JP2014023548A - System and method for evacuation action simulation - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a system and a method for evacuation action simulation which can simulate evacuation action in crowd flow conditions with high accuracy.SOLUTION: A method for evacuation action simulation to simulate action of evacuees, who evacuate from a building or in an outdoor space, with cellular automaton includes following steps of: dividing an indoor or outdoor space into cells; arranging all evacuees as simulation targets in each cell; and determining a next destination cell for each evacuee by each discrete time step on the basis of the local interaction rule with fluid coefficients in crowd flow conditions (especially 1.5/m s by horizontal walking, 1.33/m s by walking on steps) when the density of evacuees is the density of crowd flow conditions.

Description

  The present invention relates to an evacuation behavior simulation system and an evacuation behavior simulation method for simulating the behavior of an evacuee who evacuates from a building or evacuates in an outdoor space using a cellular automaton.

  Conventionally, simulations of behavior of evacuees evacuating from various buildings such as office buildings have been performed. For example, Patent Document 1 discloses a simulation of evacuation from a high-rise building that uses a multi-agent system to determine whether to use an elevator or use stairs. Non-Patent Document 1 discloses a simulation of evacuation flow using a multi-agent cellular automaton that also includes a cellular automaton.

JP 2010-14878 A

Fumio Otsuki, Motohiro Onoki, “Simulation of Evacuation Flow by Cellular Automata Method”, Japanese Journal of Operations Research 2008 51-94-111

When reproducing evacuation behavior from a building, it is necessary to consider the walking state, and the walking state has three states according to the crowd density: a free walking state, a crowd flow state, and a staying state. The free walking state is a state in which the crowd density is up to about 1 person / m ² and can walk freely at a constant walking speed (upper limit walking speed). Therefore, in the case of a free walking state, the walking speed is constant in the living room in the building, the moving space (corridor, attached room, etc.), the stairs and the exit passage of each living room (confluence when exiting from the exit to the corridor), etc. It is possible to simulate each refugee's position and evacuation action time for each step with a simple equation with high accuracy. In addition, the staying state is a state where the crowd density exceeds 4 people / m ² and an unpredictable panic or the like may occur, which is not suitable for simulation of evacuation behavior. Therefore, when reproducing the evacuation behavior, it is important to simulate a crowd flow state in which the crowd density is about 1 to 4 people / m ² .

  In the simulation described in Non-Patent Document 1, the cellular automaton local neighborhood rule is a rule for determining the direction, the walking speed is set randomly for each pedestrian, and the traveling speed of each pedestrian is constant during evacuation. It is. However, in the crowd flow state, the walking speed of each pedestrian varies according to the surrounding situation (crowd density) at each time, so the simulation described in Non-Patent Document 1 reproduces the evacuation behavior in the crowd flow state with high accuracy. It is not possible to find a reasonable evacuation action time. In particular, when reproducing the evacuation behavior in a crowd flow state, assuming that the entire building is evacuated from the building, it is important to pass through the exit of a living room where refugees are concentrated, and it is necessary to determine the walking speed of the exit passage with high accuracy. is there. However, in the simulation described in Non-Patent Document 1, the evacuation behavior to the exit of the living room is simulated, and exit passage is not taken into consideration. Also, in the simulation described in Patent Document 1, the walking speed corresponding to the crowd flow state and the crowd density when passing through the exit is not taken into consideration.

  Then, this invention makes it a subject to provide the evacuation action simulation system and evacuation action simulation method which can simulate the evacuation action in a crowd flow state with high precision.

  An evacuation behavior simulation system according to the present invention is an evacuation behavior simulation system that simulates the behavior of an evacuee who evacuates from a building or evacuates in an outdoor space using a cellular automaton, and divides the space in the building or outdoors into cells. If all the evacuees to be simulated are placed in each cell and the density of the evacuees is the density of the crowd flow state at every discrete time step, all of them are based on the local neighborhood rule that becomes the flow coefficient in the crowd flow state The next destination cell of the evacuees is determined.

  In this evacuation behavior simulation system, a cellular automaton is used to divide a space inside or outside a building into cells, all evacuees to be simulated are placed in each cell, and local neighborhood rules (neighboring neighbors) are set for each discrete time step. The next destination cell of all evacuees is sequentially determined based on the movement rules based on the positional relationship with other evacuees), and evacuation is completed until the evacuation is completed (for example, the evacuation is completed, the floor evacuation is completed, the entire evacuation is completed) Simulate behavior. Thereby, the evacuation action time until the evacuation is completed can be predicted. By using cellular automata, each evacuee is mechanically human based on only local neighborhood rules (all evacuees have the same walking characteristics, and humans who have abandoned cognitive functions such as perception, memory, and judgment) Each model can be moved and the evacuation behavior can be reproduced with high accuracy. In particular, in the evacuation behavior simulation system, when the density of evacuees is the density of the crowd flow state, a local neighborhood rule that sets the flow coefficient in the crowd flow state (number of people who pass 1 m wide per second) is set. By doing so, the evacuees are moved so as to have a flow coefficient in the crowd flow state (particularly when passing through the exit where the density increases), and the evacuation behavior in the crowd flow state is simulated. In this way, the evacuation behavior simulation system simulates the behavior of evacuees evacuating from a building or evacuating in an outdoor space using a cellular automaton based on a local neighborhood rule that is a flow coefficient in a crowd flow state. Evacuation behavior in the state can be simulated with high accuracy. As a result, evacuation action times such as room evacuation, floor evacuation, and whole building evacuation in the building can be predicted with high accuracy. In a place where the density of evacuees is low, there may be a free walking state. In this case, the evacuees are moved at a walking speed in the free walking state.

