JP2021140490A - Simulator, simulation method, and simulation program - Google Patents

Abstract

To simulate a crowd action that reflects a sequence determination and selection performed by an agent having a plurality of action purpose in consideration of a possibility of accomplishment under a time limit.SOLUTION: A simulator, which simulates a behavior that an agent having a plurality of action purposes accomplishes the plurality of action purposes respectively at a plurality of destinations, comprises: combination selection means 50 that selects a plurality of combinations of destinations accomplishing each of the plurality of action purposes; evaluation value calculation means 50 that predicts a total of probabilities of accomplishing the action purposes at the destination by an accomplishment term on the basis of a probability density distribution of required time, and calculates an evaluation value; and action simulation means 52 that simulates an action sequentially visiting the destinations selected based on the evaluation value. The evaluation value calculation means 50 is configured to: determine a probability density distribution of remaining time in accordance with the probability density distribution of the required time of the action purpose to be firstly accomplished and the accomplishment term; and predict a probability of later accomplishing the action purpose by the accomplishment term on the basis of the probability density distribution of the remaining time.SELECTED DRAWING: Figure 2

Description

The present invention relates to a simulator, a simulation method and a simulation program for predicting the behavior of a crowd.

In recent years, in a space where a large number of people gather, a simulator for predicting the behavior of a crowd has been used in order to predict the effect of space design and its improvement, and the effect of introducing operational measures and their improvement. Further, as one of the methods used in such a simulator, a multi-agent simulation is known in which each individual constituting the community is represented by data called an agent.
In general, in a space where a large number of people gather, there are places (destination) that can achieve multiple purposes (action purposes), and in multi-agent simulation, it is necessary to determine the action purpose and destination of each agent. be.
For example, in Patent Document 1, for each individual imitating a T-junction passerby, the purpose of moving to A station, moving to B station, moving to the audience seats, etc., and the west side contact port, east side contact port, and south side It is described that the user of the simulator sets the movement target position of the contact port or the like in advance. Further, for example, Patent Document 2 describes that the selection purpose of an individual whose attribute is a leader and the selection area corresponding to the selection purpose are determined by lottery from the evacuation / exit area and the fire extinguishing / combustion area.

Japanese Unexamined Patent Publication No. 2019-21980 Japanese Unexamined Patent Publication No. 2019-185073

Here, prediction accuracy is required for the simulator in order to reduce trial and error in design and measures, and one of the factors that influences prediction accuracy is the possibility that individuals with multiple behavioral objectives can achieve it under a time limit. There is a choice (decision-making) to be made in consideration of. For example, during a 15-minute break in a real-world theater, one of the spectators is likely to prioritize shopping and then head to the store to ensure that goods are available during the break, followed by extra time for the toilet. For example, if you plan to move to the toilet and give up the stool if there is little chance of surplus time, prioritize the possibility of achieving the later action purpose if you achieve the earlier action purpose. Select a destination by making decisions and selections.
However, in the multi-agent simulation performed by the conventional simulator, the user of the simulator sets the action purpose and destination of each agent in advance, or the lottery is performed at each time point. For this reason, in the prior art, it has been difficult to simulate a crowd behavior that reflects the ordering and selection performed by agents having a plurality of behavioral objectives in consideration of the possibility of achievement under a time limit.
For example, when the simulator user sets in advance, the number of combinations of individual and action purpose must be set, and each of them must be set in consideration of the achievability of the first action purpose and the second action purpose. Must be. Further, for example, the decision by lottery does not determine a plurality of action objectives and destinations, nor does it determine the achievability of the earlier action objectives and the later action objectives.

The present invention has been made in view of the above problems, and simulates a crowd behavior that reflects the ordering and selection performed by agents having a plurality of behavioral objectives in consideration of the possibility of achievement under a time limit. The purpose is a simulator, a simulation method, and a simulation program.

According to one embodiment of the present invention, an agent having a plurality of action objectives is provided with a simulator that simulates an action in which an agent having a plurality of action objectives achieves a plurality of action objectives at a plurality of destinations in a predetermined space. The simulator achieves a plurality of action objectives at a combination of a combination selection means for selecting a plurality of destination combinations that achieve each of a plurality of action objectives from a plurality of destinations and a combination of destinations. An evaluation value calculation means that predicts the total probability of achieving each by the deadline based on the probability density distribution of each time required to achieve a plurality of action objectives, and calculates the evaluation value according to the total. It is provided with an action simulating means for simulating the behavior of an agent moving so as to sequentially visit a combination of destinations selected based on an evaluation value from a plurality of combinations of destinations. The evaluation value calculation means determines the probability density distribution of the time required for the action purpose to be achieved first among the plurality of action objectives and the probability density distribution of the remaining time according to the achievement deadline, and the action objective to be achieved later is reached until the achievement deadline. Is predicted based on the probability density distribution of the remaining time.

According to another aspect of the present invention, a simulation method is provided in which an agent having a plurality of action objectives simulates an action in which an agent having a plurality of action objectives achieves a plurality of action objectives at a plurality of destinations in a predetermined space. In the simulation method, a combination selection process for selecting a plurality of combinations of destinations that achieve each of a plurality of action objectives from a plurality of destinations, and for each combination of destinations, a plurality of action objectives are selected at the combined destinations. Evaluation value calculation processing that predicts the total of the probabilities to be achieved by the achievement deadline based on the probability density distribution of each time required to achieve multiple action objectives, and calculates the evaluation value according to the total. , The computer executes a behavior simulation process that simulates the behavior of an agent who moves so as to sequentially visit a combination of destinations selected based on an evaluation value from a plurality of combinations of destinations. In the evaluation value calculation process, the probability density distribution of the time required for the action purpose to be achieved first among multiple action objectives and the probability density distribution of the remaining time according to the achievement deadline are determined, and the action objective to be achieved later is reached until the achievement deadline. Is predicted based on the probability density distribution of the remaining time.

According to still another aspect of the present invention, an agent having a plurality of behavioral objectives is provided with a simulation program that simulates an behavior in which an agent having a plurality of behavioral objectives achieves a plurality of behavioral objectives at a plurality of destinations in a predetermined space. The simulation program performs a combination selection process that selects a plurality of combinations of destinations that achieve each of a plurality of action objectives from a plurality of destinations, and multiple action objectives at the combined destinations for each combination of destinations. Evaluation value calculation processing that predicts the total of the probabilities to be achieved by the achievement deadline based on the probability density distribution of each time required to achieve multiple action objectives, and calculates the evaluation value according to the total. , The computer is made to execute an action simulation process that simulates the behavior of an agent that moves so as to sequentially visit a combination of destinations selected based on an evaluation value from a plurality of combinations of destinations. In the evaluation value calculation process, the probability density distribution of the time required for the action purpose to be achieved first among multiple action objectives and the probability density distribution of the remaining time according to the achievement deadline are determined, and the action objective to be achieved later is reached until the achievement deadline. Is predicted based on the probability density distribution of the remaining time.

According to the present invention, it is possible to simulate a crowd behavior that reflects the ordering and selection performed by agents having a plurality of behavioral objectives in consideration of the possibility of achievement under a time limit.

It is a schematic block diagram of an example of the simulator which concerns on embodiment. It is a functional block diagram of an example of the simulator which concerns on embodiment. It is a schematic diagram of an example of the map information of the target space stored in the spatial information storage means. It is a figure which shows an example of the spatial information (behavior purpose information and destination information) stored in a spatial information storage means. It is a figure which shows an example of the 1st agent information stored in the agent information storage means. It is a figure which shows an example of the 2nd agent information stored in the agent information storage means. It is a figure which shows an example of the 3rd agent information stored in the agent information storage means. It is a figure which shows an example of the 4th agent information stored in the agent information storage means. It is a figure which shows an example of the statistical information stored in the statistical information storage means. (A) and (b) are explanatory diagrams of an example of the calculation method of the success probability of achieving the first action purpose. (A) to (c) are explanatory diagrams of an example of the calculation method of the success probability of achieving the second and subsequent action objectives. (A) and (b) are diagrams for explaining how the probability distribution stored in the statistical information storage means is updated by the statistical information updating means, and a difference occurs for each agent. It is a flowchart of an example of the whole processing of the simulation method of 1st Embodiment. It is a flowchart of an example of a statistical information update process. It is a flowchart of an example of a statistical information update process. It is a flowchart of an example of an action decision process. It is a flowchart of an example of an evaluation value calculation process. It is a flowchart of an example of agent state update processing. It is explanatory drawing of the method of predicting the number of arrivals of an agent to a destination in 2nd Embodiment. It is a functional block diagram of an example of the waiting time calculation means in the 2nd Embodiment. It is a flowchart of an example of the required time distribution calculation process in 2nd Embodiment.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. The embodiments of the present invention shown below exemplify devices and methods for embodying the technical idea of the present invention, and the technical idea of the present invention includes the structure, arrangement, and the like of components. Is not specified as the following. The technical idea of the present invention may be modified in various ways within the technical scope specified by the claims stated in the claims.

As an embodiment of the present invention, an example of a simulator that predicts the behavior of a plurality of people using the audience seats, toilets and shops in a theater having audience seats, toilets and shops is shown.
That is, in the present embodiment, the space (target space) to be simulated is a theater, the individuals constituting the crowd are people, and the action objectives of each individual are "seating" in the theater seat and "sit". "Usage" and "shopping". Hereinafter, individual data in the simulator will be referred to as "agent".

Further, in this embodiment, an example in which a simulation performer (hereinafter referred to as “user”) predicts the behavior of a crowd during the inter-curricular time using a simulator will be described. In addition, an example in which the prediction of crowd behavior changes after multiple intermission times will be described. By the way, for the sake of simplicity, the pre-curtain and post-curtain are omitted.
The theater described in the present specification is merely an example of the target space, and the subject of the present invention is not limited to this.

(First Embodiment)
(composition)
FIG. 1 is a schematic configuration diagram of an example of a simulator according to the first embodiment.
The simulator 1 includes an operation input unit 2, a file input / output unit 3, a storage unit 4, a control unit 5, and a display unit 6. Of these, the file input / output unit 3, the storage unit 4, and the control unit 5 can be realized by a so-called computer, and the operation input unit 2 and the display unit 6 can be realized as peripheral devices of the computer.
The operation input unit 2 is a user interface such as a keyboard and a mouse, and is operated by a user to be used for data input and the like. The operation input unit 2 is connected to the control unit 5, converts the user's operation into an operation signal, and outputs the operation signal to the control unit 5.

The file input / output unit 3 is a DVD (Digital Versatile Disc) drive, a USB (Universal Serial Bus) interface, a network interface, etc., one of which is connected to an external device (not shown), a recording medium, a network, or the like, and the other is a control unit. It is connected to 5, the data is input to the control unit 5 as a file, and the data is output from the control unit 5 as a file.
The storage unit 4 is a memory device such as a ROM (Read Only Memory) and a RAM (Random Access Memory), and stores various programs and various data. The storage unit 4 is connected to the control unit 5 and inputs / outputs these information to and from the control unit 5.

