JPH04345478A - Group management control device for elevator - Google Patents

Abstract

PURPOSE:To improve service by letting each user input plural requisitions in a form of conversation, carrying out an allocation process based on selected optimum parameters, and thereby indicating the result of response from group management control having been assumed so far. CONSTITUTION:This device composed of a group management control section 1, a learning control section 1-1, single body control sections 2-1 through N, and of respective hall call input/output control sections 4, allows control information required by group management control to be given and/or received through a high speed transmission path 6 at high speeds. Data is given and/or received through a low speed transmission path 7. On input/output device 5-1 with a CRT is connected to the learning control section 1-1, so that optimum parameters are thereby selected in a form of conversation. Each hall call button 3 is therefore controlled by the group management control section 1 through the transmission path 7, when the button 3 is depressed, single body control is so assumed that the optimum applicable cage is determined based on information on the single body control sections 2-1 through N forwarded via the high speed transmission path 6.

Description

[Detailed description of the invention]

[Object of the invention]

0002

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an elevator group management control system for operating a plurality of elevators for a plurality of floors, and more particularly to an elevator group management and control system characterized by hall call assignment control.

0003

[Background Art] In recent years, when multiple elevators are installed in parallel, in order to improve the operating efficiency of the elevators and the service to elevator users, the elevators that respond to hall calls on each floor are controlled by a microcomputer. Currently, small-sized computers such as the above are used to allocate information rationally and quickly. That is, when a hall call occurs, the most suitable elevator to service the hall call is selected and assigned, and group management control is performed so that other elevators are not allowed to respond to the hall call.

[0004] Recently, in this type of group management control device, it has been possible to grasp inter-floor traffic for each building, such as measuring car call registration data when responding to each hall call in real time, and measuring data during boarding and alighting. The building-specific demand is determined based on the measured data and used for hall call allocation control.

Under these circumstances, elevator hall call assignment control evaluates each evaluation index of group management control performance for the occurrence of a hall call, weights and adds each evaluation value, and selects the optimal car as a comprehensive evaluation. This is done by determining the

However, since each evaluation index weighting value varies greatly depending on the purpose of the building and traffic demand, its optimum value differs depending on the building. To this end, multiple relationship models are created between each evaluation index weighted value for allocation control and the group management control response result, which is group management control performance, and a target model that responds to a large number of traffic demands is created by associating the relationship models. Using a man-machine interface that synthesizes and displays group management control response inference results, users of each building can set the optimal control parameters while checking the group management control response inference results in an interactive format. .

[0007]

[Problem to be Solved by the Invention] However, in such conventional elevator group management control devices, allocation processing is performed using optimal parameters set by the user while checking the group management control response inference results in an interactive manner. There was a problem in that it was not possible to accurately grasp the actual group management control response results.

[0008] Also, there are cases where unusual events such as exhibitions are held in the building, and in such cases,
It is necessary to change the items that serve as control target indicators, the user has to input the request immediately beforehand, and after the event has ended, it is necessary to return to the original state by inputting it again. There is no group management control device that can easily handle such problems, and it is difficult to handle customers smoothly when events are held, making it impossible to avoid crowding.

[0009] The present invention was made in view of the above-mentioned conventional problems, and utilizes a man-machine interface function to carry out allocation processing using optimal parameters selected according to a request input by a user, and to It is an object of the present invention to provide an elevator group management control device having a function of collecting management control response results and displaying them after a certain period of time has elapsed.

[0010]Furthermore, in the present invention, input settings for a specific date and time can be made in advance in the man-machine interface function, and when the specific date and time is reached, the assignment is automatically performed using the optimal parameters selected according to the request input by the user. It is an object of the present invention to provide a group management control device for elevators which can perform processing and return to the original optimal parameters to perform allocation processing at the end of a set specific date and time.

[Configuration of the invention]

0012

[Means for Solving the Problems] The specified invention operates a plurality of elevators for a plurality of floors, performs a predetermined evaluation calculation on common hall calls that occur, and selects the most suitable elevator for the hall call. In the elevator group management control device that assigns ambiguously expresses the relationship between the partial model part, the partial system model, and traffic demand using a plurality of membership functions, stores the coupling relationship of these membership functions and the partial system model, and based on the traffic demand, an inference unit that calculates a weighting value for the partial system model; a synthesis unit that calculates a group management control response inference result by combining outputs from the inference unit and the partial model unit; and a group management control response from the synthesis unit. A man-machine interface that displays the inference results and performs operations for optimizing the evaluation index weighting value by evaluating the response inference results in an interactive format, and an optimal parameter selected from the request input through the man-machine interface. The traffic data control unit collects the group management control response results when the allocation process is performed and displays them on the man-machine interface after a certain period of time has elapsed.

