Evacuation behavior prediction device, method, and program

Abstract

To provide an evacuation action prediction technique which facilitates analysis of the decision grounds of evacuation actions taken by an evacuation agent.SOLUTION: An evacuation action prediction device provided herein is configured to retain a map 16 showing evacuation routes 26 of agents 21 located in an area where an event has occurred, reflect parameters 28 on coordinate values 27 of the map 16, set an initial position 22 and target position 25 of each of multiple evacuating agents 21 on the coordinate values 27 of the map 16, and simulate evacuation routes 26 of the agents 21.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] Embodiments of the present invention relate to a technique for predicting the evacuation behavior of agents when an event such as a disaster occurs.

Background Art

[0002] In recent years, in order to cope with natural disasters and the like that are increasing, a technique for predicting evacuation behavior during disasters has been demanded. And, according to the characteristics of individuals such as age, it has been demanded to move on to early evacuation to ensure the safety of each individual. Furthermore, for wide-area disasters, it has been demanded to efficiently conduct evacuation guidance with limited resources.

[0003] Various systems for formulating an evacuation plan assuming the occurrence of a disaster have been proposed so far. As one example, a technique for simulating an optimal evacuation route with the shortest evacuation completion time according to the assumed disaster situation and formulating a quick and appropriate evacuation plan is known.

[0004] As a second example, a technique for simulating the progress of a disaster and the behavior of people on an evacuation route based on the assumed disaster and the structure of the evacuation route is known. As a third example, a technique for simulating disaster evacuation in high-rise buildings using a multi-agent system is known.

[0005] As a fourth example, a technique for simulating evacuation information using a trained model that has been reinforced with a reward according to the number of deaths and evacuation time is known. By using such simulation technology, it is possible to predict evacuation behavior for each assumed disaster scenario. And, based on such prediction of evacuation behavior, it is possible to evaluate the validity of an evacuation plan, conduct evacuation training and education, and also conduct layout design of buildings and cities with high safety that are easy to evacuate.

Prior Art Documents

Patent Documents

[0006] [Patent Document 1] Japanese Patent No. 5996689 [Patent Document 2] Japanese Patent No. 4822812 [Patent Document 3] Japanese Patent No. 5372421 [Patent Document 4] Japanese Unexamined Patent Application Publication No. 2021-047625 [Summary of the Invention] [Problems to be Solved by the Invention]

[0007] By the way, when formulating an evacuation plan, it is necessary to consider event conditions such as fires, environmental conditions such as temperature, the behavior conditions of evacuees, the cognitive results during the evacuation, and the interaction with other evacuees. Then, based on these considerations, condition settings are made, simulations are executed, and an evacuation plan that minimizes the evacuation time and the number of injured people is proposed. However, when there are many conditions to be considered in this way, there are problems such that the basis for the evacuation actions selected by each of the evacuation entities (agents) becomes unclear, or it becomes difficult to identify the conditions that affect the results of the executed simulations.

[0008] Also, in the formulation of an evacuation plan that minimizes the evacuation time using reinforcement learning as in the fourth example described above, the magnitude of future rewards becomes the basis for the selection of evacuation actions. However, in calculating such a basis, it is necessary to define and set in advance the relationship between the selection of evacuation actions and each state quantity based on many conditions (factors), which is practically difficult.

[0009] The embodiments of the present invention have been made in consideration of such circumstances, and an object thereof is to provide an evacuation behavior prediction technique that facilitates the analysis of the basis for the evacuation actions taken by an evacuation entity (agent). [Means for Solving the Problems]

[0010] In the evacuation behavior prediction device according to the embodiment, a holding unit that holds a map representing an evacuation route of an agent located in an area where an event has occurred, a reflecting unit that reflects a parameter indicating the contribution degree of the event to the determination of the evacuation route on the coordinate values of the map, a setting unit that sets the initial position and the target position of each of a plurality of the agents evacuating from the area to the coordinate values of the map, a first simulator that simulates the evacuation route from the initial position to the target position for each of the agents without applying the parameter, a second simulator that commonly applies the parameter and simulates the evacuation route of each of the agents, a third simulator that individually applies the parameter and simulates the evacuation route of each of the agents, and a collection unit that collects the evacuation routes of each of the agents determined by the first, second, and third simulators.