The control unit 5 is composed of arithmetic units such as a CPU (Central Processing Unit), a DSP (Digital Signal Processor), and an MCU (Micro Control Unit). The control unit 5 is connected to the storage unit 4 and operates as various processing units by reading and executing a program from the storage unit 4, and stores and reads various data in the storage unit 4. Further, the control unit 5 is also connected to the operation input unit 2, the file input / output unit 3, and the display unit 6, and the user operates the operation input unit 2 to acquire the input data, and the file input / output unit 3 is used. Data is acquired and output as a file from the outside via the file, simulation is performed based on the income data to predict the behavior of the crowd, and the prediction result data is output as a file from the file input / output unit 3 and / or predicted. The resulting data is imaged and displayed on the display unit 6.
The display unit 6 is a display device such as a liquid crystal display or a CRT (Cathode Ray Tube) display, is connected to the control unit 5, and displays a prediction result by the control unit 5. The user visually recognizes the displayed prediction result and considers the behavior of the crowd.

(Function of simulator 1)
FIG. 2 is a functional block diagram of an example of the simulator 1 according to the first embodiment.
The operation input unit 2 functions as a part of the condition setting means 20, and the file input / output unit 3 functions as a part of the condition setting means 20 and a part of the prediction result output means 60.
The storage unit 4 functions as a simulated condition storage means 40, a spatial information storage means 41, an agent information storage means 42, a statistical information storage means 43, and the like.

The control unit 5 executes a computer program stored in the storage unit 4 by an arithmetic unit such as a CPU, DSP, or MCU to update the evaluation value calculation means 50, the destination selection means 51, the action simulation means 52, and the statistical information. It functions as means 53 and the like.
The condition setting means 20 receives input of various conditions necessary for the simulation, and stores the input conditions in the simulated condition storage means 40, the spatial information storage means 41, the agent information storage means 42, and the statistical information storage means 43.

The condition setting means 20 is realized, for example, in collaboration with the operation input unit 2 and the control unit 5, and the control unit 5 stores the value input by the user by operating the operation input unit 2 in the storage unit 4. .. The condition setting means 20 may include a file input / output unit 3 for inputting a file in which a part or all of the conditions are described, and the control unit 5 files a file specified by the user by operating the operation input unit 2. The value acquired through the input / output unit 3 and recorded in the file is stored in the storage unit 4.
The conditions input by the condition setting means 20 and stored by the simulated condition storage means 40 are, for example, the setting of the upper limit T of the virtual time, the action determination interval, the observation interval, and the agent basic speed (distance to advance in one time). including.

The simulation of the first embodiment is performed while advancing the virtual time. For example, the process of updating the agent information for each virtual time is repeated while increasing the virtual time from 0 to 1 until the virtual time reaches (T-1). The operation input unit 2 may accept the input of the end instruction by the user, and may add a process of forcibly terminating when the end instruction is input.
The virtual time simulates the passage of the actual time, but it is advanced every time the simulator finishes the calculation for one hour, and is ticked at a higher speed than the passage of the actual time. For example, one time of virtual time corresponds to an actual 0.1 second, and the virtual time is represented by an integer starting from 0 and increasing by 1.

The spatial information storage means 41 includes map information representing the target space, action purpose information that defines an action purpose that can be achieved in the target space, and destination information regarding a destination that is a place where the action purpose can be achieved in the target space. Stores spatial information including. In the target space, there are a plurality of destinations capable of achieving this action purpose for at least one action purpose.
FIG. 3 is a schematic diagram of an example of map information 100, and FIG. 4 is a diagram showing an example of action purpose information 110 and destination information 111. These spatial information are preset by an input operation by the user using the condition setting means 20, except for the "queue" included in the destination information 111.

In the map information 100 shown in FIG. 3, the line segment represents a wall, and the area other than the line segment is a movable area in which the agent can move.
The Roman numerals I to X are destination IDs given to each destination, and indicate the location of each destination in the map information 100 of FIG.
Double circles Q1 to Q10 indicate the positions of representative points representing the positions of the destinations I to X. The positions of destinations I to X, which are actually areas, are defined as point information to facilitate handling.
The circles Q11 to Q21 indicate the waypoints that the agent can pass through to move in the movable area while avoiding the wall. The route on which the agent moves in the movable area is defined by a local route connecting representative points, transit points, and representative points and transit points (hereinafter, referred to as "local route").

See FIG. The action purpose information 110 includes an "action purpose name" of each action purpose in association with the action purpose IDs (1 to 3).
Further, the destination information 111 is associated with the destination IDs (IX), and is associated with the "destination name", "representative point", "gender restriction", "number of service counters", and "achievable" of each destination. Includes "purpose of action", "service time distribution" and "queue".

The "destination name" is the name of the destination. For example, the destination I is the "audience seat", the destination II is the "women's toilet 1", the destination III is the "women's toilet 2", the destination IV is the "men's toilet 1", and the destination V is "men's toilet 2", and destinations VI to X are "store 1" to "store 5".
The "representative point" is the XY coordinate value of the above-mentioned representative points Q1 to Q10.

"Gender restriction" is one of the attribute restrictions of the agent who can use the destination, and is the gender of the agent who can use the destination.
The "number of service counters" is the maximum number of agents who can use the destination at the same time. In the following description, "service" means the use of a destination to achieve an action purpose. For example, if the destination is a "toilet", the service is to provide a place for stool, and if the destination is a "stand", the service is a sales service.

An "achievable purpose of action" is an action purpose that can be achieved at the destination.
The "service time distribution" is statistical information representing the time required for a service at the destination (hereinafter referred to as "service time") as a probability distribution. For example, the mean value and variance when the service time is expressed by a normal distribution are stored. The service time distribution is used to determine the service time required for each agent to achieve its behavioral objective at the destination.

The "queue" is a buffer in which the agent IDs of the agents staying at the destination are arranged in the order of arrival at the destination. The queue is sequentially updated by the behavior simulating means 52.
An agent whose order in the queue is less than or equal to the number of service counters is an agent receiving service at the destination, and an agent whose order in the queue is larger than the number of service counters is an agent waiting for service. The number of agents waiting for service is the number of waits (waiting number).

In the example of the spatial information 111 of FIG. 4, four agents # 9, # 63, # 39, and # 83 have arrived at the shop 1, and since the number of service counters is 1, the agent # 9 at the head of the queue. Indicates that only are receiving service (eg, paying at the kiosk). Further, the remaining agents # 63, # 39, and # 83 are waiting for service (whether they are selecting a product or waiting for a cash register), and the number of waits is 3.

See FIG. The agent information storage means 42 stores information about each of a plurality of agents constituting the crowd together with the statistical information storage means 43. Hereinafter, the information stored in the agent information storage means 42 is referred to as "agent information".
5, FIG. 6, FIG. 7 and FIG. 8 show an example of agent information 120, 121, 122 and 123, respectively. The character strings shown in brackets next to the numerical values in each figure have meanings added in consideration of the legibility of the figure.
For example, at the initial position of the agent information 120 (4.55, 14.00), the value (4.55, 14.00) is actually stored in the place described as "door 1", which is the "door". It means the coordinates of "1". In the following description, the notation of the value may be omitted and the meaning may be explained instead.

The agent information 120 in FIG. 5 is a part of the information set by the input operation by the user using the condition setting means 20. The agent ID, generation time, initial position, and gender combination are stored as the agent information 120. The initial position is the initial value of the position of the agent, and indicates the position where the agent is generated at the generation time.
The example of FIG. 5 shows that female agent # 1 is generated at the coordinates of "door 1" at virtual time 0 in "between curtains 1". Further, the description of "intermission 2" and "intermission 3" will be omitted. Furthermore, it is shown that it is generated at the coordinates of "door 3" at the virtual time 99600 in "intermission 4". Note that in the simulation example of the first embodiment, the same agent is generated multiple times.

The agent information 121 of FIG. 6 is also set by an input operation by the user using the condition setting means 20. The agent information 121 includes a combination of "action purpose ID", "achievement deadline", "desired achievement degree", "termination purpose flag", and "achievement flag" in association with the combination of the agent ID and the generation time. .. However, among these information, the achievement flag is appropriately updated by the action simulating means 52.

The "action purpose ID" is an ID of the purpose (action purpose) in which the agent acts in the target space in each scene (between the curtains). Multiple action objectives can be set for each agent in each scene.
The "achievement deadline" is the achievement deadline for each action purpose of the agent in each scene, and a virtual time after the generation time of the agent is specified.
The "desired achievement level" is the priority level for the achievement of each action purpose of the agent in each scene. For example, the desired achievement level may be the subjective satisfaction level (utility) that the agent can obtain by achieving the action purpose. The higher the degree of achievement of hope, the higher the priority.

The "termination purpose flag" is a flag indicating the action purpose that must be achieved at the end of the action in each scene. The action purpose for which the value "1" is set in the terminal purpose flag is constrained that the achievement order is the last action purpose. On the other hand, for the action purpose for which the value "0" is set in the terminal purpose flag, the order of achievement is arbitrary and whether or not it is achieved is arbitrary.
The "achievement flag" is a flag indicating whether or not each action purpose of the agent in each scene has been achieved. The initial value of the achievement flag is "0", the value "0" indicates unachieved, and the value "1" indicates achieved.

In the example of FIG. 6, "Agent # 1" of "Makuma 1" achieves "stool" by "3 minutes later" and "seats" on the condition that "seating" is finally achieved. The goal is to achieve "" and "shopping" by "10 minutes later", and the desired achievement levels of "seating", "feces", and "shopping" are "35", "100", and "40", respectively. It also shows that the purpose of action of is not achieved.

The agent information 122 in FIG. 7 is information generated for each time. The agent information 122 includes a combination of "destination ID", "local route", "movement speed", "position", and "situation" in association with the combination of the agent ID and the virtual time.
The "destination ID" indicates the ID of the destination that each agent at the time indicated by the virtual time is the target of the movement. In the example of the simulator 1 of the first embodiment, the destination ID at the generation time is left blank.
“Local route” indicates the local route to which each agent is moving at the time indicated by the virtual time. In the example of the simulator 1 of the first embodiment, the local route at the generation time and the non-moving time is left blank.

The "movement speed" indicates the amount of movement that each agent at the time indicated by the virtual time tries to move in a unit time (1 virtual time). In the example of the agent information 122 of FIG. 7, the moving speed is represented by a vector.
The movement speed is just the amount of movement intended by each agent, and does not necessarily match the amount of movement that the agent moves at the time indicated by the virtual time. For example, due to physical restrictions between agents, the current time position may not be the value obtained by adding the current time movement speed to the previous time position.
Further, for convenience, the moving speed at the generation time and the moving speed from reaching the destination to achieving the action purpose achievable at the destination are set to 0 vectors.

The "position" is the position in the target space of each agent at the time indicated by the virtual time. The initial value of the position of each agent is the initial position set at the generation time of the agent with reference to the agent information 120 described above.
The coordinates of the representative point of the destination are set from the time when the destination is reached until the action purpose that can be achieved at the destination is achieved.

The "status" is the status of each agent at the time indicated by the virtual time, and the possible values are "moving", "waiting", "processing", and "achieved".
"Moving" indicates that the destination has not been reached and is moving.
"Waiting" indicates that the destination has been reached but is waiting for service.
"Processing" indicates that the service is being received at the destination.
"Achievement" indicates that the service has been received at the destination and the action purpose has been achieved.
In the example of the simulator 1 of the first embodiment, "achievement" is designated as the situation at the generation time for convenience.