[0013] A related invention is a group management of elevators in which a plurality of elevators are put into service for a plurality of floors, a predetermined evaluation calculation is performed on a common hall call that occurs, and the hall call is assigned to the optimum elevator. In the control device,
A partial model section modeled as a set of partial system models composed of function models synthesized by a neural network in which the relationship between each evaluation index weighting value and the group management control response result is vaguely segmented according to traffic demand; vaguely expressing the relationship between the model and traffic demand using a plurality of membership functions, storing these membership functions and the connection relationship of the partial system model, and calculating a weighting value for the partial system model based on the traffic demand. an inference section; a synthesis section that calculates a group management control response inference result by combining the outputs from the inference section and the partial model section; and a synthesis section that displays the group management control response inference result from the synthesis section, and A man-machine interface that performs operations for optimizing evaluation index weighting values by evaluating response inference results, and a specific interface that selects a specific date and time and performs operations for optimizing evaluation index weighting values by limiting that date and time. It is equipped with a date and time control section.

[0014]

[Operation] In the elevator group management control device of the specified invention,
Weighting values for the partial system model are calculated by inputting the traffic demand of the building for each fixed period at each time to the inference section. In addition, control parameter combinations in which the control parameters of each evaluation index are varied within a predetermined range are input to the partial model section, and the group management control response inference results for each control parameter combination are generated by the associative function of the synthesis section together with the weighting values. Output.

[0015] Then, in the man-machine interface, the above-mentioned inference results are displayed, and evaluation is performed based on evaluation criteria for determining "optimum" through interaction with the operator, especially the user. The inference result is selected, and the evaluation index weighting value corresponding to this inference result is selected as the optimal control parameter to perform hall call allocation control.

Furthermore, the traffic data control unit uses the man-machine interface to perform assignment control using optimal parameters selected in an interactive manner, outputs a command to collect the group management control response results, and outputs a command to collect the group management control response results at a certain period of time. The collected group management control response results are output as specific time traffic data,
To enable checking whether selected control parameters are actually optimal.

Furthermore, in the elevator group management control device of the related invention, the weighting value for the partial system model is calculated by inputting the traffic demand of the building for each fixed period at each time to the inference section. In addition, control parameter combinations in which the control parameters of each evaluation index are varied within a predetermined range are input to the partial model section, and the group management control response inference results for each control parameter combination are generated by the associative function of the synthesis section together with the weighting values. Output.

[0018]The man-machine interface then displays the above inference results and performs an evaluation based on evaluation criteria that determines ``optimal'' through interaction with the operator, especially the user. The inference result is selected, and the evaluation index weighting value corresponding to this inference result is selected as the optimal control parameter to perform hall call allocation control. In addition, in the man-machine interface, when selecting the optimal control parameters, it is possible to select the optimal control parameters by limiting a specific date and time, and transmit the limited date and time and the optimal control parameters at that time to the specific date and time control section. .

Furthermore, the specific date and time control section stores the limited specific date and time and the optimal control parameters at that time, which are transmitted from the man-machine interface, and when the limited specific date and time arrive, the learning control section selects the optimal control parameters. It is transmitted as a control parameter to control hall call allocation on a specific date and time. Then, at the end of the specific date and time, the control parameters are returned to the original optimum control parameters before the specific date and time, and normal hall call allocation control is performed.

[0020] In this way, on the day and time when a special event is held, the control is automatically switched to the specific date and time control and the elevator group management control is performed, thereby alleviating the occurrence of congestion and improving the service.

[0021]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Examples of the present invention will be explained in detail below with reference to the drawings.

FIG. 1 shows the overall system configuration of an elevator group management control device according to an embodiment of the specific invention.
1 is a group management control unit, and this group management control unit 1 is connected to single control units 2-1 to 2-N that control each unit elevator, a learning control unit 1-1, and a high-speed transmission control unit which is a first transmission control means. They are connected via a transmission line 6. This group management control unit 1,
The learning control section 1-1 and the single control sections 2-1 to 2-N are constituted by a small computer such as one or more microcomputers, and operate under the control of software. 3 is the hall call button provided on each floor,
Reference numeral 4 denotes a hall call input/output control section for inputting and outputting hall calls.

[0023] Then, the group management control section 1 and the learning control section 1-
1. The individual control units 2-1 to 2-N and each hall call input/output control unit 4 are connected via a low-speed transmission line 7, which is a second transmission control means.

[0024] The high-speed transmission line 6 connects the single control units 2-1 to 2-
This is a transmission control system that performs transmission between N, group management control section 1, and learning control section 1-1, that is, mainly between the control computers in the machine room, and is connected by a high-speed, highly intelligent network. Control information necessary for control is transmitted to the group management control unit 1, learning control unit 1-1, and each individual control unit 2-1 to 2-N.
There is a high speed transfer between the two.

The low-speed transmission line 7 is a control system that transmits information mainly to the hall call buttons 3 of each hall, the monitoring panel 5 of the building management room, etc. via the hoistway, and the high-speed transmission line 6 It is slow compared to
It is connected to the learning control section 1-1 and the single control sections 2-1 to 2-N, and exchanges data.

