variable functions obtained through a questionnaire survey or the like. Regarding the multiple control i target values obtained (and considering the weighting or priority between each item of the control target as necessary), if there is a correlation between each item, control is performed based on preset rules etc. By replacing this method with a different method, it is possible to fully incorporate the requests of users and realize group management control that is unique to each building.

In addition, for users who have some knowledge of elevator operation, the control method can be determined by directly inputting control targets without inputting sensitivity.

Furthermore, it is also possible to directly incorporate sensitivity into the elevator control system without converting it into control targets, determine the call assignment evaluation formula and parameters for group management control, and reflect it in the actual group management control. It is possible if you apply

〔Example〕

Embodiments of the present invention will be described below with reference to the drawings.

1 to 18 show an embodiment of the present invention.

FIG. 1 shows the overall configuration of an elevator control device according to an embodiment of the present invention.

A control device for group management control according to the present invention is roughly composed of three blocks. One is a sensory goal setting means 2 that sets the user's sensory target value and the elevator usage environment such as the elevator operating condition and traffic demand to achieve the sensory target value. By providing this means, even people who are not experts in elevators can easily change the control method, make operation reservations, etc. This figure shows an input terminal (computer, computer, etc.) that combines a keyboard and display.
(workstation, etc.) 2-1 to input predetermined conditions 2-2 such as sensitivity target values.
Instead, a dip switch or the like is provided in the group management control device main body 1, and the conditions can be directly set (
It is also possible to change the control target of the elevator. Furthermore, it is also possible to use a telephone line or to directly set the settings in the group management control device main body 1 using a recording medium such as an IC card. The other block is the emotional goal setting means 2.
It was set in . It converts the user's request (sensitivity target value) into a numerical value using the actual control system. For each sensitivity item,
A target conversion function created based on data obtained in advance through user questionnaire surveys, etc. is extracted, and the extracted function is used to determine actual control target values to determine the control target in the best way. used to determine the control method to achieve. These control target values are stored in tables categorized by each condition. Note that this target converting means 3 may be provided outside the group management control device main body 1 instead of inside, or may be integrated into the sensory target setting means 2.

In the next block (control execution means 4), the control method that best achieves the control objective is determined, and the hall call control section 5,
Based on the data from each unit control unit 6, 7.8, the operating status of elevators, etc. is judged, and the previously determined control method (
It extracts and executes the call allocation method and transmits the results to each machine. By configuring the system in this way, user requests can be quickly reflected in the control, improving the usability of the elevator. Furthermore, by converting the actual driving results of each machine into a sensory target value using the target converting means 3, it is also possible to inform the user of the extent to which the user's movements have satisfied the set target.

FIG. 2 is a diagram illustrating the operations of the emotional goal setting means 2 and the goal converting means 3. The sensory goal setting means 3 includes a sensory target value setting section 2a and an elevator usage environment (for example, time of day, traffic volume,
and an elevator usage environment setting section 2b that sets the usage status of the elevator, the nature of the building in which the elevator is installed, the specifications of the elevator, etc. Incidentally, the emotional target value does not directly set the desired target value of the user as a physical quantity, but indirectly indicates the degree to which the user feels psychologically, such as the strength or weakness of preference. In other words, it can be said that it is a qualitative request for elevator operation that is derived from the user's sense of value, interest, preference, feeling, likes and dislikes, etc. For example, if you are not an elevator expert (designer, maintenance worker, etc.), even if you say ``I want to get on the train quickly,'' you may say ``I want to get on the elevator quickly.''
It is generally difficult to make a request with a quantified value such as "I want the waiting time to be faster than 00 seconds, because the appropriate waiting time for an elevator with a high speed of 000m/min is about 00 seconds." Therefore, the present invention attempts to solve this problem by introducing qualitative target values such as "I want to board an empty car" or "I want to board quickly" as emotional target values. Also.

The target value of sensitivity is control data that allows you to observe the elevator control status (hall waiting time, 2 boarding times).

It can also be said that the target value (control target value) for the reservation change rate 9 (number of passengers, number of passengers transported, etc.) is normalized and expressed as an analog quantity (digital quantity). The actual control target value is the type of building (hotel, department store).

- company-occupied office buildings, administrative office buildings, etc.), elevator specifications such as the number of service floors, number of elevators, capacity, etc.
Sensitivity targets vary depending on usage conditions such as traffic demand, but since they are normalized, they can be set freely. An example of one method of normalization is as follows. For example, if the minimum serviceable average waiting time during busy hours in a company-owned building is 25 seconds and the maximum average waiting time is 70 seconds, then the minimum average waiting time is 100 seconds. Normalize the maximum average waiting time as O. In other words, the sensitivity target value is 50x50=4
A value of 7.5 seconds is obtained, and a control method will be selected to make the time smaller than this value. Furthermore, sensibility can be said to be a relative expression of the magnitude of three or more items, and considering the various conditions of the elevator, it is possible to evaluate the operating performance of the elevator between the minimum and maximum values. It can also be expressed as being divided into stages. In this way, it is a transportation method that is introduced when it is difficult to quantify the control target value for direct elevator operation and express it in concrete numerical values after considering the control method, etc., and various other expression formats are also available. It is possible.

However, even if the elevator user requests that all the input items be completely satisfied, such as "I want to get into an empty car", "I want to get on the car quickly", etc., there are conflicting input items. (For example, during busy hours, it is difficult to simultaneously satisfy the desire to ``get in an empty car'' and the desire to ``get in early.'') considering operational efficiency and the like. Therefore, it is necessary to set in advance which one to give priority to in rare cases, and to rank them. (This ranking is possible no matter how qualitative the target values are because it is a matter of comparison.) As a method for determining this priority, the first method is to rank the input order (the order in which the emotional goals are input). A second possible method is to have the ranking set at the same time as setting the sensitivity target value. It is also possible to add weighting in addition to this priority order, and it is even more effective to use both priority order and weighting.

Furthermore, as mentioned above, the sensitivity target value is a qualitative target value that simply expresses the user's desired size, so the input sensitivity target value is used as the actual control target value. Unless it is converted into a control amount, the elevator cannot be controlled. However, this target conversion is not uniquely determined by the sensory target value alone, so when inputting this sensory target value, the elevator usage environment (elevator -
Elevator specifications such as number, speed, and capacity, hotels, etc.
Enter the specifications of the building (such as company-owned buildings, department stores, multi-tenant buildings, etc., 9 floors, 9 floors, 2 seasonal hours (or hours), usage status such as traffic volume, etc.). Then, the control target determining unit 3a in the target converting means 3 determines a function or rule to be used for target conversion according to the elevator use environment, and also determines a control target value corresponding to the sensory target value, and determines the control target value for each item. Then, the value is recorded in the control target table 3b. A specific method for converting this control target will be described later.