The example of FIG. 7 illustrates the agent information 122 at the end of the simulation.
For example, "Agent # 1" of "Makuma 1" moves in the order of "Women's Toilet 1", "Shop 2", and "Audience Seat", and achieves the action purpose at "Store 2" and "Audience Seat" at times "2909" and "2969". What was done is recorded along with the movement speed, position and situation of each time during that time. In the process, the service was awaited at "Kiosk 2" from time "1322" to "2686", time "2687". It is recorded that the service was received from "2908".

The agent information 123 of FIG. 8 is generated as necessary by the behavior simulating means 52 in order to determine whether or not the service has been received. The agent information 123 includes the service completion time in association with the agent ID. The information contained in the agent information 123 is retained while the corresponding agent is receiving the service at the destination at the current time, and is deleted when the service is completed.

The "agent ID" indicates an agent receiving service at the destination (that is, an agent whose status is "in process"). The destination receiving the service can be specified by the agent information 122.
The "service completion time" is the time when the agent finishes receiving the service at the destination.
In the agent information 123 of FIG. 8, the service completion time is set for the agents # 9, # 34, # 76, # 5, # 36, etc. at the head of the queue of the shops 1 to 5 illustrated in the spatial information 111 of FIG. It is being held.

The achievement flag included in the agent information 121 and the position and status included in the agent information 122 are examples of state values and represent the state of the agent.
The action purpose included in the agent information 121, the achievement deadline for each action purpose, the desired achievement degree for each action purpose, the termination purpose flag for each action purpose, and the destination, local route, and movement speed included in the agent information 122 are described. This is an example of behavioral parameters. The behavior parameter is a parameter that affects the state value of the agent.
The agent information 120, 121, 122 and 123 shown in FIGS. 5 to 8 are merely examples, and the structure of the agent information of the present invention is not limited thereto.

See FIG. The statistical information storage means 43, together with the agent information storage means 42, stores information about each of the plurality of agents constituting the crowd.
The information stored in the statistical information storage means 43 is statistical information regarding the time required to achieve the action purpose at each destination. By storing such statistical information for each agent, it is possible to simulate the difference in knowledge that each individual in the real world has about the required time. That is, the statistical information storage means 43 stores statistical information regarding each required time required to achieve the action purpose at a plurality of destinations for each agent.

The time required for each action purpose is defined as the sum of the travel time to the destination, the waiting time at the destination, and the service time at the destination.
The waiting time at the destination is modeled as being calculated by dividing the number of waits at the destination by the number of service counters at the destination and the product of the service time.

Further, the travel time to the destination is modeled as being calculated by summing up the results obtained by dividing the route length of each local route to the destination by the travel speed of the local route.
Therefore, the statistical information storage means 43 has a probability distribution of the number of waits for each destination (wait number distribution), a probability distribution of service time for each destination (service time distribution), and a movement speed for each local route for each agent. The probability distribution (movement speed distribution) of is stored as statistical information.
For example, each of the waiting number distribution, the service time distribution, and the moving speed distribution is modeled by a normal distribution. Then, in order to update each distribution with the observed value, the joint probability distribution of the parameters of each normal distribution of the waiting number distribution, the service time distribution, and the moving speed distribution is used. The statistical information storage means 43 stores four parameters of the mean value μ, the degree of freedom ν, the hypothetical sample number κ, and the scale parameter λ, which are the parameters of the simultaneous probability distribution of the waiting number distribution, the service time distribution, and the moving speed distribution. do. The joint probability distribution is expressed as Ν ・ χ- ² (μ, ν, κ, λ).

The initial values of the waiting number distribution, the service time distribution, and the moving speed distribution are preset by the condition setting means 20. The initial value may be created based on experience, common sense, actually measured statistical information, and the like. The initial value may be common to all agents, or may be different for each agent.
Further, a part of the combination of the agent and the required time may not be set, and the initial setting may be set so that the statistical information storage means 43 does not store the combination. This makes it possible to simulate the lack of knowledge in a real-world individual, for example, agent # 1 does not know the existence of shop 4.
In the present embodiment, the initial setting is set to store the initial value of the distribution (waiting number distribution and service time distribution) related to one or more destinations corresponding to each action purpose for each agent. The destination to be stored is determined by a random lottery in advance. Further, in the present embodiment, for the local route, the initial setting for storing the initial value of the distribution (movement speed distribution) for the local route is set for all combinations of the agent and the local route.
Further, in the present embodiment, the statistical information storage means 43 stores the initial value of the distribution related to the destination for each destination, separately from the information for each agent. This information is used as an initial value when each agent first observes the destination.

As will be described later, the statistical information of each agent is individually updated and referred to by the observation values obtained with the behavior of the agent. As a result, just as the empirical knowledge of individuals differs from one another in the real world, they become different from each other as the virtual time elapses.
Further, the statistical information storage means 43 adds the unmemorized combination of the agent and the required time combination (that is, the destination or local route that the agent does not know) when the observed value is obtained. By memorizing, the acquisition of new knowledge in an individual in the real world may be simulated.

FIG. 9 shows an example of statistical information 130 and 131 stored in the statistical information storage means 43.
The statistical information 130 includes a "waiting number distribution" and a "service time distribution" associated with a combination of an agent ID and a destination ID.
The "waiting number distribution" and the "service time distribution" are joint probability distributions of the "waiting number distribution" and the "service time distribution" that the agent stores for the destination, respectively.
In the example of the statistical information 130 of FIG. 9, the waiting number distribution and the service time distribution of the audience seat, the women's toilet 1, the men's toilet 1, the shop 1, the shop 2 and the shop 3 are stored for the agent # 1, and the women's toilet 2 and the boys are stored. The distribution of the toilet 2, the shop 4, and the shop 5 is not remembered.
Statistical information 131 includes a "movement velocity distribution" associated with a combination of agent ID and local route. The "movement speed distribution" is a joint probability distribution of the "movement speed distribution" that the agent remembers for the local route.

See FIG. The evaluation value calculation means 50 calculates an evaluation value as a criterion for selection when each agent selects a destination that achieves an action purpose in each scene (between curtains).
As described above, since a plurality of action objectives to be achieved by an agent may be set in each scene, the action objectives to be achieved in each scene are expressed as a permutation of a plurality of set action objectives (hereinafter referred to as "action pattern"). ).
In addition, there may be multiple destinations that can achieve the same purpose of action. In this case, there are multiple options for the destination to achieve each action purpose.
The evaluation value calculation means 50 calculates an evaluation value for each destination option that can be selected for each action pattern. Then, the evaluation value information in which the agent, the option, and the evaluation value are associated with each other is output to the destination selection means 51. The destination options that can be selected for the behavior pattern are permutations of destinations, and the permutations are hereinafter referred to as "destination patterns".

Therefore, first, the evaluation value calculating means 50 generates information on the behavior pattern. Specifically, the evaluation value calculation means 50 uses the combination of the agent ID and the generation time (each scene) stored in the agent information 121 for each combination having the action purpose ID in which the achievement flag is 0. A permutation is generated in which the action purpose IDs having the achievement flag of 0 are arranged in order from the action purpose IDs associated with. However, the evaluation value calculating means 50 sets the action purpose for which the terminal purpose flag is set at the end of the permutation.
For example, in the case of agent # 1 in the agent information 121 of FIG. 6, the action pattern when the virtual time is 0 is P ₁ = {shopping, shopping, sitting}, P ₂ = {shopping, shopping, sitting}. , P ₃ = {stool, seated}, P ₄ = {shopping, seated}, P ₅ = {seat}.

Next, the evaluation value calculating means 50 generates information on the destination pattern. Specifically, in the evaluation value calculating means 50, for each behavior pattern, the "achievable behavioral purpose" in the spatial information 111 matches each element of the behavioral pattern, satisfies the gender restriction, and the behavioral pattern is generated. The destination ID stored in the statistical information 130 of the agent is used as a destination ID option for each element of the behavior pattern. Then, the evaluation value calculating means 50 selects one destination ID from the options acquired for each element of the action pattern for each action pattern, and selects the selected destination ID for each element of the action pattern. A permutation of destination IDs arranged in the order of is generated as a destination pattern. The evaluation value calculating means 50 is an example of the “combination selection means” described in the claims.
_{For example, for the action pattern P 1} = {toilet, shop, seating} of the agent # 1 in the inter-curtain 1 in the agent information 121 of FIG. 6, it is generated with reference to the spatial information 111 of FIG. 4 and the statistical information 130 of FIG. The destination patterns are M ₁₁ = {women's toilet 1, shop 1, audience seats}, M ₁₂ = {women's toilet 1, shop 2, audience seats}, M ₁₃ = {women's toilet 1, shop 3, audience seats} 3 It becomes a street.

Then, the evaluation value calculating means 50 calculates the evaluation value E for each destination pattern according to the following equations (1) to (4).

In the formula (1), P _i is the i-th action patterns of behavior patterns in each scene of the target agent to calculate the evaluation value, M _ij is one of the destinations pattern corresponding to the action pattern P _i it is the j-th destination pattern, N _i is the number of action purposes is included in the behavior pattern P _i.
P _IS shows a first s elements of behavior pattern _{P i} (i.e. behavioral purpose of performing the s th).
_mijs indicates the s element (that is, the sth destination) of the destination pattern M _ij.
p _{(P is, m ijs)} is, until the deadline for achieving behavior purpose _{P is,} is the probability to achieve at the destination _{m ijs} the action purpose _{P is.} Hereinafter, this probability is referred to as a success probability. As described above, _{the achievement deadline of the action purpose Pis is} set as the agent information 121, and as shown in FIG. 6, the achievement deadline is set for each agent and each action purpose ID of the action purpose to be achieved in each scene. It is specified. Success probability _{p (P is, m ijs)} for details of the calculation method of which will be described later.

In the formula (1), V ( _{Pis ) is} the desired achievement degree of the action purpose Pis. As described above, the desired achievement level V ( _Pis ) is set as the agent information 121, and as shown in FIG. 6, the desired achievement is achieved for each agent and for each action purpose ID of the action purpose to be achieved in each scene. Degree V ( _Pis ) is specified.
For example, in the agent information 121 of FIG. 6, the desired achievement degree of the agent # 1 in the inter-curtain 1 with respect to the action pattern P1 is 100 when s = 1, 40 when s = 2, and 35 when s = 3.

That is, the evaluation value can be calculated by weighting the success probability of each destination corresponding to a plurality of action objectives to be achieved by the desired achievement degree of the action objective corresponding to the destination and summing them up.
As another embodiment, weighting based on the desired achievement level can be omitted.
Further, as yet another embodiment, the evaluation value becomes higher as the success probability is higher, such as the product of the weighted success probabilities or the product of the success probabilities instead of the sum of the weighted success probabilities or the sum of the success probabilities. Functions can also be defined and used.