Therefore, when the group management control section 1 is normal, the hall call button 3 is controlled by the group management control section 1 via the hall call input/output control section 4 and the low-speed transmission line 7, and the hall call button 3 is When pressed, the hall call gate is closed, the registration lamp is set, and the optimum machine number is determined based on the information sent from the unit control units 2-1 to 2-N via the high-speed transmission line 6, and the unit number is set. A control command is given to a corresponding one of the control sections 2-1 to 2-N. Then, the unit control unit that receives the control command accepts this control command as hall call information and performs unit control.

CR as a man-machine interface
The T input/output device 5-1 is installed in a place where the user can operate it, such as a building management room, and is connected to the learning control unit 1-1.
1, allowing the user to select the optimal parameters in an interactive manner.

FIG. 2 shows the group management control unit 1 in this embodiment.
An example of the software system of the unit control units 2-1 to 2-N is shown, and this software system has a unit control function task 9, a group management control main function task 1, and a real-time OS 8 which is an operating system.
0, Group management control subfunction task 11, Transmission control task 1
2, and each task 9 to 12 is activated or held by a scheduler within the real-time OS 8.

Of these tasks 9 to 12, the single control function task 9 is a functional task for operating the single control units 2-1 to 2-N, and is given a high priority.

The group management control main function task 10 is a central function task of the group management control unit 1, and is a function task that is the central function task of the group management control unit 1. The system collects information data for each machine, performs comparative calculations, determines the optimal matching machine, issues control commands to the corresponding machine, and controls hall call registration.

The group management control sub-function task 11 is a functional task that processes information for each machine in the group management control section 1, and processes the information under the control of the group management control main function task 10. That is, the configuration is such that a computer having a group management control main function task 10 manages the startup and termination of tasks via a high-speed transmission line 6, and the group management control main function task 10 is the master.
Distributed processing is performed on a unit-by-machine basis in response to instructions from the controller, and data is transferred to the group management control main function task 10 upon completion of the processing.

The transmission control task 12 sends and receives data on the high-speed transmission line 6, starts the group management control subfunction task 11,
Controls termination.

FIG. 3 is a block diagram showing the system configuration of the high-speed transmission line 6 in FIG. 1. Transmission control is performed using the microprocessor 13.
Data transmission is performed in a highly intelligent manner using a data link controller 14 and a media access controller 15, which are configured with hardware as parts that control the data link layer of the LAN network model layer proposed by the International Organization for Standardization. The configuration is such that the ratio of transmission control software managed by the microprocessor 13 to high-speed transmission control is reduced.

For example, as the data link controller 14 which is a controller for realizing the above-mentioned highly intelligent transmission control, Intel Corporation's L
The i82586, which is an SI, and the Intel i82501, etc., have also been put into practical use as the media access controller 15, but by using these,
A high-speed transmission function such as 0 Mbit/sec can be performed relatively easily with a reduced support ratio of the microprocessor 13.

In FIG. 3, 16 is a system bus, 17 is a control line, and 18 is a serial transmission system.

FIG. 4 shows the learning control section 1 in this embodiment.
5 is a functional block diagram showing the flow of input and output signals of the learning control section 1 and the traffic data control section 5-2 of FIG.
6 and 7 show detailed system configurations of the inference section and partial model section of FIG. 5, respectively.

In FIG. 4, the group management control unit 1 is connected to the individual control units 2-1 to 2 of the elevator group system 2, as described above.
- Coordinates with the group management control subfunction task 11 in N and executes the hall call allocation control function.

The evaluation algorithm used for this hall call allocation control evaluates each evaluation index on group management performance,
A comprehensive evaluation is performed by adding each evaluation value with an optimal weight.

The learning control unit 1-1 transmits the optimum evaluation index weighting value to the group management control unit 1 as a control parameter for each fixed period at each time.

The group management control section 1 has individual control sections 2-1 to 2-2.
The evaluation value of each evaluation index is calculated based on the information from -N, optimal weighting is performed using the control parameters, and a control command is given to the optimal machine.

The learning control unit 1-1 periodically updates the elevator group system 2, that is, the individual control units 2-1 to 2.
-N and the information from the hall call input/output control unit 4, the group management control response results are input and used as base data for online learning.

Next, the detailed operation of this learning control section 1-1 will be explained using the drawings from FIG. 5 onwards.

First, the functional configuration of the learning control section 1-1 will be explained. As shown in FIG.
2, a synthesis section 23, and an inference result evaluation section 24.

Evaluation calculations in the group management control unit 1 are generally performed on a plurality of evaluation indicators q, and are expressed as g1 (i), g2 (i), . . . , gq (i) for the i-th car. The overall evaluation value is obtained by weighted addition of the assigned evaluation values for each evaluation, and the overall evaluation value Ei of the i-th machine is calculated.
teeth,

[0045]

[Math 1]

It is expressed as follows.