  In the above evacuation behavior simulation system of the present invention, the flow coefficient in the crowd flow state is preferably 1.5 person / m · s for horizontal walking and 1.33 person / m · s for stair walking. It is. The flow coefficient of 1.5 people / m · s for walking in the horizontal part and the flow coefficient of 1.33 person / m · s for walking in the staircase are calculated based on the Evacuation Safety Verification Act It is stipulated in Chapter 5 of “Explanation of fire prevention and evacuation plans for objects”. Therefore, when reproducing the evacuation behavior in the crowd flow state, the density and speed of the evacuees differ from cell to cell, and the relationship changes at each discrete time step. It is a necessary condition for all evacuees to make adjustments so that 1.33 person / m · s is established for the person / m · s and staircase walking. In addition, since it is very difficult to perform a simulation with a flow coefficient of strictly 1.5 person / m · s or 1.33 person / m · s, a value around 1.5 person / m · s or Including values of around 1.33 person / m · s. Further, in places where the density of evacuees is low, there may be cases where 1.5 persons / m · s or 1.33 persons / m · s is not achieved.

  An evacuation behavior simulation method according to the present invention is an evacuation behavior simulation method for simulating the behavior of an evacuee who evacuates from a building or evacuates in an outdoor space using a cellular automaton, and divides the space inside or outside the building into cells. If all the evacuees to be simulated are placed in each cell and the density of the evacuees is the density of the crowd flow state at every discrete time step, all of them are based on the local neighborhood rule that becomes the flow coefficient in the crowd flow state The next destination cell of the evacuees is determined. In the evacuation behavior simulation method of the present invention, the flow coefficient in the crowd flow state is preferably 1.5 person / m · s for horizontal walking and 1.33 person / m · s for stair walking. It is. This evacuation behavior simulation method has the same operations and effects as the evacuation behavior simulation system described above.

  According to the present invention, evacuation in a crowd flow state is simulated by simulating the behavior of a refugee evacuating from a building or evacuating in an outdoor space using a cellular automaton based on a local neighborhood rule that becomes a flow coefficient in a crowd flow state The behavior can be simulated with high accuracy.

It is a block diagram of the evacuation action simulation system which concerns on this Embodiment. It is a table | surface which shows an example of analysis space data. It is a table | surface which shows an example of simulation conditions. This is a relationship in horizontal walking, where (a) is the relationship between density and speed, and (b) is the relationship between density and flow coefficient. This is a relationship in walking down the stairs, (a) is the relationship between density and speed, and (b) is the relationship between density and flow coefficient. It is a relationship in staircase climbing, (a) is the relationship between density and speed, and (b) is the relationship between density and flow coefficient. It is explanatory drawing of a cell. It is explanatory drawing of the determination of a main direction axis, (a) is explanatory drawing of the calculation method of the selection probability of the direction vector with respect to the moving direction of a target, (b) is the 1st candidate of a main direction axis, and 2nd candidate It is an example, and (c) is an example of the first candidate and the second candidate of the main direction axis. It is a figure which shows the 8 * 3 = 24 direction which can move. A personal space, (a) is an orthogonal personal space, and (b) is a 45 ° personal space. It is explanatory drawing of the exit selection of a living room, (a) is explanatory drawing of the calculation method of the estimated evacuation end time with respect to the selected exit, (b) is the calculation method of the shortest evacuation end estimated time when there are a plurality of exits. It is explanatory drawing, (c) is explanatory drawing of the calculation method of the estimated evacuation completion time with respect to each other exit when changing a selection exit. It is a flowchart which shows the flow of the main process in the evacuation action simulation system which concerns on this Embodiment. It is a flowchart which shows the flow of the pedestrian evacuation action process in the main process of FIG. It is analysis space data (passage model) of the passage, (a) is a passage model in the orthogonal direction, (b) is a passage model in the oblique １ ／ direction, and (c) is a passage in the oblique ½ direction. (D) is a 45 ° direction passage model. It is the result (relationship between a density and speed) of the simulation in each channel | path model of FIG. It is the result (relationship between a density and a flow coefficient) of the simulation in each channel | path model of FIG. FIG. 15 is a screen drawing of a simulation using the orthogonal passage model in FIG. 14, where (a) is a density of 0.5 person / m ² and (b) is a density of 1.0 person / m ² . (C) is a case where the density is 2.0 person / m ² , (d) is a case where the density is 3.0 person / m ² , and (e) is a case where the density is 4.0 person / m ^2. This is the case for m ² . Staircase analysis space data (staircase model), (a) is a staircase model from the third floor to the second floor, and (b) is a staircase model from the second floor to the first floor. It is the result (relationship between a density and speed) of the simulation in the staircase model of FIG. It is the result (relationship between a density and a flow coefficient) of the simulation in the staircase model of FIG. It is analysis space data of the room. It is the result of the simulation of the room evacuation (relationship between the elapsed time and the total number of people leaving). It is the result (relationship between elapsed time and flow coefficient) of the simulation of living room evacuation. It is a screen drawing of a simulation of a room evacuation, (a) is an elapsed time of 0.0 s, (b) is an elapsed time of 4.3 s, (c) is an elapsed time of 9.2 s, d) is elapsed time is 16.1 s, (e) is elapsed time is 23.0 s, and (f) is elapsed time is 28.75 s. It is analysis space data of a living room + corridor (1 floor). It is the result of simulation of living room + corridor evacuation (relationship between elapsed time and total number of people leaving). It is the result (relationship between elapsed time and flow coefficient) of the simulation of living room + corridor evacuation. It is a screen drawing of a simulation of a living room + corridor evacuation, (a) is an elapsed time of 0.0 s, (b) is an elapsed time of 9.2 s, and (c) is an elapsed time of 19.55 s. . It is a screen drawing of a simulation of a living room + corridor evacuation, (a) is 29.90 s elapsed time, (b) is 68.77 s elapsed time, (c) is 98.21 s elapsed time. . This is analysis space data of the room + corridor + stairs (2nd and 3rd floor parts of the whole building). It is the result of simulation of living room + corridor + stair evacuation (relationship between elapsed time on 3rd floor and total number of people leaving). It is the result of the simulation of living room + corridor + stair evacuation (relationship between elapsed time on the second floor and the total number of people leaving). It is a result of simulation of living room + corridor + stair evacuation (relationship between elapsed time on 3rd floor and flow coefficient). It is a result of simulation of living room + corridor + stair evacuation (relationship between elapsed time on 3rd floor and flow coefficient). It is the result of simulation of living room + corridor + stair evacuation (relationship between elapsed time and flow coefficient on the second floor). It is the result of simulation of living room + corridor + stair evacuation (relationship between elapsed time and flow coefficient on the second floor). It is a screen drawing of a simulation of a living room + corridor + stair evacuation, (a) is an elapsed time of 0.0 s, and (b) is an elapsed time of 11.5 s. It is a screen drawing of a simulation of a living room + corridor + stair evacuation, (a) is an elapsed time of 57.50 s, and (b) is an elapsed time of 103.50 s. It is a screen drawing of a simulation of a living room + corridor + stair evacuation, (a) is elapsed time 149.50 s, (b) is elapsed time 184.46 s.

