JP3052768B2 - Elevator group control device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an elevator group management and control system for allocating calls using a neural network, and more particularly to an elevator capable of providing a neural network with high performance in all traffic conditions. The present invention relates to a group management control device.

[0002]

2. Description of the Related Art Conventionally, group management control of elevators has mainly been a call allocation method using an evaluation function or a call allocation control using an expert system using fuzzy logic. A new method of assigning calls using a neural network has been proposed.

A neural network is a network imitating the human brain. A plurality of neural cell models (neurons) are connected in a complicated manner, and the operation of each neuron and the connection form between the neurons are determined in a well-defined manner. The function and the knowledge processing function can be embedded. For example, "Nikkei Electronics" 1987
P115-P124 and 1 of August 10, issue (No. 427)
It is disclosed in the book "PDP model" published by Sangyo Tosho Co., Ltd. in February 989, and in particular, those in which neurons are arranged in a hierarchical structure can use an autonomous learning algorithm called "back propagation". There are features.

When this neural network is used, there is no need for a human to consider an assignment algorithm at all, and a decision system for deciding an optimum assigned car in response to various traffic conditions can be automatically generated. For example, Japanese Patent Application Laid-Open No. 1-275381 "Elevator group management control device" and Japanese Patent Application Laid-Open No. 3-31173 "Elevator group management control device" Japanese Patent Application No. 5-2438
No. 17, "Method of Learning Neural Network for Elevator Call Assignment" and the like.

FIG. 3 shows an example of a neural network for call assignment. As shown in FIG. 3, the neural network NN for call assignment includes an input layer NR1 corresponding to an input pattern (elevator system state data), an output layer NR3 corresponding to an output pattern (assignment suitability),
It is composed of neurons of an intermediate layer NR2 located between the input layer and the output layer.

The input pattern uses various data representing the state of the elevator system (floor and direction of landing call, position and driving direction of each car, car call, load state, etc.) as parameters necessary for call assignment. , Into a form that can be input to a neural network, and the number of neurons in the input layer corresponds to the total number of parameters.

When input data is given to each neuron in the input layer, a signal is sequentially transmitted to the output layer, and as a result, a certain value is output from each neuron in the output layer. In the output layer, there are neurons for the number of elevators, and for various input patterns, the neuron corresponding to the car that is optimal for allocation outputs “1” (maximum value), and the other neurons output “0”. Is learned in advance.

Accordingly, among the values of each neuron in the output pattern, a neuron that outputs a value closest to “1” (maximum value) indicates that it is optimal for allocation.
The unit corresponding to this neuron is selected as the assigned unit. Note that the number of neurons in the intermediate layer (one layer in the embodiment, but may be plural) is appropriately determined according to the number of elevators, the properties of the building, and the like.

Although not shown, a connection weight (synapse weight) indicating the strength of the connection between the neurons is set between the neurons. The connection weight is initially set to an appropriate value, but can be modified so that a more accurate call assignment can be performed using a learning algorithm called “back propagation”.

Since the back propagation is well known and will not be described in detail, a learning sample (an input pattern and a desired output pattern corresponding to the input pattern, ie, a teacher signal, are paired). An algorithm that compares the output pattern for the same input pattern with the teacher signal and corrects the connection weights to minimize the error. Initially, all weights are initialized (for example, random The input pattern of the learning sample is given to each neuron of the input layer. The output pattern at this time and the output pattern of the learning sample (teacher signal)
Then, using the difference (error), the values of the connection weights are sequentially corrected from the output layer side so that the difference becomes smaller.

If this is repeated using a large number of learning samples until the error converges, the call assignment function at the same level as the teacher signal is automatically embedded in the neural network, and the learning input The same level of call assignment as the teacher signal can be performed not only for patterns but also for unknown input patterns.

[0012]

Learning of the neural network for assignment is performed using a learning sample created in advance by simulation or the like or a learning sample created during operation of an elevator at an actual site.
Since the neural network needs to perform learning until the error converges sufficiently, if the learning is performed in a state where the number of learning samples does not reach a certain number, the error does not reach convergence, and a neural network with high performance You can't expect to create a net.