In Figure 2, the sensitivity target value is set as a numerical value from 1 to 5, but instead of this input method, as shown in Figures 3 and 4, you can use a radar chart or a mouse, etc. It is also possible to set it analogously. FIG. 3 shows an example of a hotel's sensory goals, which are set in advance by a designer or the like based on the usage status of elevators, etc. As shown in this figure, when guests use the elevator at a hotel, they do not want to meet other people as much as possible, or they want to ride in an empty car because they have baggage, etc. Due to the desire to minimize the amount of transportation carried by cars, there is a strong demand from people such as J, who want to be accurately notified of the arrival car, while there is a small demand for ``I want to ride quickly'' or ``I want to transport many people in a short time.'' However, at some point, an event or the like may occur and there may be a demand to transport many people in a short period of time. At this time, it is no longer possible to satisfy the request values for other items such as "I want to ride in an empty car" as in the past. The degree of this change changes depending on the amount of demand and importance of each item. The degree of importance of achieving this requirement is input by ranking in Figure 2, but there is also a method of having people rank it using a radar chart like the one shown in the figure, and a method of determining the set ranking as a ranking of importance. can also be considered. Figure 4 shows an example of sensitivity goals for a building occupied by a single company. In the case of this diagram, in consideration of work efficiency, the requests for items such as ``I want to be notified of reserved carts early'', ``I want to transport many people in a short time'', and ``I want to board quickly'' tend to be smaller.

By the way, it is necessary to set the specifications of the building in accordance with the sensitivity input, but when the properties of the building are shown in a radar chart, it can be seen that there are differences in the properties depending on the purpose as shown in Fig. 5. The solid line in FIG. 5 represents the characteristics of a single-company occupied building, and the dotted line in FIG.

FIG. 6 shows a configuration for converting the set sensitivity target value into a control target value.

As described above, when setting the sensitivity target value, the elevator usage information is also set in the elevator usage environment setting section 2b. This elevator usage information can be roughly divided into building properties 2bl, elevator specifications 2b2, and usage information 2b3. Once set, the specifications of the elevator and the nature of the building will not change significantly, but usage information 2b3 includes items that must be re-set whenever the user sets new requirements. (For example, items such as traffic volume and time zone 9th floor that need to meet new requirements). The emotional goal set by the emotional goal setting section 2a (not shown) and the elevator usage information set by the elevator usage environment setting section 2b are sent to the control goal setting section 3a of the goal converting means 3. The target conversion means 3 includes a control target setting section 3a (control target conversion function generation rule 3a1. conversion function database 3a2. target conversion section 3a3) and a control target data table 3b. Here, the control target conversion function generation rule 3al starts a conversion function selection rule set for each property of the building from the elevator usage information.

Conversion function item Sl corresponding to the emotional target item recorded in advance in the conversion function database 3a2.

S2...Function fsz...fs in Sn
+a is selected.

FIG. 7 shows an example of a conversion function, which corresponds to a fuzzy control membership function or the like. 7th
In the figure, for example, f ta(x t) or ftb(
xt) (here, fta(xt) is a function when there is a hall information guidance device, and ftb(xt) is a function when there is no hall information guidance device)
function is selected. That is, there is a hall information guide device. Considering the environment in which the elevator is used when there is no elevator, use control target conversion function generation rule 3al as rIF There is no hall information guidance device, AND I want to get on the elevator quickly THEN
The rule fl(x)=ftb(xt)J is searched, and the control target conversion function ftb(xt) is selected. Similarly, conversion functions are selected for the sensory goal items ``I want to ride in an empty car'' and ``I want to arrive early'' in consideration of the usage environment of each elevator.

In addition, in the example shown in FIG. 7, for the sake of clarity, there is only one type of judgment item for the elevator usage environment.
In reality, the conversion function to be used is determined by a combination of multiple items.Next, using the previous conversion function,
An example of converting a sensitivity target value into a control target value will be shown.

In FIG. 7, suppose that the sensitivity target value of "I want to ride quickly" is set to 4 at the time of goal setting, and that the conversion function is also selected as f s (x) = f tb (xt) according to the generation rule. From the conversion function, a target value Xk of the waiting time is specifically calculated as xt=40 seconds as a control target value.

This calculated value of 40 seconds is determined to be the maximum allowable waiting time (limit value), and is recorded in the control target table 3b such that the waiting time is set to a value of 40 seconds or less. Also, of course, the elevator must achieve this control goal.
The usage environment and weighting values are also recorded in this control target table 3b. In this way, qualitative and indirect sensitivity target values are converted into quantitative and direct control target values.

FIG. 8 shows an example of control target items corresponding to sensitivity target items. In Figure 8, sensory goals and control goals are described as having a one-to-one correspondence, but in reality, one sensory goal influences multiple control goals (for example, "I want to get on the train quickly"). In order to satisfy the items, it is not only the waiting time,
(The first-come-first-served car call rate and the amount of information provided also have a big impact.)

A plurality of control target values determined by the above method are transmitted to the control execution means 4 (see FIG. 1). The control execution means 4 shown in FIG. 9 uses control methods (for example, minimum waiting time allocation control (Min), maximum waiting time minimization allocation control (M
in-Max).

The knowledge in knowledge base 4e (
control method selection rules). (If there is only one control method that can be realized, then the control method is determined.) Several control method candidates derived in this way are:
The signal is sent to the elevator simulation experiment section 4g, which simulates the movement of the elevator using software. In the simulation experiment section 4g, data such as traffic volume and elevator specifications are retrieved from the control target table 3b and simulated using the plurality of control methods selected in 1, and the results are analyzed for each control target item. and calculate the predicted achievement value for each control objective. The calculated predicted achievement value is transmitted to the multi-goal decision making unit 4h,
Here, the control method is compared with the previously determined control target value, and the control method that achieves the highest degree of achievement is selected.

This superiority/inferiority comparison can be made using, for example, the overall target Δ, and the overall predicted achieved value as a vector f with the predicted achieved value f+(i=1...n) of each control target as an element.
= (ft, fz...fn),
A control method is selected in which the distance of the predicted achieved value vector f from the target Δ value vector f is small. The control method selected in this manner is stored in the control method database 4f for each elevator usage information.

Now, next, we will explain the online operation. In the control execution means 4, the signal S2 from each car control section 6 to 8 and the signal S□ from the hall call control section 5 are input to the elevator operation data collection section 4a. elevator
The operation data collection unit 4a uses the input data to collect elevator usage information data (traffic demand data, elevator position and direction, assigned call, car car) for a short period of time (from about 10 minutes before a new hall call occurs). number of passengers (number of passengers), etc.). The data obtained here is transmitted S3 to the learning system 4b.

The learning system uses this data to learn traffic demand by time of day, waiting times, and other data. Using the learned information S6 and the short-term usage information S3 created by the data collection unit 4a, the control method selection unit 4d selects the elevator
Determines the driving status of the vehicle.