  Hereinafter, embodiments of an evacuation behavior simulation system and an evacuation behavior simulation method according to the present invention will be described with reference to the drawings. In addition, the same code | symbol is attached | subjected about the element which is the same or it corresponds in each figure, and the overlapping description is abbreviate | omitted.

  The evacuation behavior simulation system according to the present embodiment uses a cellular automaton to simulate evacuation behavior from a building in a crowd flow state and a free walking state. In particular, the evacuation behavior simulation system according to the present embodiment expresses the evacuation behavior prediction calculation method (the theory of the evacuation prediction model) by Cell Automan in order to obtain a simulation that can be approved by the Minister of Land, Infrastructure, Transport and Tourism. A simulation is performed, and the relationship between crowd density and walking speed is simulated by applying the settings in the evacuation safety verification method and guidelines.

  Cellular automata are discrete by dividing the analysis target into divided areas called cells, defining discrete state quantities on each cell, and considering only local neighborhood rules (rules of interaction with neighboring cells). It is a discrete calculation model or a discrete simulation method that enables an analysis of an approximate phenomenon by changing the state quantity of each cell at each time step and expressing a complex phenomenon as a whole. The feature of the analysis method of the computer system by the cellular automaton is that it is not necessary to obtain the governing equation to be analyzed, all calculations are integer operations, and only the phenomenon is highlighted on the computer screen.

  Since the analysis target (pedestrian (refugee)) in the crowd flow state cannot be handled as a continuum in the space axis and time axis, it is discretized for each minute part and minute time and approximated in the crowd flow state The cellular automaton is used because it is necessary to reproduce the evacuation situation. By using cellular automata, pedestrians placed in cells move as mechanical persons based only on local neighborhood rules, and pedestrians do not move autonomously (all refugees have the same walking characteristics). , It is a human model that excludes cognitive functions such as individual perception, memory, and judgment). In addition, by using cellular automata, system development becomes easy and arithmetic processing capability is light.

  The approval of the Minister of Land, Infrastructure, Transport and Tourism is based on Article 129-2, Paragraph 1 of the Building Standards Act Enforcement Order and Article 129-2, Paragraph 1 of the Building Standards Act Enforcement Order. The evacuation behavior prediction calculation method (the theory of evacuation prediction model) is the “Comprehensive Fire Prevention Design Method for Buildings Vol. 3 (Evacuation Safety Design Method)” published by the Japan Architecture Center in 1989. It is described in Chapter 4. The evacuation safety verification method is based on the Ministry of Construction Notification Nos. 1441 and 1442 in 2000. The guidelines are those described in Chapter 5 of the "New Building Disaster Prevention Plan Guidelines-A Guide to Building Fire Prevention and Evacuation Plans" published by the Japan Architecture Center in 1995, supervised by the Housing Bureau Building Guidance Division, Ministry of Construction. It is.