For this reason, a large number of learning samples are created,
It is conceivable that learning is performed when the number reaches a predetermined number, but even if the number reaches the predetermined number, if the learning sample is biased only to those related to a specific traffic situation, the neural network after learning will be There is a problem that it can exhibit high performance in that traffic condition, but cannot exhibit sufficient performance in other traffic conditions.

For example, when learning is performed by using a large number of learning samples created with data at the time of the UP peak, the performance can be sufficiently exhibited at the time of the UP peak, but the other traffic at the time of the DOWN peak, etc. In some situations, performance may be reduced. In particular, in the case where a learning sample is created and re-learning is repeated while the elevator is operating so that it can respond to changes in traffic conditions after the elevator is installed, traffic conditions cannot be controlled at the actual site. There is a problem that the traffic conditions of the created learning samples are biased, and as a result, sufficient performance cannot be exhibited in the traffic conditions where the learning samples are insufficient.

The present invention has been made in view of such a problem, and a group management system capable of creating a neural network exhibiting high performance not only in a specific traffic condition but also in all traffic conditions. The purpose is to provide.

[0016]

According to the present invention, there is provided a learning sample creating means for creating a learning sample, a traffic condition determining means for determining a traffic situation at the time of creating the learning sample, A learning sample accumulating means for classifying and accumulating learning samples according to traffic conditions according to the discrimination result, and a predetermined ratio according to a required property of the neural network, wherein the learning samples accumulated for each traffic condition are stored And learning means for learning the neural network by using the selected learning sample.

[0017]

In the present invention, when a learning sample is created, it is classified and stored according to traffic conditions. Then, a predetermined number of learning samples are selected from the stored data at a predetermined rate for each traffic condition, and learning of the neural network is performed.

[0018]

FIG. 1 is a block diagram showing the overall configuration of the present invention. In the drawing, reference numeral 1 denotes a hall call button provided on each floor (only one floor is shown, and other floors are shown). (Abbreviated), 2 is a hall call signal, A1 is an operation control device for managing the operation of the first car, similarly A2 to An are operation control devices for the second to n car, and 3 is the state of each car (car The car information signal 10 indicating the position, the driving direction, whether the vehicle is running or stopped, the open / closed state of the door, the car call, the load, the service floor, etc.) 10 is the elevator system state data consisting of the hall call signal 2 and the car information signal 3. A group management device that assigns a hall call to an optimal car based on the assignment signal 4 and outputs it as an assignment signal 4. Each of the operation control devices A1 to An controls the hall call assigned by the assignment signal 4 and the car of the own car. In order for registered car calls To control the operation of the car so as to answer.

The group management device 10 is composed of a microcomputer or the like, and includes an input / output interface 11, learning sample creation means 12 for creating learning samples based on the elevator system state data, and traffic when creating learning samples. A traffic condition determining means 13 for determining a situation, a learning sample storing means 14 for storing learning samples for each traffic condition, and a neural network 15 for call assignment.
And assignment determining means 16 for selecting a car which seems to be optimal from the output pattern of the neural network 15 and assigning the car to a hall call, and selecting a predetermined number or more of the stored learning samples at a predetermined ratio for each traffic condition. And learning means 17 for learning the neural network 15 by using the above. Each of these means is realized on software in a microcomputer.

In the above configuration, the learning procedure in the present invention will be described with reference to the flowchart of FIG. Here, as an example, a learning procedure for creating a new neural network will be described.

First, it is determined in step S1 whether or not a new hall call has occurred.
Then, the elevator system state data at the time of occurrence of the call is converted into an input pattern of a neural network. During this time, the call is assigned and the answering machine begins to operate at the call. The assignment during this time is performed by using a conventional call assignment control by simulation or the like.

When it is confirmed in step S3 that the call has been serviced, the serviced car is converted to an output pattern of a neural network as an assigned car in step S4. At this time, if the initially assigned number is different from the actually serviced number due to an assignment change or reassignment, etc., that is, if the initial assignment is not the best as a result (however, the In the case of the case where there is no assigned device, the actually serviced device is converted into the output pattern as the assigned device. Then, the traffic condition at this time is determined in step S5. This traffic condition is determined, for example, by associating each traffic condition with the time in advance, or by a method of making a determination based on the time at that time, or the occurrence status of hall calls per unit time (number of floors, etc.). A known method such as a method of judging from the average waiting time may be used.