Further, in this control method selection section 4d, as described above, the control method selection knowledge base 4e, the simulation experiment section 4g, and the multi-objective decision making section 4h determine the control method selection for a plurality of control target values input in advance. From the control method database 4f in which control methods have been stored, a control method suitable for the actual operating condition of the elevator is selected using control method selection rules.

The control method selected here is sent to the group management control system 4c.
At this point, each machine is evaluated using the transmitted control method to determine the selected machine, and a new hall call assignment command is transmitted to the selected machine S2.

FIG. 10 shows an example of a call allocation control method selection rule. As shown in the figure, the rule table is composed of three parts. Tll is a registered rule table that indicates whether or not rules to be applied in each direction of each floor are defined, and rules are defined where a Q mark is placed between them. In this example, rule 3 is applied to the first floor ascending call.
is being WBed, and rule 1 is applied to the third floor ascending call.
This shows that Rule 3 is registered.

Other parts marked with an O mark indicate that respective rules are registered, as in the above explanation. T
12 is a rule condition table in which the condition part of each rule is recorded. The conditions include specifying the day of the week 9 time corresponding to the elevator usage information, traffic demand, etc., and the control target value (waiting penalty 9 occupancy rate).

call notification time etc.) is set. All of these conditions are handled as AND conditions, and when handled as OR conditions, they are registered as separate rules. Each of these data is written in a judgment conditional expression. For example, the condition is that the occupancy rate is 30% or less.

WEIGHT(K)=<5EKISAI(K) Umbrella 0.3
...(1) However, WEIGHT (K): Number of passengers on board K 5EKISAI (K): Capacity of K car.

T13 is a rule execution table that records the execution part when the conditions of each rule are satisfied, and records the evaluation formula or assigned machine number. For example, the evaluation formula for selecting the elevator that can arrive the earliest from elevators with a capacity of 30% or less is...(2) ASIGN=K FORMINE[VALUE(K
) Co... (3) However, VALUE (K): Array K of evaluation values
: Variable corresponding to the machine number ASIGN : Assigned machine number MAX : Maximum value. The data actually recorded in each table is obtained by converting the above equations into two-way data that can be executed by a microcomputer. Also, a blank field means that there is no condition.

The operations described above are shown in a flowchart as shown in FIG. 11. First, the elevator usage environment is set (E10). Subsequently, a sensitivity target item corresponding to EIO is selected and a target value is set (E20). According to the set elevator usage environment and sensory goals, the control target conversion function generation rule is activated and the conversion function is determined (E30
). However, if the control target value is uniquely determined, the control target value is determined by the generation rule. Subsequently, the sensitivity target value is converted into a control target value using the conversion function selected at E30 (E40). The converted control target value is recorded in the control target table (E50). From here, control execution means 4
The work will be transferred to Control method candidates are selected using the values in the control target table and the knowledge of experts set in advance (GIO).

The selected multiple control methods are sent to the simulation experiment department, which uses a system that simulates elevator operation using software.
Simulated operation according to set conditions (G20)
. As a result of this operation, predicted achievement values for each control item are calculated (G30). The predicted achievement value of each control method is compared with the input target value using the multi-goal decision-making method, and the control method with the best degree of goal achievement is selected (G40).

It should be noted that the control method is presented to the user, and if the user is not satisfied with the control method, the processes of emotional goal setting E20 to G30 are repeated.

The best control method may be selected. The selected control method is stored in the control method data table together with the elevator usage information (G50).

Above, E10 to E50. The processing from GIO to G50 can be done offline from the perspective of elevator control.

Next, the elevator operation control is started by the elevator control execution start command LIO. First, the processing of the call signal for each landing, that is, the landing call input processing is executed (L20). Furthermore, a communication process for exchanging various data from each machine control system is executed (L30). Usage information is obtained based on these data and the like, and the control method to be used is selected from the control method database using the obtained usage information (L40). Next, the optimum call allocation car is determined by the selected control method and call allocation processing is executed (L50). Based on each hall call assignment information, estimated elevator arrival time, etc., the guidance content on the guidance display of each hall is determined and notification processing is performed (L60). In addition to the above processing, processing such as outputting and displaying various data (L70) is executed.

Next, after completing this series of tasks, it is determined whether or not to continue the operation (L80), and if the operation is to be continued, L2
If the process of 0 is repeated again and the operation is terminated, the operation is terminated.

Using FIG. 12, a schematic flow of the process visible to the user in the flow shown in FIG. 11 will be shown. First, when the user turns on the switch, a screen (for example, Figures 3 and 4) in which the initial phase emotional goals are set is displayed (EO).
. Next, the user inputs the target value of the target to be changed on the initial setting screen using a mouse, etc. (E20)
. After inputting all the change values (transmitting a change completion signal), the sensory target is converted into a control target using means such as a conversion function (E40). Next, control method candidates that are likely to best achieve each control objective are selected using the control objective table and knowledge base (GIO).

There may be one control method candidate or multiple control method candidates selected here, so it is necessary to determine whether there is only one control method candidate or multiple candidates (G15). If there is only one control method candidate, that method can be immediately It is sent to section G20'.

The simulation experiment section G20' executes a simulation using the traffic flow data set at the time of inputting the sensitivity target. From the experimental results, a predicted achievement value of the previously set control target is determined (G30'). The result is reversely converted into an emotional goal and displayed together with the emotional goal previously set by the user (G45').

On the other hand, if multiple control methods are selected, mock experiments are conducted for each control method (G20
), and as a result, a predicted target achievement value for each control target is calculated for each control method (G30).

Next, the calculated predicted target achievement value and target value are compared, and the overall achievement level is calculated for each control method (G35).

This overall degree of achievement is compared for each control method, and the one with the largest (or smallest) degree of achievement is determined as the optimal control method (G40).

The predicted target achievement value of the selected control method is inversely converted into a sensory target value and displayed together with the sensory target value previously set by the user (G45). This display screen allows the user to
If it is determined that the control method determined by the system is sufficient (G6
0), send the method to the control method table (G70
). If you decide that this method is not good, try E20 again.
Light it repeatedly.

First, the off-line processing from inputting a sensory target to recording a control method in a data table will be described based on a specific example.

First, the hotel and elevator specifications (number of elevators, speed, service floors, etc.) are set as building specifications for the elevator usage environment, as shown in Figure 13, and furthermore, as elevator usage information, the reference floor ( Suppose that the traffic is normal and check-in times are set (
Figure 11 EIO).