The above description of equation (1) can be summarized as follows.
Evaluation value calculating means 50, for each agent, based on the statistical information stored in the statistical information storage section 43, to achieve the action object P _IS before deadlines behavior object P _IS in multiple destinations m _ijs each the success probability _{p (P is, m ijs)} to predict the success probability _{p (P is, m ijs)} calculates an evaluation value E in accordance with the.
In this case, in order to select the destination of the agent to try to achieve a plurality of behavior purpose _P is, the combination of the destination m _ijs to achieve each of the plurality of behavior purpose _P is from among a plurality of destination m _ijs the destination pattern _{M ij} is to select multiple, probability _{p (P is, m ijs)} to achieve a plurality of actions the object _{P iS} destinations _{m ijs} belonging to the destination pattern _{M ij} evaluation value according to E Is calculated.
Further, the evaluation value calculating means 50 has a plurality of action objectives in order to enable the agent to select a destination according to the difference in priority in relation to different priorities for the action objectives for each agent. probability _{p (P is, m ijs)} to achieve each of the P _iS calculates an evaluation value E by weighted in the plurality of action object desired achievement V _{(P is).}

Hereinafter, formula (2) to (4) of success probability _{p (P is, m ijs)} illustrating a method of calculating the illustrating an example.
First, evaluation value calculation unit 50, equations (3) and in order to use (4), each destination included in the destination pattern _{M ij (m ij1, m ij2} , ...) in each action object ( The required time distribution (f ₁ (T _d ), f ₂ (T _d ), ...), Which is the probability distribution of the required time T _{d when executing Pi 1} , _{Pi 2, ...) Is calculated.}

Here, the required time distribution f _s (T _d ) is the movement time distribution when moving from the s _{-1st destination mij (s-1) to the sth destination mijs, and the destination mijs.} It is defined by the sum of the waiting time distribution at the destination and the service time distribution related to the achievement of the action purpose at the destination _mijs. However, _mij0 is the current position of the agent.
In addition, the variance σ ² = λ / between the parameters of the simultaneous probability distribution (mean value μ, degree of freedom ν, hypothetical sample number κ, scale population λ) and the normal distribution N (μ, σ ^{2) represented by it.} There is a relationship of (ν-2).
When each distribution is represented by a normal distribution, the average value of the sum distribution is the sum of the mean values of the original distribution, and the variance of the sum distribution is the sum of the variances of the original distribution.

The evaluation value calculating means 50 calculates the travel time distribution for the calculation of _{f s} (T _d). The evaluation value calculation means 50 is a processing target in which the statistical information storage means 43 stores the parameters of the joint probability distribution of the movement speed distribution of the local route passing from the destination _{mij (s-1) to the destination mijs.} It is read from the statistical information 131 of the agent of the above, and a normal distribution representing the movement speed distribution of each local route is derived from the read parameters. The evaluation value calculation means 50 obtains the movement time distribution of each local route by converting the speed axis in the movement speed distribution of each local route into a time axis by dividing by the distance of the local route, and obtains the movement time distribution of each local route. Calculate the travel time distribution to the destination by summing the distributions.

The evaluation value calculating means 50 also calculates the waiting time distribution for the calculation of _{f s} (T _d). The evaluation value calculation means 50 _{reads out} the parameters of the joint probability distribution of the waiting number distribution and the service time distribution of the destination mijs from the statistical information 130 of the agent to be processed stored in the statistical information storage means 43, and also aims. _{Read out} the number of service counters for the local parameters. The evaluation value calculating means 50 derives a normal distribution representing a waiting number distribution and a service time distribution from the read parameters. The evaluation value calculating means 50 calculates the waiting time distribution by multiplying the result of dividing the average value of the waiting number distribution (which is the waiting number) by the number of service counters read out by the service time distribution. The mode may be used instead of the average value, or a value determined by a random lottery that wins with the probability indicated by the waiting number distribution may be used. In addition, the number of waits for the destination can be the value observed by the agent to be processed at the current time (only when the corresponding observed value is obtained). Further, the observed value of the waiting number of the destination may be held in the storage unit 4 for several hours, and the held value may be used.

The evaluation value calculating means 50 also calculates the service time distribution for the calculation of _{f s} (T _d). The evaluation value calculation means 50 _{reads out the parameters of the joint probability distribution of the service time distribution of the destination mijs} from the statistical information 130 of the agent to be processed stored in the statistical information storage means 43, and services from the read parameters. Derivate a normal distribution that represents the time distribution.
Then, the evaluation value calculating means 50 calculates the sum of the travel time distribution, the waiting time distribution, and the service time distribution for each s as the required time distribution f _s (T _d ) for the s.

Then, evaluation value calculation unit 50 calculates the remaining time _T r _{(P IS)} action to the target _{P IS} of deadlines _T lim _{(P iS)} according to equation (4).
The remaining time in the case of s = 1, ie, the deadline for achieving _T lim _{(P i1)} remaining up to the time _T r _{(P i1)} of the first action purpose _{P i1} of behavior patterns _{P i} is a scalar. When s = 1, the evaluation value calculating means 50 calculates the difference obtained by subtracting the current time t _{c from the achievement deadline Trim} ( _{Pi 1 ) as the remaining time Tr} (P _{i 1} ).
s> remaining time of the time of 1, ie, the second and subsequent behavior purpose of behavior pattern _{P i P i2, P i3,} ... time remaining until the deadline for achieving is the probability distribution of. That is, the time left to achieve the second and subsequent action objectives _{Pi2, Pi3,} ... Is the time required for the action objective executed before the action objective T ¹ _d , ..., T ^{s-. Because} _{it depends on 1 d} , the required time T ¹ _d , ..., T ^s-1 _d for the action purpose are defined as random variables according to the required time distributions f ₁ (T _d ) to f _s-1 (T _{d), respectively.} The remaining time _{Tr of the} second and subsequent action objectives _{Pi2, Pi3,} ... Is also calculated as a random variable according to the probability distribution. This probability distribution is defined as the remaining time distribution g _s ( _Tr). In the present embodiment in which the required time distribution is a normal distribution, the remaining time distribution is also a normal distribution.
Evaluation value calculating means 50, s> case 1, s-th action object _{P IS} of deadlines _T lim and _{(P IS)} obtains the difference between the current time _{t c,} further s-1 th action from the first The sum of the _{time distributions f 1} (T _d ) to f _s-1 (T _d ) required to achieve the objectives _{Pi 1 to} Pi _{(s-1) is calculated.} The obtained distribution is called the sum distribution. _{Then, the evaluation value calculating means 50 calculates the remaining time distribution g s} ( _Tr ) using the value obtained by subtracting the average value of the total distribution from the above difference and the variance of the total distribution. That is, the average value of the remaining time distribution g _s ( _Tr ) is the value obtained by subtracting the average value of the total distribution from the above difference, and the variance of the remaining time distribution g _s ( _Tr ) is the variance of the total distribution. ..

Then, evaluation value calculating means 50 calculates the success probability _{p (P is, m ijs)} a according to equation (2) and (3). Since the success probability is the probability of achieving the action purpose by the achievement deadline, it can be defined as "the probability that the time required to achieve the action purpose is less than the remaining time".

When s = 1, since the remaining time _Tr ( _{Pi1 ) calculated by the equation (4) is a scalar, the value F s} ( _Tr ) represented by the equation (3) is also the success probability p represented by the equation (2). _{(P i1, m ij1)} also becomes a scalar. Specifically, the evaluation value calculating means 50 _{integrates the probability density represented by the required time distribution f s} (T _d ) with respect to the required time T _{d in} the range of −∞ to the remaining time _Tr , and is an integrated value F _{s. (T r)} is calculated as the success probability _{p (P i1, m ij1)} .
In the present embodiment, the integrated value _F s _{(T r),} since the required time distribution _f s _{(T d)} is a normal distribution, the time required as shown in equation (3) distribution _f s of _{(T d)} It can be calculated from the average value μ, the variance σ ^2, and the remaining time _{Tr (Tr} ( _Pi1 ) calculated by the equation (4)). The erf in equation (3) is an error function. Then, as shown in equation (2), the equation (3) integral value calculated in _F s _{(T r)} is as success probability _{p (P i1, m ij1)} .
FIG. 10 (a) shows an example of the required time distribution f ₁ (T _{d ) required to achieve the first action purpose Pi} 1, and FIG. 10 (b) shows the required time distribution f ₁ (T _d ). The cumulative distribution function F ₁ ( _Tr ) of is shown. In the figure, μ ₁ is the average value of the required time distribution f ₁ (T _d ), and the value of the cumulative distribution function F ₁ (T _{r ) when the required time is Tr} (P _{i 1} ) is the success probability p (P). _i1 , _mij1 ).

When s> 1, the remaining time _Tr ( _Pis ) calculated by the equation (4) becomes the distribution g _s ( _Tr ). Therefore, success probability _{p (P is, m ijs),} as shown in equation (2), the probability that the remaining time at the End of behavioral interest to s-1 th is _T r _{(P IS)} The probability obtained by multiplying _{g s} ( _Tr ) by the _{probability F s} ( _Tr ) of achieving the sth action objective with _{the remaining time Tr} ( _{Pis) is in} the range of −∞ to ∞ _{for the remaining time Tr.} It is a value obtained by integrating up to. The probability F _s ( _Tr ) at that time is calculated by the equation (3).

This formula can not be solved analytically, evaluation value calculating means 50 of the first embodiment, for example, numerical analysis success probability p (P _{is, m ijs)} as described below is calculated. Specifically, the probability density of the remaining time distribution by the sampling method g _{s (T r)} specimens were extracted for the specimens obtained by the sampling, s-th action object P _IS duration distribution f _s required to achieve based on the _{(T d)} and, success probability _{p (P is, m ijs)} is calculated.
_{In the present embodiment, the evaluation value calculating means 50 calculates the remaining time distribution g s} ( _Tr ) until the achievement deadline of the s-th (s = 2, 3, ...) Action objective _Pis at arbitrary time intervals. Discretize and calculate the histogram.

FIG. 11 (a) _{shows an example of the remaining time distribution g s} ( _Tr ), and FIG. 11 (b) shows a histogram of _{the remaining time distribution g s} ( _Tr). FIG. 11 (c) shows the cumulative distribution function F _s ( _Tr ) of the time distribution f _s (T _{d ) required to achieve the s th action objective Pis.} In the figure, μ _rem is the average value of the remaining time distribution g _s ( _{Tr ), and μ s} is the average value of the required time distribution f _s (T _d ).
Evaluation value calculating means 50, the occurrence probability of each bin (each section) b1-b5 of the histogram (frequency) and the cumulative distribution function value _F s of the representative value _T r1 _{through T r5} of each bin b1-b5 _{(T r1)} the sum of products ~F _{s (T r5),} success probability _{p (P is, m ijs)} is calculated as. Incidentally, the probability of occurrence of bin b1, when the width of the bins and w, the rest time distribution _g s _{(T r)} rightmost _T -∞ from bins b1 of r1 + w / remaining time from the integration value of up to 2 distribution _g s ( It can be calculated by subtracting the integrated value _{from −∞ of Tr ) to Tr1} −w / 2 at the left end of bin b1. The probability of occurrence of other bins is similar. Further, in the present embodiment in which the remaining time distribution g _s ( _Tr ) is a normal distribution, the integrated values up to the right end of the bin are the average value μ and the variance σ of the _{time g s} ( _{Tr) as in the equation (3).} ^It can be calculated by applying 2 and the time at the right end of the bin to the error function, and the integrated value up to the left end of the bin is _{the average value μ of g s} ( _Tr ) and the variance σ ² and the time at the left end of the bin. Can be calculated by applying to.