Advantages of the Invention

[0011] According to the embodiment of the present invention, an evacuation behavior prediction technique is provided that facilitates the analysis of the basis for determining the evacuation behavior taken by an evacuation subject (agent).

Brief Description of the Drawings

[0012]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Embodiments for Carrying Out the Invention

[0013] (First Embodiment) Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. FIG. 1 is a block configuration diagram of an evacuation behavior prediction device 10A (10) according to the first embodiment of the present invention. FIG. 2 is a plan view of an area 20 to which each embodiment of the present invention is applied. The evacuation behavior prediction device 10 of each embodiment predicts the actions and evacuation behaviors of people and objects (such as machines) at the time of an event such as a disaster, and utilizes the technology of multi-agent simulation. Note that multi-agent simulation is a method of predicting events created by arranging a large number of models (agents 21) simulating people and objects in a virtual space and making them interact with each other.

[0014] As shown in FIG. 1, the evacuation behavior prediction device 10A (10) includes agents 21 (211, 212... 21 m …21 n …) located in the area 20 where an event 24 (for example, a fire; see FIG. 2) has occurred, a holding unit 17 that holds a map 16 (16a, 16b, 16c) representing the evacuation routes 26 (FIG. 3) of the agents, a reflection unit 15 (15b, 15c) that reflects a parameter 28 indicating the contribution degree of the event 24 to the determination of the evacuation route 26 on the coordinate values 27 of the map 16, a setting unit 35 that sets the initial positions 22 and target positions 25 of each of the plurality of agents 21 evacuating the area 20 to the coordinate values 27 of the map 16, and a simulator 30 (30a, 30b, 30c) that simulates the evacuation routes 26 of the agents 21.

[0015] Here, the first simulator 30a simulates an evacuation route 26 from each initial position 22 to the target position 25 of the agents 21 in the first map 16a where the parameter 28 is not applied. That is, the simulation in the first simulator 30a simulates the evacuation route 26 when all the agents 21 located in the area 20 are made to take evacuation actions under the assumption that the event 24 does not actually occur.

[0016] FIG. 3(A) is a conceptual diagram of a second map 16b (map 16) in which the evacuation route 26 of the agent 21 m is second-simulated. FIG. 3(B) is similarly a conceptual diagram of a second map 16c (map 16) in which the evacuation route 26 of another agent 21 n is second-simulated. Thus, the second simulator 30b simulates the respective evacuation routes 26 of the agents 21 (21 m , 21 n ) in the second map 16b to which the parameter 28 is commonly applied.

[0017] That is, the simulation in the second simulator 30b means that the parameter 28 is common for use in determining the evacuation route 26 without depending on the respective initial positions 22 of the agents 21 (21 m , 21 n ). Thereby, even when the event 24 occurs and when a new event 24 occurs thereafter, all the agents 21 (21 m , 21 n ) can share all the grasped information simultaneously and determine the evacuation route 26.

[0018] The third simulator 30c (FIG. 1) simulates the respective evacuation routes 26 of the agents 21 in the third map 16c to which the parameter 28 is individually applied. That is, the simulation in the third simulator 30c means that the parameter 28 is different for use in determining the evacuation route 26 depending on the respective initial positions 22 of the agents 21.

[0019] As a result, when Event 24 occurs and when a new Event 24 occurs after that, each of Agents 21 will make a judgment based on information with different levels of understanding and determine the evacuation route 26. That is, the evacuation route 26 of Agent 21 is simulated using information directly perceived by itself during the evacuation process, such as the location of a fire or the presence or absence of rubble, which is information not perceived by other Agents 21.

[0020] FIG. 4 is a conceptual diagram of a third map 16c (map) for explaining an example of the third simulation of the evacuation route 26 of Agent 21. As described above, in the third simulation, the coordinate values 27 of a predetermined range 29 can be recognized based on the position of Agent 21 and the parameters 28 reflected therein can be applied. m of Agent 21 m Returning to FIG. 1 to continue the explanation. The registration unit 40 stores an event scenario 41 and environmental conditions 42. The event scenario can specify the occurrence of any Event 24, such as the occurrence time of disasters such as earthquakes and fires, the time when a machine fails, or the time when a person comes from the outside, in time steps (time series). In addition to the occurrence time of these Events 24, the occurrence location can also be specified.