[0047] Here, αj is a weighting value for each evaluation index j.
control parameters sent to

Therefore, the inference result evaluation section 24 calculates the traffic demand for each fixed period for each time zone and outputs it to the inference section 21, and also generates combinations of control parameters αj within a predetermined range, and partially The group management control response inference results corresponding to the combinations of control parameters αj are output to the model unit 22, and the group management control response inference results corresponding to the combinations of control parameters αj are evaluated in accordance with the user's needs.The response inference results are output to the user via the input/output device 5-1. and set the optimal control parameters according to the user's selection.

When the group management control response result is y and the input is u, the calculations in the inference section 21, partial model section 22, and synthesis section 23 can be expressed as y=F(u) (1). In addition, here, y=(y1,y2,...,yn)T u=(us
, uq )T = (C, α)T.

In this group management control response result y, y1
, y2 , ..., yn represent the occurrence rate of hall call response time in a certain period of time, the average ridership rate, the average service time, etc., and serve as evaluation standard data for determining group management performance.

[0051] Furthermore, C in the input u represents the traffic demand, and if C=(c1, c2, c3), then c1, c2, c3 are the total average passenger generation interval [s/person] and the average passenger generation at the standard floor, respectively. This data represents the average interval between passenger arrivals heading to the standard floor, and represents the system status such as the degree of system congestion and the flow of people.

Further, α is a weighting value (control parameter) for each evaluation index, and as described above, a plurality of evaluation indexes q
is expressed as α=(α1, α2, ..., αq).

Therefore, the inference section 21, the partial model section 22,
The target model consisting of the synthesis unit 23 is expressed by the composition of m partial system models fi (α), (i = 1, 2, ..., m), and equation (1) can be rewritten as the following equation. become.

[0054]

[Math 2]

Here, ai (C) indicates the activity level of the partial system model fi (α) in the traffic demand C, and the system status obtained from the traffic demand C and the partial model part 2
It is determined by the coupling relationship with the partial system model in 2.

Next, the system configuration of each part of the inference section 21, partial model section 22, and inference result evaluation section 24 will be explained.

As shown in FIG. 5, the inference unit 21 receives the traffic demand C from the inference result evaluation unit 24, and calculates the ai of equation (2) that expresses the degree of activity based on the system status obtained from the traffic demand C.
(C), (i=1, 2, ..., m).

As shown in FIG. 6, the inference section 21 is composed of an input section 21-1, a storage section 21-2, an output section 21-3, and a gate 21-4.

The input unit 21-1 has a k-dimensional state vector V consisting of k neurons, and by passing the input traffic demand C through a membership function φi, each element is configured by its membership grade. The partial input vector ci, (i=1, 2, . . . , M) is output. These M partial input vectors ci are collectively input as an input vector C to the input state vector V.

The storage unit 21-2 is composed of an r-dimensional state vector X consisting of r neurons, and the input unit 21
This corresponds to a storage unit that associates -1 with the output unit 21-3.

The output unit 21-3 consists of an m-dimensional state vector Z consisting of m neurons, and each element Zi,
(i=1, 2, . . . , m) corresponds to the partial system model fi (α) of the partial model section 22.

[0062] Between the input section 21-1 and the storage section 21-2,
Further, there is a mutual loop between the storage section 21-2 and the output section 21-3, and each section 21-1 to 21-3 has a self-loop.

This relationship is in discrete form and can be expressed as follows.

C(k)=φ(u(k))
…(
3.1) V(k+1)=ψ(WVC・C(k)+W
VV・V(k)+WVX・X(k))


...(3.2) X(k+1)=ψ(WXV・V(
k+1)+WXX・X(k)+WXZ・Z(k))


...(3.3) Z(k+1)=ψ(
WZX・X(k+1)+WZZ・Z(k))...
(3.4) V(0)=V0, X(0)=X0,
Z(0)=Z0,k≧0Here, WVC is vector C
is a matrix representing the load from to vector V,
This is the synaptic load of the neurons forming the vector V with respect to the vector C. Also, WVV, WVX, WXV,
The same applies to WXX, WZX, and WZZ.

[0065] Also, φ is a j-dimensional membership function, ψ is a sigmoid function corresponding to each dimension, and for each input element,

[0066]

[Math 3]

The following calculation is performed.

Further, k is a parameter representing time, and each time k is increased by 1, a unit time elapses.

Demand C(u(k)) input by appropriately setting each W in the above equations (3.1) to (3.4)
subsystem model fi (α), (i=1,
2, . . . , m) appears in the state vector Z of the output section 21-3 over time.

The gate 21-4 is opened when the set time T has elapsed, and the gate 21-4 is opened to convert Zi(T) into the partial system model fi.
The activity level ai of (α) is output as (C).

Next, the partial model unit 22 has the system configuration shown in FIG. 7, and by inputting the control parameter α, the group management control response result fi is calculated for each partial system model.
It has the function of outputting (α).