The method of calculating the probability of success of the action object P _IS 2 -th of the p (P _{is, m ijs),} the present invention is not limited to this is merely illustrative. By numerical analytical approach, such as performing the approximation using other sampling methods, success probability p (P _{is, m ijs)} may be calculated. For example, a sampling method such as the Box-Muller method (when the target distribution is a normal distribution), the rejection sampling method, or the MCMC method can be used.

The explanations of equations (2) to (4) can be summarized as follows.
Evaluation value calculation unit 50, behavior patterns _P for each destination pattern _{M ij} corresponding to the _i, the probability p (P for each achieved until the deadline for achieving the action purpose _{P is} at the destination _{m ijs} belonging to the destination pattern _{M ij The sum of is} , _{mijs ) is} predicted based on the probability density distribution f _s (T _{d ) of the time T d} required to achieve each of the plurality of action objectives Pis.
The evaluation value calculation unit 50, in response to a plurality of action object _{P IS} behavior object of required time to achieve the out destination of the probability density distribution _{f 1 (T d) ~f s} -1 (T d) and deadlines Te to determine the remaining time of the probability density distribution g _{s (T r),} based on the remaining time of the probability density distribution g _{s (T r),} the probability p _(P is to achieve the achievement to act purpose to achieve the deadline after , _Mijs ).
Further, the evaluation value calculating means 50 is based on _{a histogram of the remaining time distribution g s} ( _Tr ) and a cumulative distribution function of the _{time distribution f s} (T _d ) required to achieve the action purpose to be achieved later. , The sum of the products of the frequency of the histogram and the probability of the cumulative distribution function in each interval of the histogram _is calculated as _{the probability p (Pis} , mijs).
The evaluation value E is to be calculated according to the probability _{p (P is, m ijs)} , the evaluation value E shorter the waiting time at the destination is high.

The evaluation value calculating means 50 calculates the above-mentioned evaluation value E when a predetermined timing for determining the action of the agent (hereinafter referred to as “action determination timing”) arrives.
The action decision timing is, for example, the timing at which one of the generation time of each agent, the achievement of the action purpose (excluding the last action purpose), and the predetermined action decision interval (for example, 10 virtual times) has arrived. good.

Specifically, the evaluation value calculating means 50 detects the "existing agent" existing in the target space from the agent information 120. The existing agent is an agent that is generated before the current virtual time (hereinafter referred to as the current time), has an action purpose ID with a terminal purpose flag of 1, and an achievement flag of 0.
The evaluation value calculation means 50 detects an agent whose status is "achieved" one time before in the agent information 122, so that the action determination timing of the existing agent immediately after the generation time and immediately after the action purpose is achieved arrives. Determine what you have done.
Further, since the difference between the current time and the generation time is a multiple of the action decision interval, it is determined that the action decision timing has arrived for all the existing agents.

See FIG. The destination selection means 51 selects a destination for each agent based on the evaluation value information input from the evaluation value calculation means 50, associates the selected destination with the agent ID, and associates the selected destination with the agent ID, and the agent in the agent information storage means 42. Stored in information 122.
Specifically, for example, the destination pattern _Mij having the highest evaluation value is selected for each agent, and the ID of the first destination in the selected destination pattern _Mij is associated with the agent ID and the current time. Time data is newly generated and added to the agent information 122.
For an existing agent whose current time is not the action decision timing, the destination ID one time ago is newly generated with the agent ID and the current time data associated with the current time, and added to the agent information 122.

Further, for example, a lottery for winning may be performed with a probability according to the height of the evaluation value E of _{each destination pattern Mij, and the winning destination pattern may be selected.}
Further, for example, a lottery for winning may be performed with a probability according to the rank of the evaluation value E of _{each destination pattern Mij, and the winning destination pattern may be selected.} In this case, the relationship between the rank and the probability is determined in advance.

As described above, the method of selecting the destination with the highest evaluation value corresponds to the rational selection action performed by the individual in the real world. On the other hand, the method of selecting a destination by a random lottery corresponds to an intuitive selection action performed by an individual in the real world. A destination may be selected by a method in which a lottery is drawn for each agent and which of these two methods is used. Further, which of these two methods may be used may be assigned to each agent and the destination may be selected according to the assigned method.

See FIG. The action simulating means 52 refers to the spatial information 111 stored in the spatial information storage means 41 and the agent information stored in the agent information storage means 42 at each virtual time t. Based on this information, the behavior simulating means 52 simulates the behavior in which each of the plurality of agents moves to the selected destination and achieves the behavioral purpose.
The action simulating means 52 generates an agent whose generation time has arrived as an existing agent.

The action simulating means 52 reads the agent ID whose generation time matches the current time from the agent information 120, generates the current time data associated with the agent ID, the current time, the initial position, and the status "achievement", and generates the agent. Add to information 122.
The action simulating means 52 deletes the agent whose end purpose flag is 1 and whose action purpose achievement flag is updated to 1. In the simulator 1 of the first embodiment, instead of explicitly deleting the agent, the data after the next virtual time is not added to the agent information 122.

The action simulating means 52 updates the state of the agent (that is, the existing agent) in which the data at the current time is stored from the agent information 122. As described above, for the existing agent, the data at the current time is generated by the destination selection means 51.
The behavior simulating means 52 extracts the data of the existing agent one time before and the data of the current time.

When the situation one time ago is "moving" or "achieved", the action simulating means 52 calculates the moving speed of the agent and updates the position.
Specifically, the behavior simulating means 52 determines the moving speed and position of the agent at the current time. First, the action simulating means 52 reads out the representative point, the movable area, and the waypoint of the destination at the current time t of each corresponding agent from the spatial information 111.

In the movable area, the action simulating means 52 connects the position xi (t-1) one time before the corresponding agent i to the representative point, and the position xi (t-1) via an arbitrary waypoint. Select the shortest route that connects with and the representative point.
Next, the action simulating means 52 aims at a vector whose magnitude is equal to the agent basic speed toward the closest point from the position xi (t-1) of the transit point and the representative point on the selected shortest path. Calculated as the moving speed to approach the ground.

Next, the action simulating means 52 obtains the position xi (t) of the agent i at the current time by adding the calculated movement speed to the position xi (t-1), and updates the agent information 122. At this time, a walking model such as a social force model or RVO (Reciprocal Velocity Obstacles) is applied to the agent to be processed, a wall in the vicinity thereof, or another agent to adjust the movement speed before adding. Is preferable.

Then, for the agent whose current time position does not match the representative point of the destination, the current time status in the agent information 122 is updated to "moving".
On the other hand, for agents whose current time position matches the representative point of the destination (agents who have reached the destination), the destination queue (queue of agents staying at the destination) and Update the status.

First, the queue of the destination at the current time and the number of service counters are read from the spatial information 111. The spatial information 111 is updated by adding the ID of the agent that has reached the destination to the read queue.
At this time, if the number of agents in the queue after addition is less than or equal to the number of service counters, the current time status in the agent information 122 is updated to "processing".
Further, the service time distribution related to the destination is read from the spatial information 111, and the box-Muller method is applied to the read normal distribution to calculate the service time at the destination.

That is, a random number is generated, and the service time is randomly determined with a probability of occurrence according to the normal distribution by lottery using the read normal distribution and the random number.
The set of the ID of the agent that has reached the destination and the service completion time obtained by adding the calculated service time to the current time is added to the agent information 123.
On the other hand, if the number of agents in the queue after addition is larger than the number of service counters, the current time status in the agent information 122 is updated to "waiting".

For the agent (waiting agent) whose situation one hour ago is "waiting", the action simulating means 52 sets the moving speed at the current time to 0 and maintains the position one hour ago.
Further, when the order of the waiting agents in the queue becomes less than or equal to the number of service counters due to the processing of other agents, the waiting agent status is updated to "processing". Otherwise, the situation remains "waiting".

For the agent (processing agent) whose status one time ago is "processing", the action simulating means 52 reads out the service completion time from the agent information 123.
When the current time reaches the service completion time, the action simulating means 52 updates the current time status in the agent information 122 to "achieved". Further, referring to the agent information 121 and the spatial information 111, the current time among the action purposes corresponding to the destination of the current time when the generation time of the agent who has reached the service completion time in the agent information 121 is equal to or less than the current time. Update the achievement flag of the action purpose closest to to 1. In addition, the data of the service completion time of the agent is deleted from the agent information 123.

When the current time reaches the service completion time, the behavior simulation means 52 further indicates that the agent who has reached the service completion time of the spatial information 111 is later than the agent in the queue of the destination to which the service is being serviced. Updates are performed to shift the order of agents in the order forward by one, and the agent ID of the agent is deleted from the same queue. Further, the action simulating means 52 calculates the service time for the agent whose order matches the number of service counters in the updated queue, and adds the combination of the agent ID and the service completion time to the agent information 123. , Update the current time status of the agent in the agent information 122 to "processing".

See FIG. The statistical information updating means 53 updates the statistical information 130 (waiting number distribution, service time distribution) and the statistical information 131 (moving speed distribution) based on the observed values of the required time related to each destination.
The statistical information updating means 53 may update the statistical information 130 and 131 of each agent according to the required time of the agent. In other words, it is considered that the agent has observed its own required time.
That is, the statistical information 130 and 131 stored for the target agent may be updated according to the observed value of the time required for the agent to be processed (target agent) among the plurality of agents to achieve the action purpose.

Further, the statistical information updating means 53 may update the statistical information 130 and 131 of each agent according to the value obtained by the agent observing the required time of an agent other than the agent (hereinafter, another agent).
That is, the statistical information 130 and 131 of the target agent may be updated according to the observed value regarding the time required for the other agent to achieve the action purpose.

The statistical information updating means 53 acquires observation values for a destination existing within the visible range of each agent and agents existing within the same range. Therefore, the statistical information updating means 53 sets the viewing range of each agent. The angle range and the viewing distance of the viewing range may be changed according to the situation of each agent. For example, the visible range during standby or processing may be set wider than the visible range during movement. The angle range and the viewing distance of the viewing range may be changed according to the moving speed of each agent. Further, referring to the spatial information 111, the viewing range is set by excluding the area (blind spot) blocked by obstacles such as walls.
Specifically, the statistical information updating means 53 centers on the area excluding the blind spot from the range of 90 degrees to the left and right of the moving direction for the moving agent, and the agent position for the waiting or processing agent. The area excluding the blind spot from the range of 360 degrees is set as the viewing range.

The statistical information, the observed values for updating it, and the observation timing are as follows.
The observed values for updating the movement speed distribution stored as knowledge for each local route of each agent are the movement speed of the agent and the movement speed of other agents in the local route, and are predetermined observation intervals (for example, 10). Observed at virtual time).
In the present embodiment, the observation interval common to all agents is set, but different observation intervals may be set for each agent.
Further, in the present embodiment, the starting point of the observation interval of each agent is set as the generation time of the agent. It may be a common starting point for all agents.

The observed value for updating the waiting number distribution stored as knowledge for each destination of each agent is the waiting number at the destination, and is observed by referring to the queue at the above observation interval.
The observed value for updating the service time distribution stored as knowledge for each destination of each agent is the service time required for the agent to receive the service at the destination, and the agent is at the destination. It is observed when the action purpose of is achieved.