[0021] The environmental conditions 42 are set with temperature, humidity, spatial dose, carbon monoxide concentration, etc. at any time and location. These environmental conditions 42 can be set by the user or imported from the outside. Then, the generation unit 47 of the simulator 30 updates the parameters 28 reflected in the coordinate values 27 each time a new Event 24 or the environment changes in time steps according to the time series settings of the event scenario 41 and the environmental conditions 42.

[0022]

[0023] ​The setting unit 35 is provided with an identifier 46 for identifying each of the agents 21 and a behavior model 45 involved in the algorithm for simulating the evacuation route 26. Here, in the behavior model 45, the behavior algorithms of the agents 21 identified by the respective identifiers 46 are defined. In this behavior model 45, attribute information such as gender, age, proficiency, and job position, as well as state quantities such as fatigue level and injury level, are defined together, and the behavior algorithm is determined according to this definition.

[0024] Furthermore, the setting unit 35 sets the initial position 22 and the target position 25 of each of the plurality of agents 21 evacuating the area 20 to the coordinate values 27 of the map 16. Note that the identifier 46, the number thereof, and the initial position 22 set for the agent 21 can be arbitrarily set by the user. Note that the target position 25 may be automatically set by the simulator 30 when there are a plurality of exits in the area 20.

[0025] The first map 16a is formed by the forming unit 12 based on the space model 11 of the area 20 and is formed in any dimension from one-dimensional to three-dimensional. In addition to being expressed as a plurality of grids divided into a two-dimensional plane mesh as illustrated in FIGS. 3 and 4, the map 16 may be expressed as a node graph, expressed by three-dimensional CAD, etc., or may be hierarchically arranged and expressed. Also, the evacuation route 26 is similarly represented in any dimension from one to three as a simulation space imitating the square and passageways where the agent 21 actually moves.

[0026] The second map 16b and the third map 16c are obtained by reflecting the parameter 28, which is an index of the evacuation behavior of the agent 21, to the coordinate values 27 on the first map 16a of the plane by the reflecting units 15b and 15c. As a result, obstacles are added to the passage of the agent 21 to the evacuation route 26 including the coordinate values 27 according to the contribution degree of the event 24 such as a fire that has occurred.

[0027] The second map 16b reflects all of the parameters 28 generated by the above-described generation unit 47 in the coordinate values 27 without selection via the reflection unit 15b. On the other hand, the third map 16c selects only those recognized within a predetermined range 29 (FIG. 4) of the parameters 28 generated by the generation unit 47 via the reflection unit 15c and reflects them in the coordinate values 27.

[0028] The simulator 30 (30a, 30b, 30c) has, as a common function, a calculation unit 32 that calculates the coordinate values 27 from the initial position 22 toward the target position 25 using the map 16 in which the reflection of the parameter 28 is enabled or disabled, a determination unit 36 that determines the array of the coordinate values 27 continuously calculated and reaching the target position 25 as the evacuation route 26, and an execution unit 37 that executes each evacuation action of the agent 21 in time steps using the determined evacuation route 26.

[0029] Here, the map 16 in which the reflection of the parameter 28 is disabled refers to the first map 16a, and the maps 16 in which the reflection of the parameter 28 is enabled refer to the second map 16b and the third map 16c. The first map 16a, the second map 16b, and the third map 16c are input to the respective simulators 30a, 30b, 30c after being identified by the input unit 31.

[0030] The calculation unit 32 acquires the initial position 22 of the evacuation route 26, which is the current position of the agent 21 at the time of occurrence of the event 22. Next, the calculation unit 32 acquires the parameters 28 reflected in the coordinate values 27 in the vicinity of the agent 21 (excluding the first map 16a). Next, the calculation unit 32 sets the target position 25 based on the action model 45. For example, the calculation unit 32 may set a nearby emergency exit as the target position 25 or set the final evacuation location as the target position 25. Next, the calculation unit 32 performs a route search for the evacuation route 26 from the initial position 22 to the target position 25. This route search is performed for each of the first map 16a, the second map 16b, and the third map 16c.

[0031] When there are multiple arrays of coordinate values 27 that have reached the target position 25 from the initial position 22, the decision-making unit 36 takes into account contributions such as the presence or absence of obstacles for each route, the presence or absence of other agents 21, and the density, etc., and determines one evacuation route 26.