  With reference to FIGS. 1-13, the evacuation action simulation system 1 which concerns on this Embodiment is demonstrated. FIG. 1 is a configuration diagram of an evacuation behavior simulation system. FIG. 2 is a table showing an example of analysis space data. FIG. 3 is a table showing an example of simulation conditions. FIG. 4 shows the relationship between the density and the velocity and the flow coefficient when walking horizontally. FIG. 5 shows the relationship between density, speed and flow coefficient when walking down the stairs. FIG. 6 shows the relationship between density, speed, and flow coefficient when walking up the stairs. FIG. 7 is an explanatory diagram of a cell. FIG. 8 is an explanatory diagram for determining the main direction axis. FIG. 9 is a diagram showing a movable 8 × 3 = 24 directions. FIG. 10 shows a personal space. FIG. 11 is an explanatory diagram for selecting an exit of a living room. FIG. 12 is a flowchart showing a flow of main processing in the evacuation behavior simulation system. FIG. 13 is a flowchart showing the flow of the pedestrian evacuation action process in the main process of FIG.

  The evacuation behavior simulation system 1 expresses the walking speed of each pedestrian in a free walking state and a crowd flow state by a cellular automaton. Further, the evacuation behavior simulation system 1 expresses evacuee characteristics, an evacuation space, an evacuation behavior model, and an evacuation behavior time prediction computation method based on an evacuation behavior prediction calculation method (the theory of an evacuation prediction model) by a cellular automaton. In the evacuation behavior simulation system 1, the evacuation behavior is simulated by the cellular automaton expressed as such, and the evacuation behavior situation according to the elapsed time from the start of the evacuation is visualized, and the evacuation is completed (for example, the evacuation is completed in the room) Evacuation action time until floor evacuation completion, whole building evacuation completion).

  The flow coefficient and walking speed will be described. Here, first, the flow coefficient and walking speed in the free walking state of the horizontal portion will be described with reference to FIG. 4, and then the flow coefficient in the crowd flow state of the horizontal portion with reference to FIG. The walking speed will be described. Furthermore, with reference to FIG.5 and FIG.6, the flow coefficient and walking speed in a staircase part are demonstrated.

The flow coefficient and walking speed in the free walking state of the horizontal part will be described. The free walking state is a state in which the density of the front and surroundings is small, and there is no action restriction due to the influence of other pedestrians, and the user can walk freely. In the free walking state, the crowd density is up to about 1 person / m ² . The traveling speed in this free walking state is the free walking speed. In the present embodiment, the walking speed in the free walking state has an upper limit of 1.3 m / s (seconds) according to the evacuation safety verification method and guidelines (see FIG. 4A), and is less than 1.3 m / s. In this case, the speed is adjusted with the movement probability P = v / 1.3. In the free walking state, the relationship between the density and the flow coefficient is linear (flow coefficient = speed × density) (see FIG. 4B). In the free walking state, the walking speed is constant as described above, so the position and evacuation action time for each discrete time step of each pedestrian can be analyzed by a simple equation.

The flow coefficient and the walking speed in the horizontal part crowd flow state will be described. As the density of pedestrians increases, the distance from the pedestrians in front is reduced, so that the stride is small, the walking pitch becomes gentle, and the walking speed decreases. This phenomenon has been shown in previous studies on the relationship between density, walking speed, and flow coefficient (see, for example, “Building Architectural Design Materials [Human] (2003 Architectural Institute of Japan)” P128). Shows a state change from a free walking state to a crowd flow state from a density exceeding 1 person / m ² . The crowd flow state has a density of about 1 to 4 people / m ² . In the case of the crowd flow state, in the range of about 1 to 4 people / m ² by three indicators of density ρ (person / m ² ), velocity v (m / s), and flow coefficient N (person / m · s). , N = ρ × v is established.

  In the crowd flow state, the relationship between the density and the walking speed is nonlinear, and the walking speed of each pedestrian is sequentially determined according to the position of each discrete time step. Therefore, the whole pedestrian (refugee) cannot be bundled at a uniform walking speed. Therefore, the walking speed in the horizontal part in the crowd flow state is calculated as the density using the relational expression between the density and the walking speed shown in the safety section of “Architecture Design Material Collection 10 Technology (1983, Architectural Institute of Japan)”. The walking speed is set accordingly. The flow coefficient is 1.5 person / m · s in the evacuation safety verification method and guidelines.

Even in a free walking state, the density of pedestrians increases in the narrow neck portion such as the exit, and a crowd flow state may occur. In this case as well, the flow coefficient is obtained by actual measurement in a past study (for example, see “Building Architectural Design Data [Human] (edited by the Architectural Institute of Japan) P128 crowd behavior: density, flow velocity, flow coefficient”). It is shown to be about 5 people / m · s. This flow coefficient coincides with the flow coefficient in the above guideline = 1.5 person / m · s and Neff = 90 person / m · minute in the evacuation safety verification method, which is the basis for setting these standards. . In accordance with these past studies and current standards, in this embodiment, in the crowd flow state where the density is about 1 to 4 people / m ² , the flow coefficient of the horizontal portion is 1.5 people / m · s (see FIG. 4 (b)). Therefore, in the crowd flow state, the walking speed of the horizontal portion is v = 1.5 / ρ (see FIG. 4A). In the present embodiment, in the horizontal portion, the crowd flow state and the flow coefficient at the exit passage are reproduced so as to be 1.5 person / m · s.