The individual partial system models fi (α), (i=1, 2, . . . , m) in the partial model section 22 are composed of multilayer neural networks as shown in FIG. Each corresponds to a specific demand Ci, and each stores a control parameter α for that demand and input/output data i of the group management control response result of the actual system. And the partial system model f
i (α) is learned using the back propagation method using this input/output data i as teacher data.

As shown in detail in FIG. 8, each subsystem model has y(k) when input u is given.
=fi (u(k)) is executed. This calculation process is as follows: net(k)=Whu(k)・u(k)
…(4.
1) h(k)=ψ( net(k)+θh(k
)) …(4.2
) nety(k)=Wyh(k)・h(k)+W
yu(k)・u(k)...(4.3) y(k)=
ψ(nety(k)+θy(k))
...(4.4) k≧0 Here, Whu, Wyh, and Wyu are matrices representing synaptic loads. Further, θh and θy are vectors representing bias values for the intermediate layer h and the output layer y, respectively.

Each partial system model fi (u), (i
=1, 2, . . . , m) have different synaptic loads and biases, and perform calculations.

In the synthesis unit 23 shown in FIG. 5, the partial system model f1 (α
), f2 (α), ..., fm (α), and inference part 2
Activities a1 (C), a2 (C), ..., am (C) for each partial system model input from 1 are expressed as (
2) Synthesize according to the formula, group management control response inference result y
It is output to the inference result evaluation unit 24 as

The inference result evaluation unit 24 includes the inference unit 21 and the partial model unit 2 as described above, as will be explained in detail later.
2. For the target model consisting of the synthesis unit 23, input u(
= (C, α)T ), by generating combinations of control parameters α within a predetermined range for the traffic demand of the actual system and giving them as input u, it is possible to correspond to each combination of control parameters α as a result. The group management control response result is evaluated in an interactive manner, an optimal control parameter α is set, and the result is sent to the group management control unit 1.

Next, the inference unit 2 uses the group management control response results from the elevator group system 2 obtained online.
1. Online learning of the partial model section 22 will be explained.

Regarding learning, based on the difference between the inference result and the response result from the elevator group system 2, the inference unit 2
1 “certainty” and related partial system model 22-
Modify the value fi of i. The amount of correction is made in proportion to the magnitude and activity of the difference, and inversely proportional to its certainty.

The loop for correcting the certainty in the inference unit 21 shown in FIG.
loop to the matrix WXZ. Further, it is assumed that the matrix WXZ is square, that is, r=m. At this time, the (i, j) element wij of the matrix WXZ
, wii=pi , (i=1,2,...,m)...(
5.1) wji=-pi, (j≠i)
...Modify as shown in (5.2).

Here, pi≧0 means the partial system model f
This is a parameter representing the reliability of memory for i (α), and is calculated by the following equation.

[0081] pi =ξ・Ri (k+1)+ζ
…(
6.1) Ri (k+1)=1-exp[-β(N
i (k+1)+γ)] …(6.2) Ni (
k+1)=Ni (k)+δNi
...(6.3)

008
2]

[Math 4]

Here, η, β, γ, ξ, and ζ are constants, and Ri and Ni are the proficiency level and learning progress of the partial system model fi (α), respectively. and,
This learning progress Ni (k) is determined by the partial system model fi
It represents the learning progress after learning (α) k times, and is proportional to the activation level ai, and the current proficiency level Ri (k)
It changes by a degree δNi (maximum 1) that is inversely proportional to . Then, the new learning progress Ni (
k+1), the new proficiency Ri (k+1) is (
6.2) The final revised certainty p determined by the formula
i is determined by equation (6.1).

On the other hand, the partial model portion 22 is modified by the back propagation method. At this time, the input/output data held by each partial system model is rewritten. When a certain period of time for each predetermined time period ends, a group management control response result is calculated based on the response result for the time period, and input/output data Do = (uo
, yo ), uo = (Co*, αo )T is generated.

At this time, the input/output data (D1, D2, . . . , DL) held by the partial system model is rewritten.

Scan all input/output data and calculate the square of the distance between α and αo dα=|α−αo |2
...(7.1) is obtained, and the two data D (1st), D are ordered from the nearest one.
(2nd) Extract.

These two data D(1st) and D(2
nd), the square of the distance between y and yo dy=|y-yo |2
...(7.2) is obtained, and then D(1st) and D(2nd) are corrected by the following formula.

y(1st) = ρ1 ・yo + (1−ρ1)
・y(1st) ...(7.3) y(2nd
) = ρ2 ・yo + (1-ρ2 ) ・y (2nd)
...(7.4)

[0089]

[Math 5]

The input/output data in the partial system model 22-i is rewritten according to the above equations (7.3) and (7.4), and the partial system model is created by the back propagation method using the rewritten input/output data as training data. Modify the load matrix using the following steps.