Next, in step S6, it is determined whether or not the number of learning samples for traffic conditions determined in step S5 has reached a predetermined number. For example, if the traffic condition determined in step S5 is the UP peak, while the number of training samples for the UP peak that have already been accumulated has reached a predetermined number, there is no more need for the UP peak. The process returns to step S1, but if the number has not reached the predetermined number yet, the process proceeds to step S7 to create a learning sample by pairing the input pattern and the output pattern.
Then, in step S8, it is classified according to traffic conditions, that is, in this case, it is stored under the classification of the UP peak.

In step S9, it is confirmed whether or not all the data necessary for learning the neural network has been prepared.
That is, the above procedure is repeated until a predetermined number of learning samples are accumulated for each traffic situation, and when all are completed, the process proceeds to step S10, and learning of the neural network is started.

For example, it is assumed that a neural network capable of coping with all traffic conditions is created. In this case, the total number of learning samples required for one learning is 1000.
Individually, UP peak, DOWN peak, 2WAY (time when traffic demand is high in both directions like lunch)
The ratio to normal traffic conditions is 1: 1: 1:
If it is 2, the required number for each traffic situation is set in advance as follows.

Learning sample for UP peak: 200 pieces for DOWN peak 200 pieces for 2WAY (at lunch) 200 pieces for normal time 400 pieces for normal For example, if you want to create a dedicated neural net that only uses the allocation at the peak of the UP,
The following settings are made so as to select only the data of the P peak.

Learning sample of UP peak: 1000 pieces DOWN peak ・ ・ ・: 0 pieces 2WAY pieces: ０ 0 pieces Normal time 〃: 0 pieces Neural to be created in this way Depending on the nature of the net, the total number of learning samples and the ratio for each traffic situation are determined in advance, and then the above procedure is executed to create a neural net that can fully demonstrate the required performance can do.

Although the above description has been made on the learning procedure when a new neural network is created, the same procedure as described above can be used when the neural network is re-learned after the operation of the elevator. . In that case, the assignment at the time of creating the learning sample is performed in the same procedure as above using the neural network, and the re-training of the neural network after the creation of the learning sample can be performed in parallel with the call assignment or It may be performed during off-peak hours such as at night.

[0029]

According to the present invention, it is possible to perform learning according to the nature of a neural network to be newly created or to be re-learned. Can be surely created or re-learned, and a neural network dedicated to a specific traffic condition can be easily created or re-learned.

[Brief description of the drawings]

FIG. 1 is a block diagram showing the overall configuration of the present invention.

FIG. 2 is a flowchart illustrating an example of a learning procedure according to the present invention.

FIG. 3 is a diagram showing an example of a neural network for call assignment.

[Explanation of symbols]

 A1 to An 1st to nth operation control devices 1 hall call button 2 hall call signal 3 car information signal 4 assignment signal 10 group management device 11 input / output interface 12 learning sample creation means 13 traffic situation determination means 14 learning sample Storage means 15 neural network 16 assignment determination means 17 learning means

Continued on the front page (72) Inventor Mami Nakagawa 1-28-10 Sho, Ibaraki-shi, Osaka Prefecture Inside Fujitec R & D Co., Ltd. (72) Inventor Takeshi Midoriya 1-28-10 Sho, Ibaraki-shi, Osaka Stock Company In Fujitech R & D Co., Ltd. (72) Inventor Tomoaki Tanabe 1-228-10 Shojo, Ibaraki-shi, Osaka Investigator in Fujitec R & D Co., Ltd. Yuji Shimizu (56) References JP-A-4-32472 (JP, A (58) Field surveyed (Int.Cl. ⁷ , DB name) B66B 1/18-1/20

Claims (1)

(57) [Claims] When a plurality of elevators are put into service on a plurality of floors and a hall call occurs, various data representing the state of the elevator system at that time are converted into a form that can be used as an input pattern of a neural network. A learning sample creating means for creating a learning sample in an elevator group management control device configured to input the neural network for assignment and to select an optimum car from the output pattern and assign the car to the hall call. ,
Traffic condition determining means for determining a traffic condition at the time of creating a learning sample; learning sample storing means for classifying and storing learning samples according to traffic conditions according to the determination result; said a sample news
-At a predetermined rate according to the required properties of the
And a learning means for learning the neural network by using the selected learning sample.