Alternatively, the service time of another agent at the destination where the agent is waiting may be used as the observed value. In this case, the observed value may be observed when the rank of the agent in the queue of the destination changes. In this case, for example, the behavior simulating means 52 notifies the statistical information updating means 53 of the agent whose queue has changed and the destination, and the notified statistical information updating means 53 stores the notification time in the storage unit 4. Then, the elapsed time from the previous notification is calculated as the service time.

In addition, the statistical information updating means 53 simulates how each individual adds a destination and a local route for the first time to knowledge.
That is, the statistical information updating means 53 statistically collects information on destinations that are within the visible range of each agent but are not included in the statistical information 130 of the agent and information on local routes that are not included in the statistical information 131 of the agent. Add to information 130.

Specifically, the statistical information updating means 53 sets the agent ID of the agent and the destination of the destination when there is a destination not included in the statistical information 130 of the agent within the visible range of each agent. Data in which the initial values of the parameters of the joint probability distribution of the waiting number distribution and the service time distribution are associated with the combination with the ID is added to the statistical information 130. As described above, this initial value is stored in the statistical information storage means 43 as an initial value for each destination.
Further, the statistical information updating means 53 is a combination of the agent ID of the agent and both end points of the local route when there is a local route not included in the statistical information 131 of the agent within the visible range of each agent. The data associated with the initial values of the parameters of the joint probability distribution of the moving speed distribution is added to the statistical information 131. As described above, this initial value is stored in the statistical information storage means 43 as an initial value for each destination.

The update of the parameters of the simultaneous probability distribution (mean value μ, degree of freedom ν, hypothetical sample number κ, scale parameter λ) by the statistical information updating means 53 is, for example, Bayes represented by the following equations (5) to (8). It may be done by an estimation method.

In the above equation, μ ₀ , ν ₀ , κ ₀ , and λ ₀ are parameters of the simultaneous probability distribution stored in the statistical information storage means 43, and μ ₁ , ν ₁ , κ ₁ , and λ ₁ are the updated parameters. This is the simultaneous probability distribution parameter. N is the number of observed values, and x _a is the average value of the observed values.

Next, referring to (a) of FIG. 12 and (b) of FIG. 12, the probability distribution stored in the statistical information storage means 43 is updated by the statistical information updating means 53, and a state in which a difference occurs for each agent is observed. explain.
The probability distributions 140 and 150 indicate the initial values of the service time distributions of the shop 2 of the agents # 1 and # 59. The initial values of agents # 1 and # 59 are both Ν (600, 2500).

The probability distributions 141 and 151 show the service time distributions of the shop 2 of the agents # 1 and # 59 after the virtual time has elapsed and the statistical information update means 53 has sufficiently updated the probability distributions 141 and 151. The service time distribution 141 of agent # 1 is Ν (480, 25700), and the service time distribution 151 of agent # 59 is Ν (550, 17100). That is, the service time distribution 151 has a larger average value and a smaller variance than the service time distribution 141.

For this reason, the kiosk 2 for agent # 1 is, on average, faster than the kiosk 2 for agent # 59, but is often significantly faster or significantly delayed than that average. It turns out that it is a stable store. As a result, for example, when the remaining time is short and the service time of the other shop is shorter on average, the agent # 1 will select the shop 2 in the first place.
On the other hand, it can be seen that the shop 2 for agent # 59 is a shop that is slightly slower on average than the shop 2 for agent # 1, but the service time is stable. As a result, for example, agent # 59 can easily select the shop 2 when the remaining time is 550 hours or more.

In this way, even if the destination is the same, the information on the required time differs depending on the agent. Similarly, even for the same agent, there is a difference in the information on the required time between the destinations for achieving the same action purpose.
It is possible to make a difference in behavior between agents due to a difference in the information of the required time, and it is possible to simulate that a difference in behavior is caused by a difference in experience and knowledge of each individual in the real world.

See FIG. The prediction result output means 60 outputs the prediction result which is the result of the simulation. The prediction result includes at least a part or all of the information stored in the agent information storage means 42, and preferably further includes a part or all of the information stored in the spatial information storage means 41.
The prediction result output means 60 is realized, for example, in collaboration with the control unit 5 and the display unit 6, and causes the display unit 6 to display the prediction result read from the storage unit 4. Further, the prediction result output means 60 may output the prediction result as a data file, may be output as an image file, may be displayed as an image, and may be displayed in one or more of these formats. You may output it.

For example, the prediction result output means 60 creates an image of the prediction result by imaging a map of the target space and superimposing and drawing a figure representing the agent on the coordinates corresponding to the position of each agent on the image. The image is output as a file and / or is output to the display unit 6.
The output timing of the prediction result may include the middle of the simulation in addition to the time when the simulation is completed. That is, a part or all of the above information obtained up to the middle of the simulation may be output as an intermediate prediction result.

(motion)
Next, a simulation method using the simulator 1 of the first embodiment will be described.
FIG. 13 is a flowchart of an example of the overall processing of the simulation method by the simulator 1 of the first embodiment.
First, in step S1, the user sets simulated conditions, spatial information, initial values of agent information, and initial values of statistical information as simulation conditions. The setting is performed through the condition setting means 20. Then, in step S2, the virtual time t is initialized to 0.

In step S3, an agent whose generation time has been reached is generated. Specifically, the behavior simulating means 52 searches the agent information 120 for an agent for which the same generation time as the current time t is set. When the corresponding agent is detected, the ID of the agent is added to the agent information 122 together with the initial position. The added agent becomes an existing agent.
In step S4, it is determined whether or not there is an existing agent (an agent that is generated before the virtual time and has an action purpose ID whose achievement flag is 0). When the existing agent exists (step S4: Y), the process proceeds to step S5. If the existing agent does not exist (step S4: N), the process proceeds to step S8.

In step S5, the statistical information update process is performed. In the statistical information update process, the statistical information 130 and 131 stored in the statistical information storage means 43 are updated based on the observed values observed by the existing agent. The details of the statistical information update process will be described later.
In step S6, the action determination process is performed. In the action determination process, the destination to be reached by the agent is selected based on the action purpose set in the agent information 121 and the probability distribution set in the statistical information 130 and 131. The details of the action decision process will be described later.

In step S7, the agent state update process is performed. In the agent state update process, the state of the existing agent is updated according to the destination determined in step S6. Details of the agent status update process will be described later.
In step S8, the virtual time t is increased by "1".
In step S9, it is determined whether or not the virtual time t has reached the simulation end time T stored in the simulated condition storage means 40. When the virtual time t reaches the end time T (step S9: Y), the process proceeds to step S10. When the virtual time t has not reached the end time T (step S9: N), the process returns to step S3.
In step S10, the prediction result, which is the result of the simulation, is output, and the process ends.

Next, the statistical information update process in step S5 will be described with reference to FIGS. 14 and 15. The statistical information update process is executed by the behavior simulation means 52 and the statistical information update means 53.
In step S20, the action simulating means 52 sequentially selects existing agents and sets them as the agents of interest to be processed.

In step S21, the behavior simulating means 52 determines whether or not the observation timing of the observed values for updating the statistical information 130 and 131 has arrived with respect to the agent of interest. When the observation timing arrives (step S21: Y), the process proceeds to step S22. When the observation timing has not arrived (step S21: N), the process proceeds to step S35.
In step S22, the statistical information updating means 53 sets the viewing range of the agent of interest. The statistical information updating means 53 determines whether or not a destination not included in the statistical information 130 of the agent of interest exists within the visible range.

When the destination not included in the statistical information 130 exists within the visible range (step S22: Y), the process proceeds to step S23. When the destination not included in the statistical information 130 does not exist within the visible range (step S22: N), the process proceeds to step S24.
In step S23, the statistical information updating means 53 adds the data of the destination to the statistical information 130 of the agent of interest. After that, the process proceeds to step S24.

In step S24, the statistical information updating means 53 determines whether or not a local route that is not included in the statistical information 131 of the agent of interest exists within the visible range. When the local route not included in the statistical information 131 exists within the visible range (step S24: Y), the process proceeds to step S25. When the local route not included in the statistical information 131 does not exist within the visible range (step S24: N), the process proceeds to step S26.
In step S25, the statistical information updating means 53 adds the data of the local route to the statistical information 131 of the agent of interest. After that, the process proceeds to step S26.

In step S26, the statistical information updating means 53 determines whether or not the status of the agent of interest is “moving”. When the status of the attention agent is “moving” (step S26: Y), the process proceeds to step S27. If the status of the agent of interest is not "moving" (step S26: N), the process proceeds to step S28.
In step S27, the statistical information updating means 53 acquires the moving speed and the local route of the agent of interest, and calculates the speed in the local route. After that, the process proceeds to step S28.

In step S28, the statistical information updating means 53 determines whether or not the status of another agent existing within the visible range of the agent of interest is “moving”. When the status of the other agent within the viewing range is "moving" (step S28: Y), the process proceeds to step S29. If the status of the other agent within the viewing range is not "moving" (step S28: N), the process proceeds to step S30.
In step S29, the statistical information updating means 53 acquires the moving speed and the local route of another agent that is moving and exists within the visible range, and calculates the speed in the local route. After that, the process proceeds to step S30.

In step S30, the statistical information updating means 53 determines whether or not the speed is calculated in either step S27 or S29. When the speed is calculated (step S30: Y), the process proceeds to step S31. If the speed is not calculated (step S30: N), the process proceeds to step S32.
In step S31, the statistical information updating means 53 uses the calculated speed as an observed value to update the moving speed distribution of the local route in which the speed is observed in the statistical information 131 of the agent of interest. After that, the process proceeds to step S32.

In step S32, the statistical information updating means 53 determines whether or not any of the destinations exists within the visible range of the agent of interest. When the destination exists within the viewing range (step S32: Y), the process proceeds to step S33. If the destination does not exist within the viewing range (step S32: N), the process proceeds to step S35.
In step S33, the statistical information updating means 53 acquires the number of waits for the destination existing within the visible range. In step S34, the statistical information updating means 53 updates the waiting number distribution of the destination where the waiting number is observed in the statistical information 130 of the attention agent by using the acquired waiting number as an observed value. After that, the process proceeds to step S35.

In step S35, the statistical information updating means 53 determines whether or not the service completion time of the attention agent has arrived. When the service completion time arrives (step S35: Y), the process proceeds to step S36. If the service completion time does not arrive (step S35: N), the process proceeds to step S37.
In step S36, the statistical information updating means 53 calculates the service time required for the agent of interest. After that, the process proceeds to step S37.

In step S37, the statistical information updating means 53 determines whether or not the order in the queue has been updated if the status of the agent of interest is waiting. When the order is updated (step S37: Y), the process proceeds to step S38. If the order is not updated (step S37: N), the process proceeds to step S39.
In step S38, the statistical information updating means 53 calculates the service time required for another agent whose service has been completed at the destination waiting for the service of the agent of interest. After that, the process proceeds to step S39.