In a state where the density exceeds 4 persons / m ² , the gait must be changed frequently, and the state changes to a staying state that may stop even at a low speed. In the case of this staying state, there is a possibility that “shogi toppling” or the like may occur due to conflicting points. In the above evacuation safety verification method and guideline, the density is limited to 3.3 to 5 persons / m ² or more, but in this embodiment, the stay state is difficult to reproduce in accordance with the idea. The following phenomena are not considered, and the phenomena up to free walking and crowd flow are handled.

  The flow coefficient and walking speed in the staircase will be described. The walking speed in the Evacuation Safety Verification Act depends on the use of the building or room: 1: theaters, 2: department stores / exhibition halls, apartment houses, hotels, etc. (excluding hospitals, etc.) 3: schools, offices, etc. Although it is classified and set in three types, the descending walking speed and the ascending walking speed of the stairs are 0.6 times and 0.45 times the walking speed in the horizontal part, respectively. In this embodiment, according to this evacuation safety verification method, the walking speed in the free walking state down the stairs is a value obtained by multiplying the free walking speed in the horizontal part by a movement probability = 0.6 times, When the free walking speed of the horizontal portion is 1.3 m / s, the speed is 0.78 m / s (see FIG. 5A). Further, the walking speed in the free walking state up the stairs is set to a value obtained by multiplying the free walking speed in the horizontal part by the movement probability = 0.45 times, and the free walking speed in the horizontal part = 1.3 m / In the case of s, it is set to 0.585 m / s (see FIG. 6A). Also in the case of the staircase, in the free walking state, the relationship between the density and the flow coefficient is linear (see FIGS. 5B and 6B).

In the staircase, the walking speed decreases as the density increases (see FIGS. 5A and 6A). The flow coefficient of the staircase in practical evacuation calculation is 1.3 person / m · s in the above guidelines, and the effective flow coefficient of the evacuation safety verification method is Neff in the range of density up to 4 person / m ^2. = 80 people / m · min (= 1.33 people / m · s). In the present embodiment, the flow coefficient of the crowd flow state at the staircase portion is set to 1.33 person / m · s in accordance with 1.33 person / m · s in the evacuation safety verification method (FIG. 5B). FIG. 6 (b)). Therefore, in the crowd flow state, the walking speed of the staircase portion is v = 1.33 / ρ (see FIGS. 5A and 6A). In the present embodiment, in the staircase portion, the local flow law of the crowd flow coefficient set in the horizontal portion is multiplied by the movement probability P = 1.33 / 1.5, and the flow coefficient in the crowd flow state of the staircase portion. To reproduce.

  In this embodiment, in the case of horizontal walking in a living room or a corridor, the relationship between the density and the flow coefficient shown in FIG. 4B (particularly, the flow coefficient is 1.5 people / m in the crowd flow state).・ A simulation is conducted with s) as a target, and in the case of walking down a staircase, the relationship between the density and the flow coefficient shown in FIG. 5B (particularly in the crowd flow state, the flow coefficient is 1.33 person / m · s). ) And the relationship between the density and flow coefficient shown in FIG. 6 (b) (particularly, the flow coefficient is 1.33 person / m · s in the crowd flow state) Simulate as a goal.

  The following describes the evacuee characteristics, evacuation space, and evacuation behavior model based on the evacuation behavior prediction calculation method (the theory of evacuation prediction model). Evacuee characteristics will be described. As for the characteristics of evacuees, all evacuees (pedestrians) can evacuate by themselves, and the characteristics of evacuees are not individually distinguished. Therefore, all evacuees have the same evacuee characteristics. This is realized by moving each pedestrian considering only the local neighborhood rule.

  The evacuation space model will be described. The evacuation space model is composed of space elements such as a living room, a moving space (corridor, attached room, etc.), stairs, and a final evacuation area as an evacuation space, and an evacuation model is set for each evacuation space. In the case of evacuation in a living room, there may be one exit or multiple exits. If there is only one exit, the evacuees uniformly distributed in the room after the start of evacuation start evacuation all at once toward the exit. The exit is reached by a centripetal path that goes straight in the shortest distance, and passes through the exit at a rate of flow coefficient≈1.5 person / m · s. Note that the L-shaped path is possible by arranging fixtures or the like. When there are a plurality of exits, after the start of evacuation, evacuees uniformly distributed in the living room start evacuation all at once toward the nearest exit. During the evacuation action, each refugee selects the exit that is predicted to be able to leave the earliest in consideration of the evacuee outflow speed of each exit and the amount of residence before the exit, and changes the target exit. This exit selection method will be described in detail later. In the case of evacuation in the moving space, the evacuees who flow into the moving space evacuate in an orderly manner toward the target exit without overtaking and returning. This is also realized by moving each pedestrian considering only the local neighborhood rule. The movement calculation is executed by dividing the movement space into units (cells) for each v × Δt and moving the evacuees in the evacuation direction at discrete time steps. The evacuation on the stairs is the same as the evacuation in the moving space, and the walking speed and the flow coefficient are described above.

  The evacuation behavior model will be described. It is assumed that the evacuees evacuate the passage (evacuation route) planned by the designer as an evacuation route in an orderly and confusing manner. Before the start of evacuation, the residents in the building are uniformly distributed in the offices, assembly rooms, etc., and after the start of evacuation, the residents leave the room as refugees, corridors, attached rooms, It shall move to the final evacuation site through an evacuation route composed of stairs and the like. In addition, since all have the same evacuee characteristics, there is no arbitrary overtaking or backtracking. Although the evacuees are distributed uniformly, other than the uniform distribution may be arbitrarily arranged.