δy (k)=ψ′(nety(k)+θy(k))
*(y*(k)−y(k))

…(8
．． 1) δh (k)=ψ′( neth(k)+θ
h (k)) *W´yh・δy (k)


...(8.2) ΔWyh (k+1) = ηyh(k)・δy (k)・h′(k)
+αyh・ΔWyh(k)

…(8.3
) ΔWhu (k+1) = ηhu(k)・δh (k)・u′(k)
+αhu・ΔWhu(k)

…(8.4
) ΔWyu (k+1) = ηyu (k)・δy (k)・u′(k)
+αyu・ΔWyu(k)

…(8.5
) Wyh(k+1)=Wyh(k)+ΔWyh(k
+1) …(8.6) Wh
u(k+1)=Whu(k)+ΔWhu(k+1)
...(8.7) Wyu(k+1
)=Wyu(k)+ΔWyu(k+1)
...(8.8) ΔWyh(0)=0,
ΔWhu(0)=0, ΔWyu(0)=0, Δ
θy (0)=0, Δθh (0)=0 Here, y* is the teaching data,

[0092]

[Math 6]

k until holds true for all y*
Increment and repeat. Further, A*B represents the product of each element of matrices A and B, and η and α are learning parameters.

By the above calculation, the load matrix of the partial system model is corrected and re-learned every time the group management control response result is input.

Next, the detailed configuration of the inference result evaluation section 24 and the optimization setting operation of the control parameter α will be explained with reference to FIG. The inference result evaluation unit 24 includes a control parameter combination generation unit 24-1 and a traffic demand detection unit 24-2.
, an inference result evaluation parameter setting section 24-3, an inference result evaluation calculation section 24-4, and a control parameter setting section 24-
It consists of 5.

As described above, the inference unit 2 of the learning control unit 1-1
1. The partial model unit 22 and the synthesis unit 23 configure the relationship between the control parameters and the group management control response inference results for arbitrary traffic demand for each building, and calculate the group management control response result y using equation (1). It was possible to estimate by association.

Therefore, the inference result evaluation section 24 sets the optimum value of the control parameter based on the group management control response result y with respect to the "optimal" standard suitable for each building in accordance with the customer's needs. For this purpose, the inference result evaluation section 24 detects traffic demand in a predetermined time period and outputs it to the inference section 21, and also allows the user to interact with the input/output device 5-1.
The optimum value of the control parameter is determined based on the inference result evaluation parameter set by .

[0098] The inference result evaluation parameter here is a parameter table for evaluating the group management control response result y, and is set for each building according to traffic demand. Further, the group management control response result y is an index representing the group management control performance, and is evaluation standard data such as the occurrence rate of hall call response time, the average ridership rate, and the average service time as described above.

Next, the operation of the traffic data control section 5-2 will be explained.

First, FIG. 10 shows an example of the input screen of the input/output device 5-1. The user inputs a request in an interactive format, as shown in FIG. This input value is converted into an evaluation formula for selecting optimal parameters, and the group management control response inference result selected by this evaluation formula is displayed.

FIGS. 11 and 12 show examples of converting user input. That is, FIGS. 10(a) to (c
) is converted into a numerical value with the entire value being 10 (in this example, it is converted into a numerical value of "7"), and then (a) ~
The values of the three items in (c) are normalized to a total of 10. That is, as shown in FIG. 11, for example, (a
) is input as an intermediate value between "neutral" and "very important", the input value is digitized as 7, and then for input values a, b, and c,

[0102]

[Math 7]

The normalization formula is used to normalize the total number to 10.

Next, the coefficients k1, k2, k3 are determined by referring to the related table shown in FIG. 12 showing the relationship between the input items, evaluation index weight items, and coefficients, and the evaluation formula for selecting the optimal parameter is calculated. E(i) is created as follows.

E(i)=t・k1 +sd・k2+m・
k3 =2t+3sd+5m The group management control response inference result selected by the evaluation formula created in such a procedure is displayed on the input/output device 5-1. So when the user sees this,
Check to see if it meets your requirements.

Next, in the traffic data control section 5-2,
As shown in the flowchart No. 3, first, the input/output device 5-1 of the traffic data control unit 5-2 sends out the time for performing the allocation process using the optimal parameters selected by request input (
Steps S1 to S3), t seconds before the collection start time, the learning data of the said time is received from the learning control unit 1-1 and saved (steps S4, S5), and after the collection start time,
New group management control results are collected (step S7).

[0107] For a certain period of time, for example, if the user's request is for a time period of 10:00 am to 12:00 am, then a certain period of time from 10:00 to 12:00 (here, one week) After the collection is completed (step S8), the traffic data control unit 5 collects the previously saved learning data and the specific time traffic data collected during the time period from 10:00 am to 12:00 am for one week. -2, and when the traffic data control unit 5-2 receives the specific time traffic data, it sends two types of data to the input/output device 5-1 as a man-machine interface, that is, the optimal data selected according to the user's request. The learning data and specific time traffic data for the time period in which the parameter allocation process is to be performed are transmitted (step S9).