[0032] Furthermore, the simulator 30 includes the above-described generation unit 47 and a recognition unit 38 that recognizes the coordinate values 27 within a predetermined range 29 (FIG. 4) based on the positions of the agents 21 executing the evacuation behavior. This recognition unit 38 selects the parameters 28 corresponding to the coordinate values 27 recognized within this predetermined range 29 and reflects them on the third map 16c via the reflection unit 15c.

[0033] The execution unit 37 executes each evacuation route 26 of the agent 21 in each of the first map 16a, the second map 16b, and the third map 16c at time steps, and outputs the evacuation route 26 indicating the execution result. When the application of the parameter 28 changes at an intermediate position of the evacuation route 26, the simulation is retried and the evacuation route 26 from the intermediate position to the target position 25 is updated. The output execution result also includes the state quantities of each agent 21 that change during the process of the evacuation route 26 and the parameters 28 representing the surrounding situations.

[0034] Furthermore, the evacuation behavior prediction device 10 includes a collection unit 18 that collects each evacuation route 26 of the agent 21 determined by the first, second, and third simulators 30. The collected evacuation routes 26 include macroscopic information in addition to the location positions, state quantities, and recognition results of all the agents 21 at an arbitrary time step. Here, the macroscopic information includes, for example, the number of injured people at an arbitrary point in time and the evacuation time required for each agent 21 to reach the target position 25.

[0035] In the analysis unit 48, the evacuation routes 26 of the agent 21 of interest are compared in each of the first map 16a, the second map 16b, and the third map 16c. Then, based on the respective differences, the basis for determining the evacuation route 26 of the agent 21 is analyzed. In the display unit 49, the evacuation route 26 of the agent 21 and the basis for the determination in the first map 16a, the second map 16b, and the third map 16c are displayed.

[0036] (Second Embodiment) Next, a second embodiment of the present invention will be described with reference to FIG. 5. FIG. 5 is a block diagram showing the configuration of an evacuation behavior prediction device 10B (10) according to the second embodiment of the present invention. In FIG. 5, parts having the same configuration or function as those in FIG. 1 are denoted by the same reference numerals, and redundant explanations are omitted.

[0037] The evacuation behavior prediction device 10B in the second embodiment has, as a configuration common to the evacuation behavior prediction device 10A in the first embodiment, a holding unit 17 for the maps 16 (16a, 16b, 16c), a reflection unit 15 (15b, 15c) for reflecting the parameter 28 on the coordinate values 27 of the map 16, a setting unit 35 for setting the respective initial positions 22 and target positions 25 of the agents 21 on the coordinate values 27 of the map 16, and a simulator 30 (30a, 30b, 30c) for simulating the evacuation behavior of the agents 21.

[0038] The evacuation behavior prediction device 10B in the second embodiment is different from the evacuation behavior prediction device 10A in the first embodiment in that a function of a fourth simulator 30d is added to the simulator 30. This fourth simulator 30d shares the parameter 28 applied to the counterpart of the agents 21 in close proximity to each other during the execution of the evacuation behavior by the plurality of agents 21 and simulates the evacuation route 26.

[0039] Accordingly, a reflection unit 15d is added to reflect the newly shared parameter 28 on the coordinate values 27 of the third map 16c to form a fourth map 16d. Then, the holding unit 17 further holds the fourth map 16d.

[0040] In addition to the above-described calculation unit 32, determination unit 36, execution unit 37, and recognition unit 38, a sharing unit 39 is provided in this fourth simulator 30d. In the execution unit 37 of this fourth simulator 30d, it is assumed that the time steps of the evacuation routes 26 of different agents 21 coincide at the same or adjacent coordinate values 27. At this time, the sharing unit 39 extracts the parameter 28 applied to one agent 21.

[0041] Then, the reflection unit 15d reflects and shares the extracted parameter 28 at the same coordinate value 27 of the fourth map 16d of the other agent 21. Since the fourth map 16d is newly held in the holding unit 17, it means that the parameter 28 applied to the search has changed at an intermediate position of the evacuation route 26. Therefore, the simulation is retried, and the evacuation route 26 from the intermediate position to the target position 25 is updated.

[0042] Due to the function of this fourth simulator 30d, information uniquely recognized by a certain agent 21, such as the location of a fire or information on blocked passages due to rubble, can be shared with other agents 21. As a result, for other agents 21, the judgment information necessary to determine an effective evacuation route 26 will increase. As a specific method of such information sharing with other agents 21, in addition to the case of exchanging information at the coordinate value 27 where the timings of the evacuation routes 26 coincide as described above, the case of uploading information to a platform for sharing information such as an SNS is also assumed.