At this time, assuming that the emotional target items are the six items shown in FIG. 14, it is assumed that the values shown by the solid line are set in advance as initial values on the system side. (Of course, it goes without saying that the configuration may be configured such that the user directly sets the value.) The user can input the sensitivity target value "I want to ride quickly" in this figure using an input means such as a mouse. Assume that the setting is changed from the value of 4 to the value of 4 (E20 in FIG. 11). At the same time as this setting change, enter the priority of the achievement as well. If the priority of the achievement is the same as the initial setting and the priority of the item for which the setting change is requested is lower, then it can be achieved. The control method remains the same, so customers are asked to change their priorities. Here, if the target value becomes larger than the initial setting value due to a setting change, the priority will be placed first, and if it becomes smaller, the priority will be placed sixth (last). .

Next, a control target conversion function generation rule is activated to convert the sensitivity target value of "I want to ride quickly" into a control target value. An example of this rule (for hotels) is shown below.

1) Rule group rule I for the item “I want to board early” IF hotel AND check-in time T HE N fl (X) = f th (x t)
Rule 2 IF Hotel A N D Lunch Time T HE N f 1 (x ) = f th (x 2)
A target conversion function with conditions matching the rule is selected (010 in FIG. 11). This selected target conversion function (in this case, it is assumed that f 5hcx t in FIG. 15) has been selected is sent to the target conversion means 3, where a control target is calculated. In other words, the initially set sensitivity target value is 1.
Therefore, the waiting time was previously set to 40 seconds or less, but with this setting, the sensitivity target value was set to 4, so this target value was set to 25 seconds or less (first
1 Figure 010). The previously set control target table value is changed based on this target value (E50 in FIG. 11).

At this time, the weight of the control target table is also changed.

FIG. 16 shows the state before and after the current control target change.

Control method candidates that satisfy these control objectives are selected using a knowledge base that has been set in advance based on the knowledge of experts (FIG. 11, 010). An example of this selection rule is shown below.

Rule I IF time zone = T1 and traffic volume ≧a and front floor THENφ” T s Rule 2 IF time zone = Ttand traffic volume ≧a and general floor TH
ENφ” T z Here, Ts is the check-in time, the traffic volume is a person/vehicle,
5 minutes are shown, and TI and T2 show the call allocation method. Here, the call allocation method T□ is the minimum waiting time allocation method (Min), and its evaluation value is determined by the following equation.

To φ = T to one α・TAK+α′ ・TIIK ・
...(4) Tt=min(φ1. ”””φK)
-(5)- Now, φ is given. Second is the hall waiting time evaluation value T of the machine. : is the time T for the machine to arrive at the newly assigned hall call floor. The stop call evaluation value Ta takes into consideration: load concentration evaluation value α according to the elevator condition, α′: weighting coefficient. Next, Tz is the maximum waiting time minimum call allocation method (M
i n −M a x) and can be obtained using a linear equation.

φ to demon: TKI + TPI-α・T^ to elixir + α'
TBKI...(5) 1 assigned to machine number 1 to φ
Evaluation value of the floor hall call TKI: The time from the current time until the car arrives at the i floor Tpt: Longitudinal travel time after the 1st floor hall call occurs TAKI: The stop call evaluation value TB of the car: Load concentration evaluation value α・α′ according to elevator condition:・Weighting coefficient φx=max (φKl, “””φKi)
””(6) T2=llin (φwork, φZ”””
φK) ”(7).In other words, find the evaluation value for each hall call assigned to each car (formula (5)), and select the maximum evaluation value (formula (6)).
, the call is assigned to the one with the smallest evaluation value.

In addition, rules for selecting formalized allocation evaluation methods are prepared in the knowledge base, such as the average waiting time allocation evaluation method, the waiting time distribution allocation evaluation method 9, the equally spaced car allocation evaluation method, and the evaluation method that combines these evaluation methods. has been done.

Now, in this control method selection rule, in addition to the call assignment method, various operating specifications are also selected. In the hotel example mentioned above, a lobby control method is also commanded to return the elevator, which does not have a call reception, to the standard floor. Here, it is assumed that as a result of the user inputting the sensitivity target value, two control methods, the minimum waiting time evaluation method and the average waiting time evaluation method, are derived as possible to be realized. In addition, in the initial setting on the system side, the initial set target value was obtained by the longest waiting time minimum call allocation method that suppresses the occurrence of long waiting times. However, since the minimum waiting time allocation evaluation method and the average waiting time allocation evaluation method were selected as candidates, the selected two
The type of call allocation evaluation method and operation specifications are sent to the simulation experiment department and a simulation experiment is conducted (G20 in FIG. 11). As a result, results obtained using the two types of allocation evaluation methods and other operating specifications are obtained. If only the allocation evaluation method is changed this time, predicted achievement values for each control target can be obtained as shown in FIG. 17 (630 in FIG. 11). From this result, we will calculate Qp for each allocation method for i=0 in the multi-objective decision-making method, which is one of the methods for evaluating the goal achievement value described above.

This time, as a method for evaluating the target achieved value, the value of the item that satisfied the set target value is evaluated as zero, and the unachieved value is multiplied by each weighting coefficient value, and the magnitude of the target unachieved value is calculated. Find it with Also, the weighting coefficient ω.

Let be the weight set at the time of setting each control target. In front of
The degree of achievement is determined for each control method shown in FIG. 17, which was previously obtained through a simulation experiment. For example, the achievement level (lpt) for the minimum waiting time method is as follows.

Qpt=8(4030)+7(5-3)+6(33
)+9(25-25)+5(-35+30)+4(0,
1-1) In the above formula, the first item is the occupancy rate, and the second item is the reservation change rate.

The third item is the first-come, first-served rate, and the fourth item is the hall waiting time.

Item 5 is transportation capacity, and item 6 is reservation notification time.

Here, the target value of the five items of transportation capacity is 30 people/min or more, and it is better if it is larger than the target value, and the smaller the remaining terms are, the better the remaining terms are, so the equations for adjustment are different.

Qpl is determined using the above conditions.

Qpx=8X10+7X2+6XO+9XO+5XO+
4XO=94. Similarly, when calculating the average waiting time method, 12
pz=147. As mentioned above, in order to make the explanation easier to understand, the negative term was set to zero, but in reality, the weighting coefficient ωi for each evaluation item was set to zero this time as a coefficient for flattening each control item. There is also a method of evaluating in addition to the evaluation value the good achievement rate you set.

As a result, the call allocation method with the smallest waiting time is selected. The control target value of the selected control method is again converted back into a sensory target and displayed together with the initial setting value in the form of a radar chart (dotted line in FIG. 14). If the user determines that this method is acceptable, the selected control method is registered and recorded in the control method data table.

The processing from EIO to G50 in FIG. 11 has been described above using a specific example. The explanation so far has been based on the assumption that all processing is performed within the group management control device.

A modification of one embodiment of the invention is shown in FIG. The present embodiment comprises a sensory goal setting support system 9 and a group management control device 1, and the first group management control device is the multi-S decision making means 4h of the control execution means 4 described above.