[0108] Then, when the learning data and the specific time traffic data are transmitted to the input/output device 5-1,
As shown in a), a graph display 5-1a is output at the bottom right of the screen to notify the user of the presence of data. The user can view traffic data at a specific time by selecting the graph display 5-1a with the mouse. In addition, in this example, as shown in FIG. 14(a), an example was shown in which the waiting time distribution is compared with learning data and specific time traffic data.

[0109] This allows the user to compare data while looking at the displayed learning data and traffic data at a specific time, and to see what group management control response results will actually result from his or her request. This makes it possible to understand whether the

[0110] This specific invention is not limited to the above embodiment, and as shown in Fig. 14, the actually measured traffic data can display not only waiting time but also service time, average value, maximum value, etc. Furthermore, it may be in a tabular format instead of a graphical format.

FIGS. 15 and 16 show the overall system configuration of an embodiment of the related invention. In addition, Figures 1 to 14
The parts common to the embodiments of the specific invention shown in FIG.

[0112] The input/output device 5-3 with CRT, which serves as a man-machine interface where the user can input requests in an interactive format, can specify a specific date and time, and the optimal input for that specific date and time can be made in an interactive format. A parameter is selected, and the specific date and time and optimum parameter data are transmitted to the specific date and time control section 5-4. In the specific date and time control unit 5-4,
Save the specific date and time and optimal parameter data,
At a specific date and time, the stored optimal parameter data is transmitted to the learning control section 1-1, and the optimal parameters are further transmitted to the group management control section 1 to perform call allocation control.

The data outputted from the input/output device 5-3 includes a specific date and time and optimal parameters determined by the user for that date and time, and is sent to the specific date and time control section 5-4.

[0114] The specific date and time control unit 5-4 manages the date and time, constantly confirms whether it is the specific date and time specified by the user, and at the time the specific date and time arrives, performs learning control on the optimal parameters determined by the user. 1-1. The learning control section 1-1 then transmits the optimal control parameters sent from the specific date and time control section 5-4 to the group management control section 1.

[0115] When the specific date and time ends, the specific date and time control unit 5-4 again outputs a signal indicating the end to the learning control unit 1-1, and the learning control unit 1-1 determines that the specific date and time has ended. After receiving the signal, the original optimum control parameters before the specific time are transmitted to the group management control section 1. In other words, the learning control section 1-1 performs allocation control with priority given to the optimum control parameters only when there is an interrupt command from the specific date and time control section 5-4.

The above operation will be explained in more detail based on the flowchart of FIG.

FIGS. 17 and 18 are examples of specifying a specific date and time using the man-machine interface 5-3.
First, select a specific date. In this embodiment, a calendar-like screen is displayed, and a specific day (the 11th) is selected using the mouse to specify the date. Further, in the embodiment, only one day is designated, but even if the schedule spans several days, the entire schedule can be selected using the mouse. and,
In the figure, buttons with several functions are provided on the screen, and the buttons are used to switch to the next screen and proceed with the operation. Next, on this screen, select event settings and
Switch to screen 8. Here, the user selects a time period on the date of the event for which the user would like to enter and set the request (steps S11 to S14).

[0118] When this setting is completed, select the registration button on the right side with the mouse, determine the optimal control parameters for the specific day of the event in an interactive manner according to the procedure explained in the first embodiment, and set the specific date and time control section. 5-4 (steps S15 to S19).

Next, the flowchart in FIG. 20 shows the operation of the specific date and time control unit 5-4, and when data is transmitted from the man-machine interface 5-3, the data is processed as shown in FIG. It is saved in the specific date and time table T (steps S21 to S23).

[0120] Then, when the designated date arrives, the stored data is sent to the learning control section 1-1. That is, while managing the date and time, the timing of the specific date and time specified by the user is detected (steps S24, S2
5) When the designated day arrives, the optimal control parameters corresponding to the designated day are extracted from the specific day table T shown in FIG. 21 and transmitted to the learning control unit 1-1 (step S26,
S27).

By performing such control, when an event is held, it becomes possible to perform control using the specified control parameters.

[0122] This related invention is not limited to the above-described embodiments, and if an event is held regularly every week, the day of the week may be specified.
For example, it may be possible to set Tuesdays to be controlled using different control parameters from other days of the week. Further, for example, specific control parameters may be set by selecting specific days of the week only during the first two weeks of the month.

[0123]

[Effects of the Invention] As described above, according to the specific invention, the user inputs multiple requests in an interactive format, executes the allocation process using the selected optimal parameters, and displays the group management control response results during this process. This allows users to understand the actual group management control response results to their own inputs, allowing them to judge for themselves whether the optimal control parameters are appropriate, and to make adjustments to improve services. .