[0043] Then, the execution unit 37 executes each evacuation action of the agent 21 in each of the first map 16a, the second map 16b, the third map 16c, and the fourth map 16d in time steps, and outputs an evacuation route 26 indicating the execution result.

[0044] Based on the flowchart of FIG. 6 (refer to FIG. 5 as appropriate), the steps of the evacuation behavior prediction method according to each embodiment and the algorithm of the evacuation behavior prediction method will be described. First, a first map 16a is formed and held from the space model of area 20 (S11).

[0045] Next, a parameter 28 is reflected in the coordinate value 27 of the first map 16a, and a second map 16b commonly applied to each of the agents 21 is held (S12). Then, a parameter 28 is reflected in the coordinate value 27 of a predetermined range 29 based on the agent 21 located in the first map 16a, and a third map 16c individually applied to each of the agents 21 is held (S13). Further, the parameter 28 applied to other agents 21 is shared, and a fourth map 16d reflected in the coordinate value 27 of the third map 16c is held (S14).

[0046] Next, an initial position 22 and a target position 25 are set in the coordinate value 27 of the maps 16 (16a, 16b, 16c, 16d) (S15), and the evacuation route 26 of the agent 21 is simulated (S16). When the parameter 28 applied at an intermediate position of the evacuation route 26 changes (S17 Yes), the simulation is retried (S16), and the evacuation route 26 from the intermediate position to the target position 25 is determined (S17 No, S18).

[0047] Next, the evacuation routes 26 of each of the agents 21 determined by the first, second, third, and fourth simulations are collected (S19). Then, in each of the first map 16a, the second map 16b, the third map 16c, and the fourth map 16d, the evacuation routes 26 of the agent 21 to be focused on are compared and analyzed (S20). Then, from the respective differences, the basis for the determination of the agent 21 that determined the evacuation route 26 in the first map 16a, the second map 16b, the third map 16c, and the fourth map 16d is analyzed and displayed (S21).

[0048] According to the evacuation behavior prediction device of at least one of the above-described embodiments, by collecting and comparing the evacuation routes of each agent determined by the first, second, and third simulations, it is possible to facilitate the analysis of the basis for the evacuation behavior taken by the agent.

[0049] Although several embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, changes, and combinations can be made without departing from the gist of the invention. These embodiments and their modifications are included in the scope and gist of the invention, as well as in the invention described in the claims and the equivalent scope thereof.

[0050] The evacuation behavior prediction device described above includes a highly integrated control device such as a dedicated chip, FPGA (Field Programmable Gate Array), GPU (Graphics Processing Unit), or CPU (Central Processing Unit), a storage device such as ROM (Read Only Memory) and RAM (Random Access Memory), an external storage device such as HDD (Hard Disk Drive) and SSD (Solid State Drive), a display device such as a display, an input device such as a mouse and a keyboard, and a communication I / F, and can be realized with a hardware configuration using an ordinary computer. Therefore, the components of the evacuation behavior prediction device can also be realized by a computer processor and can be operated by an evacuation behavior prediction program.

[0051] Also, the evacuation behavior prediction program is provided by being pre-embedded in ROM or the like. Alternatively, this program may be stored and provided in a computer-readable storage medium such as a CD-ROM, CD-R, memory card, DVD, flexible disk (FD), etc. in an installable or executable file format.

[0052] In addition, the evacuation behavior prediction program according to the present embodiment may be stored on a computer connected to a network such as the Internet and provided by being downloaded via the network. Further, the evacuation behavior prediction device can also be configured by interconnecting and combining separate modules that independently exhibit each function of the components with a network or a dedicated line.

Description of Reference Numerals

[0053] 10(10A, 10B)… Evacuation behavior prediction device, 11… Space model, 12… Formation unit, 15(15b, 15c, 15d)… Reflection unit, 16a(16)… First map (map), 16b(16)… Second map (map), 16c(16)… Third map (map), 16d(16)… Fourth map (map), 17… Holding unit, 18… Collection unit, 20… Area, 21… Agent, 22… Initial position, 24… Event, 25… Target position, 26… Evacuation route, 27… Coordinate value, 28… Parameter, 29… Predetermined range, 30(30a, 30b, 30c, 30d)… Simulator, 31… Input unit, 32… Calculation unit, 35… Setting unit, 36… Decision unit, 37… Execution unit, 38… Cognition unit, 39… Sharing unit, 40… Registration unit, 41… Event scenario, 42… Environmental conditions, 45… Behavior model, 46… Identifier, 47… Generation unit, 48… Analysis unit.