The parts are the same except for the simulation experiment section 4g and the knowledge base 4e. That is, the off-line operation part is taken out and made into a separate device.

The sensory goal setting support system 9 includes a sensory goal input/output section 2a' corresponding to the aforementioned sensory goal setting means 2, an elevator usage environment input setting section 2b', and a control target conversion function generation unit corresponding to the goal conversion means 3. It was included in the rule 3al, the conversion function database 3a2, the target conversion unit 3a3, and the control execution means 4. It consists of a control method selection rule 4he that is a combination of a multi-purpose decision making section and a knowledge base, a simulation experiment section 4g, and a control method database section 4f.

Here, the function is to determine a control method that satisfies a plurality of input sensory goals, and to record the determined control method in a control method database within the group management control device. With this configuration, user requests can be input in the form of sensitivities and converted to control values completely separately from the main body of the group management control device, so, for example, the sensibility target setting support system can be connected to a handheld computer (laptop). etc.), and that it is possible to conduct meetings with users using various terminal devices even when the elevator is completely away from the building where the elevator is installed.

Furthermore, as another modification, if the control target value can be directly inputted in advance, it is also possible to have a configuration in which the sensitivity target value setting and the conversion unit to the control target are removed.

Other embodiments are shown from FIG. 19 onwards. This has almost the same configuration as the sensory goal setting support system 9' and the group management control device 1 explained in FIG.
AHP (Analytic Hi)
It is characterized by the fact that it is carried out using a method such as error process.

FIG. 19 is an overall configuration diagram, and the individualization support device 9' has almost the same configuration as the emotional goal setting support system 9 of the embodiment shown in FIG. 18.

In other words, input the user's emotional target value and the usage environment of the elevator, such as the elevator operating conditions that should achieve the emotional target value, to the terminal ■ (this is the same as 2-1 in Figure 2).
Kansei input knowledge base 9'-4, environment/
A sensitivity input unit 9'-1 that converts into a control target value (and 1 weighting or priority order) using a traffic demand database 9'-5, etc., and an environment/traffic demand database 9'-5 and a control method determination knowledge base 9. The group management control device 1' includes a control method inference section 9'-2 that selects and determines a control method by inference using the individualization support device 9'-6.
′ data table (
Usually called a decision table. (not shown)
Based on the signals from the machine control devices 6'-1 to 6'-4 of each machine and the signals from each hall, the operating status of the elevator is determined, and the control method in the data table is retrieved and executed. For example, there is a learning system 1'-1 that collects usage data such as waiting time and number of passengers and determines characteristic modes, and an intelligent system that generates new characteristic modes and creates new driving programs. 1'-2, etc. (see Japanese Patent Laid-Open No. 59-48369).

In FIG. 20, the knowledge acquisition section 9'-3 is included in the individualization support device 9'.
An example is shown below.

The execution results for each control target and the situation at that time collected by the school system 1'-1 of the group management control device 1' are sent to the individualization support device 9', and if the degree of achievement of the control target value is low, The knowledge acquisition unit 9'-3 requests a review of the control method determination knowledge base, or requests permission to register it as new data if the environmental/traffic demand data is significantly different from the set data. By adding this, it is possible to improve the accuracy of control method selection and determination in the individualization support device 9'.

Next, the specific operation of the individualization support device 9' will be explained.

Figure 21 shows the emotional input section 9'-1 and the emotional input knowledge base 9.
'-4 etc. is an explanatory diagram of the operation.

The user (customer) first inputs the type of building (for example, (1) office building, (2) hotel, (3) department store, (4) hospital, 5) Government offices, (6) Multi-tenant buildings, etc.)
Select. In FIG. 21, an explanation will be given assuming that an office building has been selected.

Next, typical elevator specifications as shown in FIG. 22 are presented based on the type of building.

If the specifications need to be changed, they can be modified from the input terminal I using a keyboard, mouse, or the like. After completing this correction, enter environmental and transportation demands as necessary6.
Then, the user's sensitivity is input in a question-and-answer format (guidance format). An example of this emotional input is shown in FIG. 23. In other words, for each of the emotional goal items such as "Q.1. Do you want to get on the elevator quickly?", you can say, "1) Especially 2) Somewhat 3) Average 4) Not that much, 5) I don't care." Prepare 5-level answers, and use this answer as 4) (equivalent to the sensitivity target value)
is selected, inference is made using the emotional input knowledge base 9'-4 (9b') of the individualization support device 9', and rule 1 is selected, which sets the specific control goal of waiting time of 40 seconds or less. Convert to control target value. (In this example, there is a one-to-one correspondence between the sensory goal items and the control goal items.)
In addition to this method of sensory input, various methods can be considered, and as explained in the previous embodiment, input can also be made using a radar chart or the like. For example, standard patterns of sensitivity target items and target values are presented in a radar chart according to the type of building, specifications, usage environment, etc., and based on that pattern, users can determine the desired size. The setting (correction) method is effective for users who are not familiar with elevator control (see Figure 4 for an example of a laser chart for an office (-company-occupied) building). . In addition, since it is impossible to achieve the optimal wish for all items that have been emotionally input, (1) inform the user of the range of target values that can be changed for that item (for example, ``Waiting time is 3
"Can be set in the range of 0 to 40 seconds" is displayed. (2) It is also possible to calculate the control target value by giving the user an overall score (e.g. 20 points) and having the user allocate it proportionally between each item. It will be done. Of course, users who are able to set the settings may be allowed to directly set input items and target values according to their sensibilities.

By the way, as I mentioned earlier, this emotional goal is.

It expresses the words used in machines in words that the general public can understand, and this value is control data (hall waiting time) that allows you to observe the elevator control status.

Boarding time, reservation change rate, complete pass rate9 Number of passengers.

It can be said that the target value (control target value) for the number of people transported, etc.) is normalized (divided into five stages, etc.) and expressed as an analog quantity or a digital quantity.

Furthermore, one sensory goal may include two or more control goals.

Furthermore, in consideration of the input sensory target value and the environment/traffic demand that must achieve the request, inference is made using the sensory input knowledge base 9'-4 (9'b) in Fig. 21, and the control target is determined. The production rule shown in the emotional input knowledge base 9'-4 (9'b) in Fig. 21 uses the conversion numbers and conversion tables (Fig. 7, etc.) explained in the previous example. ), whereas it is necessary to convert uncertain elements (including climate, etc.) into control target values using a dedicated conversion function or conversion table. (The explanation of this point is omitted as it is the same as in the previous example.) Qualitative sensitivity goals are converted into quantitative control goals using this method, but the sensitivity goal items and control goal items There is not always a one-to-one correspondence between the two, and in reality, a sensory target item often influences a plurality of control target items. For example, when an emotional goal item such as "I want to arrive early" is input, in addition to a request to shorten the waiting time and a request to shorten the riding time (boarding time), Various requests are included, such as a request for not wanting to wait for a long time (reducing the longevity rate).