[0124] Furthermore, according to the related invention, since it is possible to input a request at a certain limited specific date and time, even if the elevator demand situation is significantly different from usual due to an event such as an exhibition, By specifying the specific date and time and setting special control parameters, it will automatically respond, and when the event ends, it will be possible to return to the normal control parameters and control the special elevator. Customers can be handled smoothly to meet demand without crowding, and service can be improved.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an embodiment of a specific invention.

FIG. 2 is a configuration diagram of a software system of a group management control section and a single control section in the above embodiment.

FIG. 3 is a system configuration diagram of a high-speed transmission line in the above embodiment.

FIG. 4 is a functional block diagram showing the flow of input and output signals of the learning control section in the above embodiment.

FIG. 5 is a system configuration diagram of a learning control section in the above embodiment.

FIG. 6 is a system configuration diagram of an inference section in the above embodiment.

FIG. 7 is a system configuration diagram of a partial model section in the above embodiment.

FIG. 8 is a configuration diagram of each partial system model of the partial model section.

FIG. 9 is a system configuration diagram of an inference result evaluation section in the above embodiment.

FIG. 10 is an explanatory diagram showing an example of a user input screen by the input/output device in the above embodiment.

FIG. 11 is an explanatory diagram illustrating the operation of the traffic data control unit in the above embodiment.

FIG. 12 is an explanatory diagram showing a relation table between user input items and evaluation index weight items in the above embodiment.

FIG. 13 is a flowchart showing the operation of the traffic data control unit in the above embodiment.

FIG. 14 is an explanatory diagram of an example of displaying traffic data by the traffic data control unit in the above embodiment.

FIG. 15 is a block diagram of an embodiment of a related invention.

FIG. 16 is a functional block diagram showing the flow of input and output signals of the learning control section in the above embodiment.

FIG. 17 is an explanatory diagram showing an example of an input screen of the input/output device in the above embodiment.

FIG. 18 is an explanatory diagram showing a next screen of an example of the input screen of the input/output device in the above embodiment.

FIG. 19 is a flowchart showing a specific date and time setting operation in the above embodiment.

FIG. 20 is a flowchart showing the operation of the specific date and time control section in the above embodiment.

FIG. 21 is an explanatory diagram of a specific date and time table created by the specific date and time control section.

[Explanation of symbols]

1...Group management control unit 1-1...Learning control unit 2...Elevator group system 2-1 to 2-N...Single control unit 3...Hall call button 4...Hall call input/output control section 5…Monitoring board 5-1...I/O device 5-2...Traffic data control unit 5-3...I/O device 5-4...Specific date and time control section 6...High-speed transmission line 7...Low speed transmission line 21... Reasoning part 22...Partial model part 23...Synthesis section 24...Inference result evaluation section

Claims (2)

[Claims] 1. An elevator group management control device that operates a plurality of elevators for a plurality of floors, performs a predetermined evaluation calculation on a common hall call that occurs, and allocates the hall call to the optimum elevator. a partial model unit modeled as a set of partial system models composed of function models synthesized by a neural network in which the relationship between each evaluation index weighting value and the group management control response result is vaguely segmented according to traffic demand; The relationship between the partial system model and traffic demand is vaguely expressed by a plurality of membership functions, these membership functions and the connection relationship of the partial system model are stored, and the weighting value for the partial system model is determined based on the traffic demand. an inference unit that performs calculations; a synthesis unit that calculates a group management control response inference result by combining the outputs from the inference unit and the partial model unit; and a synthesis unit that displays the group management control response inference result from this synthesis unit, and A man-machine interface that performs an operation for optimizing the evaluation index weighting value by evaluating the response inference results, and a group when assignment processing is performed using optimal parameters selected from requests inputted at the man-machine interface. A group management control device for elevators, comprising a traffic data control unit that collects management control response results and displays them on the man-machine interface after a preset period of time has elapsed. 2. An elevator group management control device that operates a plurality of elevators for a plurality of floors, performs a predetermined evaluation calculation on a common hall call that occurs, and allocates the hall call to the optimum elevator. a partial model section modeled as a set of partial system models composed of function models synthesized by a neural network in which the relationship between each evaluation index weighting value and the group management control response result is vaguely segmented according to traffic demand; The relationship between the partial system model and traffic demand is vaguely expressed by a plurality of membership functions, these membership functions and the connection relationship of the partial system model are stored, and the weighting value for the partial system model is determined based on the traffic demand. an inference unit that performs calculations; a synthesis unit that calculates a group management control response inference result by combining the outputs from the inference unit and the partial model unit; and a synthesis unit that displays the group management control response inference result from this synthesis unit, and A man-machine interface that performs an operation for optimizing the evaluation index weight value by evaluating the response inference result, and a man-machine interface that performs an operation for optimizing the evaluation index weight value by selecting a specific date and time and limiting the date and time. A group management control device for an elevator, comprising a specific date and time control unit.