  The prediction calculation method of evacuation action time is demonstrated. As the evacuation action time, the evacuation action time indicated in the evacuation action prediction calculation method is applied. In an actual evacuation site, if many evacuees are concentrated and stay at the exit, the evacuation time is determined by the exit passage time, and if the density is low and not staying at the exit, walking time is determined. The time to evacuate is determined by. Therefore, as the evacuation action time, max [walking time, exit passage time] indicated in the guideline based on the evacuation action prediction calculation method matches the actual situation of evacuation, and it is possible to predict realistic and rational evacuation action time. To do.

  Explain evacuation on the stairs. As for the stairs, the simulation of the crowd flow state at the stairs is performed according to the above movement probability in the same manner as the simulation of the crowd flow state in the horizontal movement space, and the approximate evacuation action time at the stairs is predicted.

  The cell will be described with reference to FIG. In the present embodiment, the size of one cell of the cellular automaton is 30 cm × 30 cm. It is set that only one pedestrian can exist in one cell. In principle, the pedestrians P and P do not enter the cells adjacent to each other in the orthogonal direction with respect to the arrangement of the cells S,. Each cell state defines the following three types. 1: It is a pedestrian cell (the state where the pedestrian P exists in the cell S as shown in FIG. 7). 2: It is a space cell (as shown in FIG. 7, a pedestrian P or the like does not exist in the cell S, and the space where the pedestrian P can move). 3: An obstacle cell (as shown in FIG. 7, the cell S has a fixed object such as a pillar, a wall W, a fixture F, etc., and is in an unmovable space state).

The direction vector will be described with reference to FIG. In the present embodiment, the directions of adjacent cells that can be moved for each discrete time step of the cellular automaton are 8 directions in total, that is, an orthogonal direction and a 45 ° oblique direction with respect to the cell arrangement. The eight directions are roughly classified into two types, that is, an orthogonal direction and a 45 ° oblique direction, which are defined as two types of direction vectors as an orthogonal direction vector and a 45 ° direction vector, respectively. Since the pedestrian's target moving direction is the direction to the exit of the living room, hallway, etc., these two types of directions may not always match, and the moving direction is selected according to the priority order according to the following rules. As shown in FIG. 8 (a), the moving direction vector to the outlet direction between the target and V, and the orthogonal direction vector of length a and V _A, the 45 ° direction vector of length b and V _B. In this case, the probability of selecting the orthogonal direction vector V _A is a / a + b, and the probability of selecting the 45 ° direction vector V _B is b / a + b. The one with the higher probability is the first candidate vector, and the one with the lower probability is the second candidate vector. However, a can be calculated by the following formula (1), and b can be calculated by the following formula (2).

In the example shown in FIG. 8B, in the case of the pedestrian P1, the target moving direction vector V ₁ toward the exit E of the living room is the first candidate is the orthogonal direction vector V _A1 , and the second candidate is the 45 ° direction. The vector V _B1 is obtained. For pedestrian P2, a movement direction vector V ₂ of the target toward the room exit E, the first candidate is 45 ° direction vector V _B2, and the second candidate a perpendicular direction vector V _A2. In the example shown in FIG. 8 (c), when the pedestrian P3, a movement direction vector V ₃ goal towards the room exit E, first candidate orthogonal direction vector V _A3, and the second candidate two The 45 ° direction vectors V _B31 and V _B32 are obtained. For pedestrian P4, a target moving direction vector _{V 4} of towards the room exit E, the first candidate is 45 ° direction vector _{V B4,} and the orthogonal direction of the second candidate are two vectors _{V A41, V A42} and Become.

  When the first candidate vector and the second candidate vector are determined, it is first determined whether or not the cell can be moved to the cell in the direction of the first candidate vector. If the vector can be moved, the first candidate vector is selected and cannot be moved. In this case, a stop or second candidate vector is selected. The selected direction vector is called a main direction axis.

  In this embodiment, the actual time of one discrete time step of the cellular automaton is 0.23 so that the free walking speed in the orthogonal vector is 1.3 m / s because one cell is 30 cm × 30 cm. Set to seconds. When the free walking speed is less than 1.3 m / s, the speed is adjusted with the movement probability P = v / 1.3. For the 45 ° direction vector, the movement probability P = 1 / √2 (where √2 indicates the square root of 2) is set so that the free walking speeds are equal.

  By the way, as a human walking behavior, it seems unnatural at first glance that the direction is only the direction orthogonal to the direction perpendicular to the cell arrangement. Regarding that, first, the average step length of an adult male is about 75 to 80 cm, so that the 2-3 discrete time steps in the present embodiment correspond to one step of a normal pedestrian. In this case, the free walking speed per step other than the orthogonal direction and the 45 ° direction is slightly reduced to about 1.2 m / s, but it is assumed that there is no sudden increase or decrease in speed, and the direction is 8 × 3 = 24. On the other hand, a simulation is performed on the assumption that a natural walking state can be reproduced. FIG. 9 shows the current pedestrian cell with a black circle located in the center, 28 white circles showing the pedestrian cell that can be moved after 2 to 3 discrete time steps, and a dashed circle with an adult male The average step length is shown, and the direction of movement is indicated by 24 arrows. As shown in FIG. 9, the movement in the orthogonal direction and 45 ° direction has a free walking speed of 1.3 m / s, and the movement in the other direction has a free walking speed of 1.208 m / s. .