Therefore, we divided the emotional goal items into multiple control goal items.
It is necessary to divide it into 9 parts and further find the target value for each. - As an example, the degree of mutual relationship between sensory requests is determined in advance by means such as simulation according to the usage environment of the elevator and traffic demand, and the relationship speed table shown in FIG. For example, there is a request to "arrive early".

Assume that 4 is input as a normalized numerical value between 0 and 5.

Assuming that the usage environment at this time is ■ in the speed table in Fig. 24, in this example, the speed is expressed as a numerical value from 0 to 5, so the demand for each control target is What is the degree?

I want to ride quickly (waiting time): 4X-=2.4 I want to shorten the time I ride (boarding time): 4X-=
4. I want to reduce the number of long waits (long-lasting rate): 4X-=1.
6 I want to ride in an empty car (crowd level in car): 4X-=0.
The sensitivity target value is reset as shown in 8, and this value is used to convert it into the control target value using the method described above.

Furthermore, by making full use of the sensory input knowledge database 9'-4, it is possible to uniquely determine the control target value from the speed of the vehicle without determining the control target value in two stages.

As described above, the control target value is determined from the sensitivity target, but there are cases where it is impossible to perform control that satisfies all the control target values, and the control method for performing group 1 control cannot be uniquely determined. It is highly likely that the control target value that has been set will not be fully reflected. Therefore, it becomes necessary to set weights or priorities for each control goal (sensitivity goal) as to what is considered important. Of course, this weighting etc. can be determined directly by the user as explained in the previous embodiment, but it is also possible to directly determine this weighting etc.
It is effective to use a method called ``erarchy process''.

This AHP will be explained using FIG. 25 and FIG. 26. Customers' requests include, ``I want to get on the train quickly (waiting time)''
”, “I want to carry many people (transportation capacity)”, “I want to reduce the time spent riding (multiplying time)”, “I want to be informed of the reservation cart accurately (predictive accuracy rate)”, “I want to know the empty cart Assume that there are five items: ``I want to ride in the car (crowd level)''. For these requests, as shown in FIG. 25, the importance of each request is input interactively by pairwise comparison. From these set values, a pair comparison table as shown in Figure 26 is created, a pair comparison matrix A is created from this table, and the eigenvector V of this pair comparison matrix A is determined.・The importance of each requirement item ( weighting coefficient value).

In the above example, the waiting time at the hall is 0.513, and the boarding time is 0.261.

Note that if the association degree table shown in FIG. 23 is used, 1 weight is included in the association degree value, so there is no need for new weighting such as AHP or setting of priority order. Further, when a method such as limiting the target value range or assigning points during emotional input is adopted, weighting or priority setting may be omitted in some cases.

By the above method, the customer request table 9'c shown in FIG. 21 is created. This customer requirement table 9'c
A control target value and weighting (priority order) are stored in , and a control method reasoning unit 9'-2 then determines a control method that satisfies the control target value (see FIG. 20).

In making this selection decision, first, various operation methods selected for each building type (for example, operation with an express zone, skip operation that services every other floor or every multiple floors, direct operation to a specific floor, and output order) A rule is selected that determines a preset operation, an operation that preferentially services a specific floor, and an operation that provides priority service to a specific floor. For example, if a request is input that emphasizes transportation capacity during dispatch hours for an office building, if there are 6 elevators installed, 3 will be allocated to lower floors and the remaining 3 will be allocated to higher floors. If multiple vehicles are waiting on a standard floor, the departure order can be determined and the number of passengers or departures set at certain time intervals can be used to improve transportation capacity. It is also possible to determine one or a combination of two or more of these operating methods. Next, the call assignment control method and parameters to be used for the determined operation method are selected.

In selecting these control methods, the environment/traffic demand data table 9'-5 is referred to and the control method determination knowledge base 9'-6 (rules) as shown in FIG. 27 is used. The predicted achieved value when using the selected driving method and call allocation control method and the parameters used is estimated, and sent and converted to a sensory target by the sensory input section 9'-1, and displayed on the input terminal I. That is, after presenting the predicted achievement value to the customer and obtaining the customer's approval, a decision table 1'-- provided in the group management control device 1' shown in FIG.
Record in 3. Although FIG. 28 is expressed based on time zones, it is also possible to create a table divided by traffic flow pattern, and driving directions can also be recorded in the same table. The group management control device uses the traffic flow patterns learned by the learning system and the built-in clock.
It judges the situation, calls up a decision table corresponding to each situation, and executes control appropriate to the contents.

As modifications of the other embodiments of the present invention described above, the following may be made as appropriate.

First, as shown in FIG. 20, a knowledge acquisition section 9'-3 is provided within the individualization support device 9'.

The learning system 1'-1 collects data corresponding to each control objective as a result of actually controlling the elevators with the group management control device 1', and calculates the data for each environment, traffic demand, demand, vehicle position per day, and one week. Record the results of statistical processing in units. In addition, in this learning system 1'-1, items that are not directly related to customer requests but have an indirect influence, such as boarding refusal rate, travel time for floor height measurement, abnormal waiting times, and the elevator at that time. state.

Also learn data such as the occurrence of long-term stoppages.

The above learning data can be presented to the customer via the individualization support device 9'. In addition, individualization support device 9
' receives the learning data, and the knowledge acquisition unit 9'-3 compares the learning data with the control target value previously created based on the sensory input, and if the target value is satisfied, the customer understands the actual measurement result. Convert it back into the form of sensibility that you can and present it using a radar chart, etc. If a term that does not satisfy the target value occurs, are there any unusual phenomena occurring in the usage situation at that time (for example, are boarding refusals occurring frequently? If this phenomenon occurs, the control method inference unit 9'-2 reselects another control method using the usage environment data. If there is no selection rule that matches the usage environment data, inference is made using the closest usage environment data.

Compare the control method and parameters obtained by this inference with the previous one, and if they are the same as the previous one,
Notify to create a new rule. If a different control method or parameter is selected, we will present to the customer whether it is acceptable to use it in the actual measurement environment, and after receiving permission, we will install it in the intelligent system 1'-2 of the group management control device 1'. Record in decision table 1'-3. Note that this knowledge acquisition section 9'-3 has a learning system 1'-1 of the group management control device 1'.
When it is detected that new environmental or traffic demand data has been generated, it has a function to add the new data to the environment/traffic demand database, and a means to notify that it has been added.

As another modification, as shown in FIG. 29, a simulation experiment (simulator) function 9'-
7 is added, the knowledge acquisition section 9'-3, the sensitivity input section 9'-
The degree of achievement is determined for the control target value set in step 1, and if this degree of achievement is small (or large), the control method inference unit 9' uses the actually measured environment/traffic demand data.
-2, reselect the control method, predict the execution result using a simulator, confirm that the control objective is satisfied, and then change the control method to the decision table 1'-3 in the group management control device 1'.
The control method is sent to the

In addition, if the control method inference unit 9'-2 selects the reselected control method and evaluation parameters that are the same as those used in actual control, the target value The control parameters related to the terms that do not satisfy are extracted, the extracted control parameters are varied, a simulation is executed, and the value that best matches the target value is registered as a new parameter.