In step S39, the statistical information updating means 53 determines whether or not the service time has been calculated in either step S36 or S38. When the service time is calculated (step S39: Y), the process proceeds to step S40. When the service time is not calculated (step S39: N), the process proceeds to step S41.
In step S40, the statistical information updating means 53 uses the calculated service time as an observed value, and uses the service time of the destination (that is, the destination where the attention agent receives the service) in which the service time is observed in the statistical information 130 of the attention agent. Update the distribution. After that, the process proceeds to step S41.
In step S41, the statistical information updating means 53 determines whether or not the processes of steps S21 to S40 have been executed for all the existing agents. When the processing is executed for all the existing agents (step S41: Y), the statistical information update processing ends. If there is an unprocessed existing agent (step S41: N), the process returns to step S20.

Next, the action determination process in step S6 will be described with reference to FIG. The action determination process is executed by the evaluation value calculation means 50 and the destination selection means 51.
In step S50, the evaluation value calculating means 50 sequentially selects existing agents and sets them as the agents of interest to be processed.
In step S51, the evaluation value calculating means 50 determines whether or not the action determination timing of the attention agent has arrived. When the action determination timing has arrived (step S51: Y), the process proceeds to step S52. When the action determination timing has not arrived (step S51: N), the process proceeds to step S59.

Evaluation value calculating means 50 in step S52 generates one or more behavioral patterns _{P i.}
In step S53, the evaluation value calculating means 50 generates one or a plurality of destination patterns M _{ij for each of the action patterns P i.}
In step S54, the evaluation value calculating means 50 _{sequentially selects the destination pattern Mij} and sets it as the attention pattern to be processed.

In step S55, the evaluation value calculation means 50 performs the evaluation value calculation process. In the evaluation value calculation process, the evaluation value E is calculated for the pattern of interest. The details of the evaluation value calculation process will be described later.
In step S56, the evaluation value calculating means 50 _{determines whether or not all the destination patterns Mij} are set as the attention patterns. When all the destination patterns _Mij are set as the attention patterns (step S56: Y), the process proceeds to step S57. _{When the destination pattern Mij} that is not set in the attention pattern exists (step S56: N), the process returns to step S54.

In step S57, the destination selection means 51 selects one of the destination patterns _Mij based on the evaluation value.
In step S58, the destination selection means 51 _{newly generates current time data in which the ID of the first destination in the selected destination pattern Mij} is associated with the agent ID and the current virtual time, and agent information. Add to 122. After that, the process proceeds to step S60.

On the other hand, in step S59, the destination selection means 51 newly generates current time data in which the destination ID one time ago is associated with the agent ID and the current virtual time, and adds it to the agent information 122. After that, the process proceeds to step S60.
In step S60, the evaluation value calculating means 50 determines whether or not the processes of steps S51 to S59 have been executed for all the existing agents. When the processes are executed for all the existing agents (step S60: Y), the action decision process ends. If there is an unprocessed existing agent (step S60: N), the process returns to step S50.

The evaluation value calculation process in step S55 will be described with reference to FIG. The evaluation value calculation process is executed by the evaluation value calculation means 50.
In step S70, the evaluation value calculating means 50 calculates the required time distribution, which is the probability distribution of the required time when the action purpose is achieved at the destination, for each destination included in the attention pattern.
In step S71, the evaluation value calculating means 50 initializes the variable s indicating the rank of the destination included in the attention pattern to “1”. In the following steps S73 to S77, the success probability of achieving the action purpose by the achievement deadline at the sth destination included in the attention pattern is calculated. Further, the evaluation value calculating means 50 initializes the evaluation value E to “0”.

In step S72, the evaluation value calculating means 50 determines whether or not the variable s is "1", that is, whether or not to calculate the success probability of the first destination. When the variable s is "1" (step S72: Y), the process proceeds to step S73. If the variable s is not "1" (step S72: N), the process proceeds to step S75.
In step S73, the evaluation value calculating means 50 calculates the remaining time until the achievement deadline of the action purpose to be achieved at the first destination.

In step S74, the evaluation value calculating means 50 calculates the success probability based on the remaining time and the action purpose required time distribution at the first destination. After that, the process proceeds to step S78.
On the other hand, in step S75, the evaluation value calculating means 50 calculates the total time distribution required to achieve the action purpose at the first to s-1st destinations.

In step S76, the evaluation value calculating means 50 calculates the remaining time distribution until the achievement deadline of the action purpose to be achieved at the sth destination based on the total of the required time distributions.
In step S77, the evaluation value calculating means 50 calculates the success probability based on the remaining time distribution and the required time distribution of the sth destination. After that, the process proceeds to step S78.
In step S78, the evaluation value calculating means 50 adds the product obtained by multiplying the desired achievement degree by the success probability calculated in step S74 or S77 to the evaluation value E.

In step S79, the evaluation value calculating means 50 increases the variable s by "1".
In step S80, the evaluation value calculating means 50 determines whether or not the processing of steps S72 to S79 has been completed for all the destinations in the pattern of interest. When the processing is completed for all the destinations (step S80: Y), the processing proceeds to step S81. If there is an unprocessed destination (step S80: N), the process returns to step S72.
In step S81, the evaluation value calculating means 50 outputs the evaluation value information in which the agent, the attention pattern, and the evaluation value are associated with each other to the destination selection means 51. After that, the evaluation value calculation process ends.

Next, the agent state update process in step S7 will be described with reference to FIG. The agent state update process is realized by the destination selection means 51 and the action simulation means 52.
In step S90, the action simulating means 52 sequentially selects existing agents and sets them as the agents of interest to be processed.
In step S91, the action simulating means 52 determines whether or not the destination selection means 51 newly selects the destination pattern of the agent of interest. When a new destination pattern is selected (step S91: Y), the process proceeds to step S92. When the destination pattern is not newly selected (step S91: N), the process proceeds to step S93.

In step S92, the destination selection means 51 updates the destination ID stored as an action parameter in the agent information 122 to the destination ID at the head of the newly selected destination pattern. After that, the process proceeds to step S93.
In step S93, the action simulating means 52 updates the status of the attention agent.

In step S94, the behavior simulating means 52 determines whether or not the status of the agent of interest is “moving” or “achieved”. When the situation is "moving" or "achieved" (step S94: Y), the process proceeds to step S95. If the situation is not "moving" or "achieved" (step S94: N), the process proceeds to step S96.
In step S95, the action simulating means 52 calculates the moving speed of the agent of interest and updates the position. After that, the process proceeds to step S96.

In step S96, the action simulating means 52 determines whether or not the attention agent has achieved all the action objectives (that is, whether or not the action objective for which the terminal purpose flag is 1 has been achieved). When all the action objectives have been achieved (step S96: Y), the process proceeds to step S97. When there is an unachieved action purpose (step S96: N), the process proceeds to step S98.
In step S97, the behavior simulating means 52 erases the attention agent. After that, the process proceeds to step S98.
In step S98, the action simulating means 52 determines whether or not the processing of steps S91 to S97 has been completed for all the existing agents. If there is an unprocessed existing agent (step S98: N), the process returns to step S90. When the processing for all the existing agents is completed (step S98: Y), the agent status update processing ends.

(Modification example)
(1) In the first embodiment, the required time is defined as the sum of the travel time, the waiting time, and the service time.
Instead, the travel time may be the required time (ie, the waiting time and the service time are 0). This makes it possible to simulate an action purpose such as delivery in which the waiting time and the service time can be regarded as zero.

(2) In the first embodiment, the case where the relationship between the action purpose and the destination is a one-to-one relationship and a one-to-many relationship is illustrated. The relationship between the purpose of action and the destination may be a many-to-one or many-to-many relationship. For example, it is possible to simulate a situation in which the service time differs between different services at the same destination. In this case, it is desirable to store or update the statistical information of the service time distribution for each combination of the action purpose and the destination.

(3) In the first embodiment, an example in which the evaluation value calculation by the evaluation value calculation means 50 and the destination selection by the destination selection means 51 are repeated at the action determination interval has been described. Instead of this, it may be omitted to repeat the evaluation value calculation and the destination selection at the action determination interval.
In this case, for example, the destination pattern selected by the destination selection means 51 is stored in the agent information storage means 42, and the action simulation means 52 sequentially corresponds to the first destination of the destination pattern stored for each agent. It may act to move the agent.

(4) In the first embodiment, an example of setting an achievement deadline for each action purpose has been described. Alternatively or additionally, a single achievement deadline (ie, a common achievement deadline for multiple action objectives) may be set for each inter-curtain and for each agent. For example, the end time between intermissions may be the deadline for achievement. In this case, it is assumed that the same achievement deadline is set for all the action objectives of the agents between the intermissions, and the same processing as in the above embodiment may be performed.
Furthermore, a single deadline common to all agents may be set for each inter-curtain.

(5) In the first embodiment, when there is a destination that is not included in the statistical information 130 of the agent within the visible range of the agent, the data of the destination is added to the statistical information 130 of the agent. Alternatively or additionally, an information source such as a signboard that informs the agent of information about the destination may be set in the target space. If there is an information source that informs the destination that is not included in the statistical information 130 of the agent within the visible range of the agent, the data of the destination may be added to the statistical information 130 of the agent. Similarly, an announcement that broadcasts information about the destination in the target space may be simulated. When an announcement is broadcast that informs a destination that is not included in the agent's statistical information 130, the data of the destination may be added to the agent's statistical information 130.

(Second Embodiment)
When calculating the required time distribution, the evaluation value calculating means 50 of the first embodiment predicts the waiting time based on the waiting number distribution of the destination stored in the statistical information 130. For destinations within the visible range, the waiting time is predicted from the observed values of the number of waits.
However, observing the movements of other agents towards their destination can change the wait time prediction.
Therefore, the evaluation value calculating means 50 of the second embodiment predicts the waiting time at the destination based on the actions of other agents when selecting the destination pattern of the agent to be processed. For example, the waiting time at the destination is predicted based on the moving direction of another agent.

See FIG. Now, the reference code 200 indicates the attention agent of the processing target for calculating the waiting number distribution (that is, the attention agent of the action determination processing of FIG. 16), and the broken line indicates the visual range of the attention agent 200.
Reference numerals 201a to 201c indicate other agents within the visible range of the agent of interest 200, and reference numeral 202 indicates other agents outside the visible range.

Further, reference numerals 203a, 204b, and 205a indicate agents receiving services at the destinations VI, VII, and VIII, respectively, reference numerals 203b and 203c indicate agents waiting for services at the destination VI, and reference numerals 204b are objects. Agents waiting for service at destination VII, reference numerals 205b and 205c indicate agents waiting for service at destination VIII.
The evaluation value calculating means 50 of the second embodiment predicts the number of waits for the destinations VI, VII, and VIII based on the moving directions of the other agents 201a to 201c within the visible range.

The evaluation value calculation means 50 of the second embodiment includes the waiting time calculation means 70 shown in FIG. The waiting time calculating means 70 includes an agent selecting means 71, a viewing range setting means 72, a moving direction acquisition means 73, a arrival number predicting means 74, and a waiting time predicting means 75.
The waiting time calculation means 70 selects the agent of interest to be processed from the existing agents.

The viewing range setting means 72 sets the viewing range of the agent of interest. The angle range and the viewing distance of the viewing range may be changed according to the situation of the agent of interest. For example, the visible range during standby or processing may be set wider than the visible range during movement.
For example, if the agent of interest is moving, the range of 90 degrees to the left and right of the moving direction may be set, and if the agent is waiting or processing, the range may be 360 degrees centered on the agent position.
The angle range and the viewing distance of the viewing range may be changed according to the moving speed of each agent. Further, preferably, the spatial information 111 stored in the spatial information storage means 41 is used, and the visible range is set by excluding the area blocked by obstacles such as walls within the visible range.