  Next, the local neighborhood rule will be described with reference to FIG. Around the human individual, an invisible kind of so-called personal space (personal space, individual space) is formed. It has been pointed out that when a personal space is invaded due to crowding or the like, humans want to be away from others due to clogging and discomfort. In addition, as for the space during walking, each pedestrian is connected to pedestrians in the front, back, left, and right in a walking zone with a width of about 70 cm and a length of about 100 to 200 cm allowing for a smooth walking motion, and in crowd flow. The distance between pedestrians representing the distance between them (approximately 1m left and right, approximately 1.5m forward) is a past study (for example, “Building Architectural Design Materials“ Human ”(2003 Architectural Institute of Japan)” P58 walking.・ Exercise: Motor ability Relationship between height and stride, gait zone and perception zone, “Architectural design material collection“ Human ”(2003 Architectural Institute of Japan)” P126 Crowd behavior: Flow characteristics Pedestrians (non-commuting crowd) (See spacing pattern (I)).

The personal space is set in the present embodiment in consideration of past studies and actual measurement results and the start of a state change from about 1 person / m ² or more to the crowd flow state influenced by other pedestrians. This personal space is set in a vector in two types of directions, an orthogonal direction vector and a 45 ° direction vector. If the main axis is perpendicular to the direction vector, as shown in FIG. 10 (a), sets the personal space PS _A in the direction perpendicular type. When the main direction axis is a 45 ° direction vector, a 45 ° direction type personal space PS _B is set as shown in FIG. Based on these two types of personal spaces, action determination rules based on local neighborhood rules are set.

Personal space PS _A in the direction perpendicular type, as shown in FIG. 10 (a), is composed of a front direction corresponding to three cells, composed of three pieces of the cell left direction about the pedestrian cell P Is done. Personal space PS _A in the direction perpendicular type, as shown in FIG. 10 (a), consists of the pedestrian cell nine cells forward of P, given the summary number of 0-8 on the nine cells ing. As the distance from the pedestrian cell P along the main direction axis, the first, second, and third columns are set. As shown in FIG. 10 (b), the 45 ° direction type personal space is composed of three cells including the pedestrian cell P in the 45 ° direction, and the cell is centered on the 45 ° direction cell. Is placed. As shown in FIG. 10 (b), the 45 ° direction type personal space PS _B is composed of 8 cells in the 45 ° direction of the pedestrian cell P, and the abbreviated number of 0 to 7 is assigned to the 8 cells. It is attached. As moving away from the pedestrian cell P along the main direction axis, the 0th column, the 1st column, the 2nd column, and the 3rd column are set.

  The action judgment of the local neighborhood rule will be described. When determining where to move in the next step of a pedestrian existing in the pedestrian cell P in the current step, it is determined from the presence / absence and position of other pedestrians in the personal space at the current step, so that the following conditions are satisfied Select Condition 1 is a cell in which the movable cells are indicated by numbers 6, 7, and 8 in the case of the orthogonal direction type personal space, the selection priority is in the order of 7, 6 or 8, and the 45 ° direction type personal space. In this case, the movable cells are cells indicated by the numbers 4, 6, and 7, and the selection priority is in the order of 4, 6 or 7. Condition 2 is a condition in which the movement destination is an enterable cell in the current step. As shown in FIG. 2, the enterable cell is, for example, a floor cell, a staircase cell, an entrance cell, an exit cell, or a destination cell. Condition 3 is a condition that does not adjoin another pedestrian when moving to the next step. Condition 4 is a condition for stopping or changing the main axis when the above three conditions cannot be satisfied at the same time.

When the movement destination cell is determined, the movement probability of the horizontal portion is calculated according to the following equation (3). The basic stop probability is 0.0 when there are no other pedestrians in the personal space, and 1.0 when the nearest pedestrian is in the first row in the personal space. 0.4 when there is a closest pedestrian in the second row, and 0.2 when there is a closest pedestrian in the third row in the personal space. The density coefficient is 1.0 when there are 0 to 2 other pedestrians in the personal space, and 0.6 when there are 3 other pedestrians in the personal space. When there are four other pedestrians, it is 0.3, and when there are five or more other pedestrians in the personal space, it is 0.0. The movement direction probability is 1.0 when the movement destination is an orthogonal cell, and is 1 / √2 when the movement destination is a 45 ° direction cell. Moreover, the movement probability in the staircase is calculated according to the following equation (4). The stair factor is 0.6 when there are 2 or less other pedestrians in the personal space in the downward direction, and 0.8 when there are 3 or more other pedestrians in the personal space in the downward direction. Yes, it is 0.4 when there are two or less other pedestrians in the personal space in the upward direction, and 0.9 when there are three or more other pedestrians in the personal space in the upward direction.