Next, an example of the overall processing flow will be explained using FIGS. 30 and 31.

First, set the building type using the individualization support device (101)
. Next, typical elevator specifications of the same building type are presented (102), and the customer decides whether or not the presented specifications need to be modified (103), and if modifications are necessary, the specifications are modified (104).

Next, the customer is asked to input the customer's emotional goals (105).

The input target is converted into a control target value using the sensory input knowledge base (106). Next, inquire whether the priority order between each control target value can be set by the customer (107).
, if it cannot be set, a relative comparison value is set between the two targets using a pairwise comparison using AHP (108). Based on this input result, a pairwise comparison matrix is created and eigenvalues are calculated (109).

This eigenvalue is used as a weight. If one weight has been determined, the customer is asked to input the weight (110).

The control target values and weights determined here are recorded for each control target as a customer request table (111).

Next, this table is sent to the control method reasoning section (112).

The control method inference unit selects a control method using the control method knowledge base and the environment/traffic demand database (113
), if there is only one type of control method selected, the best control method is determined (116), and the selected control method is transmitted to the control device with the consent of the user (sensitivity input person) (
117). If it is determined that there are multiple
4) Calculate the predicted achieved value of each fiI control target value by simulation (115), compare this value with the previously determined control target value, and determine the control method with the highest degree of achievement as the optimal control method ( 116). The predicted achievement value of the determined control method is presented to the customer and, with the customer's consent, transmitted to a decision table provided in the group management control device (117).

Figure 31 shows the moving body of the knowledge acquisition unit. First, read the data actually measured by the group management control device (2
01). Next, among the read data, the usage environment/traffic demand data is compared with the environment/traffic demand for which target values were previously set (202).

From this comparison result, it is determined whether the condition is new (2
03). If it is a new state, the control target value is compared with the actual measurement data (204), and it is determined whether the target value is satisfied. If the target value is not satisfied, record the new usage environment in the environment/traffic demand database (206
). Next, the control method inference unit selects a control method using the new usage environment data (207). It is determined whether there are multiple selected control methods (208). If multiple control methods are selected, a simulation is performed for each method using the measured new usage environment data (
209). The result is compared with the control target value, and the control method with the highest degree of target achievement is selected (21Q). In addition, if the previous control method selection result is a single control method, if it is different from the previous control method (211), that control method is considered to be the best control method, and that control method is used as the decision of the group management control device. Send to the table (212). Also, if the usage environment matches the environment/traffic demand database (203), compare the control target value with the measured data (21).
3) If the target value is satisfied (214), the work of the knowledge acquisition unit ends. If the target value is not satisfied, extract the target item that does not satisfy the target value (215)L, extract the rule related to the target item, change the parameters described in the rule, and run the simulation.217
), the parameters that best satisfy the goal are registered as new rules, and the old rules are deleted. Further, if the target is satisfied in (205), the new usage environment and the control method at that time are registered (218).

As described above, it is possible to extract unsatisfied rules for new environments and rule corrections based on actual driving results, and fine-grained driving control can be achieved.

Finally, although the present invention has been described with reference to group-controlled elevators, it is also applicable to individual elevators and the like. In other words, the so-called user's needs are required for operation control methods (for example, setting express zones and non-stop floors, realizing VIP services, etc.), guidance displays for hall information, which has been increasing recently, and control of door opening/closing speeds. It is also possible to implement control by introducing sensitivity, and to improve man-machine characteristics by introducing support tools.

〔Effect of the invention〕

According to the present invention, in order to automatically convert an elevator user's sensory goals (qualitative quantities) into actual control goals, a desired control method is executed without the help of an elevator expert. It becomes possible to do so.

Furthermore, regarding the results of actually operating the elevator after delivery, we first check whether the control desired by the user is being executed or not, and use the operating data to convert it into a sensory target or control target value. It is also possible to diagnose the elevator usage situation. Furthermore, since the control method used by the control execution means is adopted after the user is satisfied with it, the usability for the user is improved.

In addition, as explained in the embodiments, it is possible to realize various controls that suit the individuality of the elevator for the user without using the expression sensibility, and it is also possible to implement various controls that suit the individuality of the elevator for the user, and also as an input device and input method for the control device. It has excellent effects.

[Brief explanation of the drawing]

FIG. 1 is an overall configuration diagram of an embodiment of the present invention, FIG. 2 is an explanatory diagram of the operation of the emotional goal setting means and goal conversion means, FIGS. 3 and 4 are examples of emotional goal setting, and FIG. Laser chart showing the properties of the building, Figure 6 is an explanatory diagram of the control target conversion method, Figure 7 is an example of the target conversion function, Figure 8 is a correspondence diagram between the sensory target and control target, and Figure 9 is the control target. An explanatory diagram of the operation of the execution means, FIG. 10 is an example of a control method selection rule, FIG. 11 is an operation flowchart of an embodiment of the present invention, FIG. 12 is a flowchart of the part visible to the user, and FIG. 13 is an example of a control method selection rule. Example of elevator specification settings, Figure 14 is an example of change in sensitivity target settings, Figure 15 is an example of a target conversion function, Figure 16 is an example of a control target table, Figure 17 is an example of predicted achievement values, and Figure 18 is an example of a predicted achievement value. FIG. 19 is an overall configuration diagram of a modification of one embodiment of the present invention, FIG. 20 is an overall configuration diagram of another embodiment of the invention, and FIG. An explanatory diagram of the operation of the sensibility input section of another embodiment, FIG. 22 is an example of typical specifications at the time of sensibility input, FIG. 23 is a diagram showing a method of inputting a sensibility target, and FIG. An example of a relevance table for determining a control target, Figure 25 is a diagram showing how to input weights using AHP, Figure 26 is a diagram showing pairwise comparison using Al-I P, and Figure 27 is a knowledge base for determining control method. 28 is an example of a decision table, FIG. 29 is an overall configuration diagram of a modification of another embodiment, and FIG. 30 is an example of a decision table.
31 and 31 are processing flowcharts of another embodiment. 1...Group management control device, 1'...Group management control device,
2... Emotional goal setting means, 3... Goal conversion means, 4
... control execution means, 5 ... hall call control section, 6,
7.8... Unit control unit, 9'... Individualization support device,
9'-1... Emotional input section, 9'-2... Control method inference section, 9'-3... Knowledge acquisition section, 9'-4... Emotional input knowledge base, 9'-5 ...Environmental/traffic demand database, 9'-6...Control method decision knowledge base, 9'
-7... 3rd Hote Shino, Mud\, 10th O1 Iy'j 4th Eye 1 company-occupied building formula I raw material 11 5th mouth 6th eye 7 Go to figure (αtsu 1i No times 8th mouth 11th prisoner 12th concave 13th mouth 74th mouth Hotel's inclusion raw goal change 1 dream 15th mouth 76th mouth 17th day 19th mouth 9' 20th country 22nd country 23rd ward 24th 25th 27th day of Showa 28th day 1'-3 29th ward 31st day