The movement direction acquisition means 73 acquires the movement speed of another agent within the visible range as the movement direction of the other agent from the agent information 122 stored in the agent information storage means 42.
The arrival number predicting means 74 is the destination d stored in the statistical information 130 of the agent of interest based on the moving direction of the other agent within the visible range and the representative point of the destination stored in the spatial information 111. Predict the number of arrivals that other agents will reach _i. For example, the arrival number predicting means 74 may predict the _{arrival number R di} that another agent arrives at the destination based on the following equation (9).

In the formula (9), a is the other agents that are determined to be heading to the destination d _i stored in the statistical information 130. For example, the arrival number prediction means 74, if the destination d _i is present in within the predetermined angle θ of the moving direction left and right other agents 201b as shown in FIG. 19, the other agents 201b is towards the destination d _i You may judge that it is. Further, in the equation (9), r _a denotes the number of all destinations within a predetermined angle θ of the moving direction lateral of another agent a of the destinations of the stored statistics 130 target agent.

The waiting time prediction means 75 predicts the waiting time distribution for the destination stored in the statistical information 130 of the agent of interest. The waiting time distribution is calculated by dividing the number of future waits at the destination by the number of service counters and the product of the service time distribution required to execute the action purpose at the destination.
When the destination is within the visible range of the agent of interest, the waiting time prediction means 75 sets the sum of the waiting number of the destination and the arrival number of the destination as the future waiting number of the destination. That is, the waiting time distribution is predicted based on the observation result and the number of arrivals of the number of other agents existing at the destination.

When the destination is not within the visible range of the attention agent, the representative value is sampled from the waiting number distribution of the destination stored in the statistical information 130 of the attention agent, and the number of arrivals at the destination is used as the representative value. The sum of the sums is used as the number of future waits for the destination.
That is, the waiting time distribution is predicted based on the number of arrivals at the destination and the statistical information of the waiting time at the destination stored in the statistical information 130.

Next, with reference to FIG. 21, the required time distribution calculation process by the evaluation value calculation means 50 of the second embodiment will be described.
In step S100, the evaluation value calculating means 50 moves to each destination included in the attention pattern (destination pattern) based on the movement speed distribution of the local route stored in the statistical information 131 of the attention agent. Calculate the time distribution.

In step S101, the evaluation value calculating means 50 has a service time distribution at each destination included in the attention pattern (destination pattern) based on the service time distribution at the destination stored in the statistical information 130 of the attention agent. Is calculated.
In step S102, the arrival number predicting means 74 predicts the arrival number of other agents heading for the destination stored in the statistical information 130 of the agent of interest.

In step S103, the waiting time prediction means 75 is stored in the observed value of the number of arrivals of the agent toward the destination and the number of waits at the destination, or the number of arrivals of the agent toward the destination and the statistical information 130 of the agent of interest. Predict the number of future waits at the destination based on the representative value of the number of people waiting at the destination.

In step S104, the waiting time predicting means 75 calculates the waiting time distribution at the destination based on the number of future waits, the number of service counters, and the service time distribution.
In step S105, the evaluation value calculating means 50 calculates the sum of the travel time distribution, the service time distribution, and the waiting time distribution as the required time. After that, the required time distribution calculation process ends.

In each of the above embodiments, an example of holding the waiting number distribution, the service time distribution, the moving speed distribution, and each distribution associated therewith as a normal distribution is shown, but these distributions may be held as a histogram. In that case, the calculation between the distributions is the calculation between the corresponding bins, and the update of the distribution by the observed value is the update of the bin value. Further, each distribution can be held as a formalized distribution other than the normal distribution. In that case, an calculation method or an update method suitable for the distribution to be adopted may be used, such as performing an operation on the distribution by a sampling method.

(Effect of embodiment)
(1) The simulator 1 simulates an action in which an agent having a plurality of action objectives achieves a plurality of action objectives at a plurality of destinations in a predetermined space.
The evaluation value calculation means 50 selects a plurality of combinations of destinations that achieve each of a plurality of action objectives from a plurality of destinations, and achieves a plurality of action objectives at the combined destinations for each combination of destinations. The total probability of achieving each by the deadline is predicted based on the probability density distribution of each time required to achieve a plurality of action objectives, and the evaluation value is calculated according to the total.

The behavior simulating means 52 simulates the behavior of an agent moving so as to sequentially visit a combination of destinations selected based on an evaluation value from a plurality of combinations of destinations.
The evaluation value calculation means 50 determines the probability density distribution of the time required for the action purpose to be achieved first among the plurality of action objectives and the probability density distribution of the remaining time according to the achievement deadline, and achieves the action purpose to be achieved later. The probability to be achieved by is predicted based on the probability density distribution of the remaining time.

As a result, the agent himself judges the required time for the action purpose and flexibly changes the action purpose and the destination by reflecting the bias of the experience and knowledge of each agent regarding the time required for the action purpose. Can be simulated.

(2) The evaluation value calculation means samples a sample of the probability density distribution of the remaining time by a predetermined sampling method, and is based on the sample and the probability density distribution of the time required to achieve the action purpose to be achieved later. Then, the probability of achieving the action purpose to be achieved later by the achievement deadline may be calculated.
As a result, the remaining time for the action purpose to be achieved later among a plurality of action purposes becomes a probability density distribution, and even if the calculation formula of the probability of achieving the action purpose to be achieved later cannot be analytically solved, it is a numerical value. The probability can be calculated analytically.

(3) The evaluation value calculation means is based on the histogram of the remaining time and the cumulative distribution function of the time required to achieve the action purpose to be achieved later, and the frequency and the cumulative distribution function of the histogram in each section of the histogram. The sum of the products with the probabilities may be calculated as the probability of achieving the action objective to be achieved later by the achievement deadline.
As a result, the remaining time for the action purpose to be achieved later among a plurality of action purposes becomes a probability density distribution, and even if the calculation formula of the probability of achieving the action purpose to be achieved later cannot be analytically solved, it is a numerical value. The probability can be calculated analytically.

(4) The evaluation value calculating means 50 may calculate an evaluation value which is the sum of the probabilities of weighting the probabilities of achieving each of the plurality of action objectives by the utility of the plurality of action objectives.
As a result, it is possible to simulate an action of selecting a destination according to the priority of an action purpose different for each agent.

1 ... Simulator, 2 ... Operation input unit, 3 ... File input / output unit, 4 ... Storage unit 5 ... Control unit, 6 ... Display unit, 20 ... Condition setting means, 40 ... Simulated condition storage means, 41 ... Spatial information storage means , 42 ... Agent information storage means, 43 ... Statistical information storage means, 50 ... Evaluation value calculation means, 51 ... Destination selection means, 52 ... Behavior simulation means, 53 ... Statistical information update means, 60 ... Prediction result output means, 70 ... Time calculation means, 71 ... Agent selection means, 72 ... Visual range setting means, 73 ... Movement direction acquisition means, 74 ... Arrival number prediction means, 75 ... Time prediction means, 100 ... Map information, 110 ... Action purpose information, 111 ... Spatial information, 120, 121, 122, 123 ... Agent information, 130, 131 ... Statistical information

Claims (6)

A simulator that simulates an action in which an agent having a plurality of action objectives achieves the plurality of action objectives at a plurality of destinations in a predetermined space.
A combination selection means for selecting a plurality of combinations of destinations that achieve each of the plurality of action objectives from the plurality of destinations.
For each combination of the destinations, the sum of the probabilities of achieving the plurality of action objectives at the combined destinations by the achievement deadline is the probability density of each time required to achieve the plurality of action objectives. An evaluation value calculation means that makes a prediction based on the distribution and calculates an evaluation value according to the total.
It is provided with an action simulating means for simulating the behavior of the agent who moves so as to sequentially visit a combination of destinations selected based on the evaluation value from a plurality of combinations of the destinations.
The evaluation value calculating means determines the probability density distribution of the time required for the action purpose to be achieved first among the plurality of action purposes and the probability density distribution of the remaining time according to the achievement deadline, and determines the action purpose to be achieved later. The probability of achieving the achievement by the deadline is predicted based on the probability density distribution of the remaining time.
A simulator that features that. The evaluation value calculating means extracts a sample of the probability density distribution of the remaining time by a predetermined sampling method, and is based on the sample and the probability density distribution of the time required to achieve the action purpose to be achieved later. The simulator according to claim 1, wherein the probability of achieving the action purpose to be achieved later by the achievement deadline is calculated. The evaluation value calculation means is based on the histogram of the remaining time and the cumulative distribution function of the time required to achieve the action purpose to be achieved later, and the frequency of the histogram and the cumulative number in each section of the histogram. The sum of the products with the probabilities of the distribution function is calculated as the probability of achieving the action objective to be achieved after the above by the achievement deadline.
The simulator according to claim 2. Claims 1 to 3 characterized in that the evaluation value calculating means calculates the evaluation value which is the sum of the probabilities of weighting the probabilities of achieving each of the plurality of action objectives by the utility of the plurality of action objectives. The simulator described in any one of the above. It is a simulation method that simulates an action in which an agent having a plurality of action objectives achieves the plurality of action objectives at a plurality of destinations in a predetermined space.
A combination selection process for selecting a plurality of combinations of destinations that achieve each of the plurality of action objectives from the plurality of destinations.
For each combination of the destinations, the sum of the probabilities of achieving the plurality of action objectives at the combined destinations by the achievement deadline is the probability density of each time required to achieve the plurality of action objectives. Evaluation value calculation processing that predicts based on the distribution and calculates the evaluation value according to the total,
An action simulation process that simulates the behavior of the agent who moves so as to sequentially visit a combination of destinations selected based on the evaluation value from a plurality of combinations of the destinations.
The computer runs,
In the evaluation value calculation process, the probability density distribution of the time required for the action purpose to be achieved first among the plurality of action purposes and the probability density distribution of the remaining time according to the achievement deadline are determined, and the action purpose to be achieved later is determined. A simulation method characterized in that the probability of achievement by the achievement deadline is predicted based on the probability density distribution of the remaining time. It is a simulation program that simulates an action in which an agent having a plurality of action objectives achieves the plurality of action objectives at a plurality of destinations in a predetermined space.
A combination selection process for selecting a plurality of combinations of destinations that achieve each of the plurality of action objectives from the plurality of destinations.
For each combination of the destinations, the sum of the probabilities of achieving the plurality of action objectives at the combined destinations by the achievement deadline is the probability density of each time required to achieve the plurality of action objectives. Evaluation value calculation processing that predicts based on the distribution and calculates the evaluation value according to the total,
An action simulation process that simulates the behavior of the agent who moves so as to sequentially visit a combination of destinations selected based on the evaluation value from a plurality of combinations of the destinations.
Let the computer run
In the evaluation value calculation process, the probability density distribution of the time required for the action purpose to be achieved first among the plurality of action purposes and the probability density distribution of the remaining time according to the achievement deadline are determined, and the action purpose to be achieved later is determined. A simulation program characterized in that the probability of achievement by the achievement deadline is predicted based on the probability density distribution of the remaining time.

Images