Claims

1. A holding unit that holds a map representing an evacuation route of an agent located in an area where an event has occurred; A reflecting unit that reflects a parameter indicating the contribution degree of the event to the determination of the evacuation route on the coordinate values of the map; A setting unit that sets the initial position and the target position of each of the plurality of agents evacuating from the area to the coordinate values of the map; A first simulator that simulates the evacuation route from the initial position to the target position for each of the agents without applying the parameter; A second simulator that commonly applies the parameter and simulates the evacuation route of each of the agents; A third simulator that individually applies the parameter and simulates the evacuation route of each of the agents; An evacuation behavior prediction device comprising a collection unit that collects the evacuation routes of each of the agents determined by the first, second, and third simulators.

2. In the evacuation behavior prediction device according to Claim 1, The simulation in the third simulator is an evacuation behavior prediction device that recognizes the coordinate values within a predetermined range based on the position of the agent and to which the parameter reflected therein is applied.

3. In the evacuation behavior prediction device according to Claim 1 or Claim 2, An evacuation behavior prediction device comprising a fourth simulator that shares the parameter applied to the other party of the agents in proximity to each other and simulates the evacuation route.

4. In the evacuation behavior prediction device according to any one of Claims 1 to 3, The first, second, and third simulators are: A calculation unit that calculates the coordinate values from the initial position to the target position using the map with the reflection of the parameter enabled / disabled; A determination unit that determines the array of the coordinate values continuously calculated and reaching the target position as the evacuation route; An evacuation behavior prediction device having an execution unit that executes the evacuation behavior of each of the agents in time steps according to the determined evacuation route.

5. In the evacuation behavior prediction device according to any one of Claims 1 to 4, When the parameter to be applied changes at an intermediate position of the evacuation route, the simulation in the first, second, and third simulators is retried and the evacuation route from the intermediate position to the target position is updated.

6. In the evacuation behavior prediction device according to any one of claims 1 to 5, a registration unit for registering an event scenario for generating the event at a predetermined position in the area at a predetermined time; an evacuation behavior prediction device in which the parameter is updated and reflected in the coordinate value each time a new event occurs according to the event scenario.

7. In the evacuation behavior prediction device according to any one of claims 1 to 6, each agent is provided with an identifier for identifying each agent and a behavior model involved in the simulation algorithm in the first, second, and third simulators of the evacuation route.

8. A step of holding, by a holding unit, a map representing an evacuation route of an agent located in an area where an event has occurred; a step of reflecting, by a reflecting unit, a parameter indicating the contribution degree of the event to the determination of the evacuation route in the coordinate value of the map; a step of setting, by a setting unit, the initial position and the target position of each of a plurality of agents evacuating from the area in the coordinate value of the map; a step of first simulating, by a first simulator, the evacuation route from the initial position to the target position for each of the agents without applying the parameter; a step of second simulating, by a second simulator, the evacuation route of each of the agents by commonly applying the parameter; a step of third simulating, by a third simulator, the evacuation route of each of the agents by individually applying the parameter; a step of collecting, by a collection unit, the evacuation route of each of the agents determined by the first, second, and third simulations. An evacuation behavior prediction method including.

9. On a computer, a step of holding a map representing an evacuation route of an agent located in an area where an event has occurred; a step of reflecting a parameter indicating the contribution degree of the event to the determination of the evacuation route in the coordinate value of the map; a step of setting the initial position and the target position of each of a plurality of agents evacuating from the area in the coordinate value of the map; a step of first simulating the evacuation route from the initial position to the target position for each of the agents without applying the parameter; Step of commonly applying the parameters and second-simulating each of the evacuation routes of the agents, Step of individually applying the parameters and third-simulating each of the evacuation routes of the agents, Evacuation behavior prediction program that causes the steps of collecting each of the evacuation routes of the agents determined by the first, second, and third simulations to be executed.

Images