Claims (1)

[Scope of Claims] 1. In an elevator control device that performs group management control of a plurality of elevators in service between multiple floors, means for inputting a user's sensitivity toward the elevator; Elevator control device with means for carrying out administrative control. 2. The elevator control device according to claim 1, wherein the sensitivity is a normalized representation of a qualitative request for the elevator operation. 3. The elevator control device according to claim 1, wherein the sensitivity is a relative representation of magnitude between three or more items. 4. The elevator control device according to claim 1, wherein the sensitivity represents the operational performance of the elevator divided into a plurality of stages between a minimum value and a maximum value. 5. The elevator control device according to any one of claims 1 to 4, wherein the sensory input means inputs a plurality of sensory goals and the environment in which the elevator is used. 6. The elevator control device according to claim 5, further comprising means for converting the sensitivity into a plurality of control targets for the elevator, and executing the group management control using each control target. 7. The elevator control device according to claim 7, wherein the control target conversion means converts at least the control target value using knowledge-based rules. 8. The knowledge base includes a conversion function for each usage environment of the elevator for the predetermined item of the sensitivity goal,
8. The elevator control device according to claim 7, wherein the conversion function is used to convert each of the control target values for the elevator. 9. The elevator control device according to claim 7, wherein the knowledge base includes a correlation speed table of the sensory target and the control target, and the control target value is determined from the correlation speed table. 10. The elevator control device according to claim 6, further comprising means for setting weights or priorities for each item of the control target. 11. In an elevator control device that performs group management control of a plurality of elevators in service between multiple floors, means for inputting two or more control targets for the elevator, and performing group management control using each control target. Elevator control device with means. 12. The elevator control device according to claim 6 or 11, wherein weighting or priority among each item of the control target is determined using AHP. 13. The elevator control device according to claim 6 or 11, further comprising means for determining a control method for achieving the control target, and executing the group management control based on the control method. 14. The elevator control device according to claim 13, wherein the control method determining means determines the control method based on the control target value of the control target and weighting or priority among each item of the control target. 15. The control method determining means includes means for selecting at least one control method candidate for achieving the control target, and means for determining a predicted achievement value of the control target for the selected control method candidate; 14. The elevator control device according to claim 13, further comprising means for determining a control method based on these predicted achieved values. 16. The elevator control device according to claim 15, wherein the means for determining the predicted achieved value calculates the predicted achieved value by performing a simulation on the selected candidate. 17. The elevator control device according to claim 15, wherein the predicted achieved value is displayed in correspondence with the inputted control target or sensory target. 18. In an elevator control device that performs group management control of a plurality of elevators in service between multiple floors, a registration means is provided to enable selection of an operation method according to the operating status of the elevator or each time period, and An elevator control device for adding, changing, or deleting the contents of the registration means in response to a signal. 19. The elevator control device according to claim 19, wherein the operating method includes at least a call assignment control method and parameters. 20. In an elevator control method that performs group management control of a plurality of elevators in service between multiple floors, the user's sensitivity toward the elevator is input, and from this sensitivity, at least the call assignment evaluation formula and parameters for the group management control are determined. An elevator control method that calculates the evaluation formula and performs the group management control according to the evaluation formula and parameters. 21. The elevator control method according to claim 20, wherein the sensitivity is converted into a control target for the elevator, and the evaluation formula and parameters are calculated using the control target. 22. An elevator control device comprising means for inputting a user's qualitative sensibility regarding a plurality of elevators in service between multiple floors, and a means for converting this sensibility into a plurality of control targets for the elevator. Sensitive input device. 23. The sensibility input device for an elevator control device according to claim 22, further comprising means for determining a control method for achieving the control target. 24. Means for inputting the qualitative sensitivity of users regarding multiple elevators in service between multiple floors and converting it into multiple control targets, and inferring a control method to achieve each control target value. A sensibility input device for an elevator control device, comprising means for determining and means for registering this control method in a control device that performs group management control of the elevator. 25. The sensitivity input device for an elevator control device according to claim 24, further comprising means for transmitting the control method to the elevator control device. 26. The means for inputting the sensitivity and converting it into a plurality of control targets determines the control target value by an inference function using a preset sensitivity input knowledge base and the environment/traffic demand database, and determines the control target value for each item of the control target. 25. The sensory input device for an elevator control device according to claim 24, wherein the sensitivity input device is configured to determine the weighting or priority among the elevators using AHP. 27. The means for inferring and determining the control method determines the control method based on the control target value and the weighting or priority order between each item, using the environment/traffic demand database and the control method determination knowledge base. The sensory input device for an elevator control device according to claim 24, configured as follows. 28. Based on the actual measurement data of the elevator control device, detect the occurrence of new data that does not match the data in the environment/traffic demand database and the control method determination knowledge base, and use this data as the environment/traffic demand database. 28. The sensibility input device for an elevator control device according to claim 27, further comprising knowledge acquisition means for additionally registering each in a database and a control method determination knowledge base. 29. If two or more control methods that achieve the control target value are selected in the control method determination knowledge base, a simulation experiment is performed using a simulator, and based on the data of the experimental results obtained, the control method is determined after the control target is achieved. 28. The sensibility input device for an elevator control device according to claim 27, wherein the sensibility input device is configured to determine a control method with a good control method as the optimum control method. 30. If the actual measurement data does not achieve the control target value, the control method is reselected by the means for inferring and determining the control method using the environment/traffic demand data of the actual measurement data, and the selected control method is reselected. 29. The sensibility input device for an elevator control device according to claim 28, wherein the predicted achievement value is determined using a simulator, and the result is presented to the user for approval and then registered in the elevator control device. 31. A sensitivity input method for an elevator control device in which a user's qualitative sensitivity toward a plurality of elevators operating between multiple floors is input in a predetermined guidance format. 32. The sensitivity input method for an elevator control device according to claim 31, wherein the sensitivity items are displayed on a radar chart, and the sensitivity target value is input using the radar chart. 33. Claim 31: A predetermined pattern is displayed according to the usage environment of the building, and the sensitivity is input according to this pattern.
Sensitivity input method for the elevator control device described. 34. Claim 3, wherein a predetermined range is set for the sensitivity target value for the sensitivity item, and the sensitivity target value is input within the range.
The sensory input device for an elevator control device according to item 1.