  For example, if the movement probability is 0.3, the cell moves to the movement destination cell with a probability of 3 times during 10 steps, and stops at the current cell with a probability of 7 times. When the movement probability is 1.0, the movement to the movement destination cell is always performed in the next step. When the movement probability is 0.5, the cell moves to the movement destination cell in either the next step or the step after step. However, since the movement probability is calculated for each discrete time step, even if the current movement probability is 0.5, the next movement probability may be a value other than 0.5. It may not move at any of these steps. Based on the behavior judgment based on the movement probability in the local neighborhood rule, 0.23 s of real time of one discrete time step and 30 cm × 30 cm of one cell, the flow coefficient in the crowd flow state is 1.5 person / m · s. Reproduce (1.33 person / m · s on the stairs).

  With reference to FIG. 11, the movement rule (especially exit selection rule) between the residences in the residence before the exit of the pedestrian (refugee) in a living room is demonstrated. In accordance with the theory of the above evacuation prediction model, it is reasonable to think that “refugees want to evacuate to a safe place as soon as possible in an emergency”. Therefore, in the present embodiment, a pedestrian in the living room first selects the nearest evacuable exit and starts evacuation behavior toward the exit. And, from time to time, considering the outflow speed of pedestrians at each exit and the number of people staying in front of the exit, select the exit that is predicted to be able to leave the earliest, and change the target exit to join the stay at another exit Suppose that The basic concept of this selection exit change rule is shown below. In addition, when there is one exit in the living room, that one exit is the target exit, and there is no change of the exit.

First, the pedestrian in the room is set to select the nearest exit, but it seems that the outflow state of his / her selected exit is anxious compared to other exits immediately after the start of evacuation when the first target exit is selected. As shown in FIG. 11A, when an arbitrary pedestrian P (a resident and a refugee) is set, the width of the selected exit E ₁ is B ₁ , the flow coefficient is N ₁ , the exit When the number of people who have flowed out of E ₁ is P _{1_out} and the number of pedestrians near the selected exit E ₁ is P _{1_flo} , the remaining estimated evacuation end time at any discrete time step t of any pedestrian P T ₁ can be calculated by equation (5). However, _{N 1} is calculated by equation (6).

On the other hand, the remaining at any discrete time step t the shortest evacuation completion expected time T _min of the entire room is a time when all ends evacuated simultaneously from all exit the room. Therefore, as shown in FIG. 11B, when the initial number of people in the room is P and the room has _three exits E ₁ , E ₂ , and E ₃ with widths B ₁ , B ₂ , and B ₃ , the shortest The estimated evacuation end time T _min can be calculated by Equation (7). However, the number of pedestrians P _in being in the room at any discrete time step t, are calculated by the equation (8), the flow coefficient N _i of each outlet is calculated by Equation (9).

Here, when evacuating, the occupant (especially in the case of the fire room) “want to evacuate as soon as possible from the room”, so the intuitive evacuation end time T ₁ and the shortest evacuation end prediction time T _min for the selected exit are intuitively determined. I think to compare. However, the person in the room is not a precise computer, because it is human, and the evacuation end expected time T ₁ from their own choice exit to change the selection exit at the time you feel "apparently slower" than the other exit Think.

When evacuation completion expected time T ₁ is considered a case longer than the shortest evacuation estimated end time T _min, although the difference is T ₁ -T _min, the difference of the time difference between the outlet flow coefficient due to the density of the exit destination space Only if it is not significant, it is considered that there will be no major changes until the end of the evacuation. In addition, since the evacuation is started first from the fire chamber, it is considered that the change in the density of the exit space in the fire chamber is unlikely.

Therefore, when M = (T ₁ −T _min ) / T _min , the ratio of M increases with every passing discrete time step. The point in time when M reaches a certain value (for example, 0.1 (= 10%)) is set as the point in time when it is determined that the selection exit is to be changed, because it is felt “slightly later” than the other exits. That is, when (T ₁ −T _min ) / T _min becomes a certain value or more, the change of the selection exit is started.

After starting the change of the selected exit, it is considered that the exit is expected to be evacuated earliest among other exits and the movement is started. As shown in FIG. 11C, when there are exits E ₂ and E ₃ having widths B ₂ and B _{3 in} addition to the initial selection exit E ₁ , the evacuation ends when each exit E ₂ and E ₃ is selected. The expected times T ₂ and T ₃ can be calculated by equations (10) and (12), respectively. However, N ₂ and N ₃ are calculated by the equations (11) and (13), respectively. By comparing the evacuation completion expected time T ₃ for evacuation completion expected time T ₂ and an outlet E ₃ for the exit E _2, selects the exit of the shorter time. Thus, if there are n outlet, calculates the evacuation completion expected time T _i for each outlet, the equation (14), selecting a new outlet is expected to be evacuated earliest, it starts to move. It should be noted that the comparison determination may be made after leaving the initial selection exit as an option, or the comparison decision may be made without leaving it as an option. The above exit selection rule can also be applied in a moving space having a plurality of exits.

  Now, the evacuation behavior simulation system 1 will be specifically described. The evacuation behavior simulation system 1 may be configured by incorporating application software for evacuation behavior simulation into a general-purpose computer such as a personal computer, or may be configured as a dedicated device.

  The evacuation behavior simulation system 1 includes analysis space data reading means 2, simulation condition input means 3, database 4, simulation result output means 5, control means 6, and simulation execution means 7.

  The analysis space data reading means 2 is means for reading analysis space data to be analyzed from the database 4. This reading is performed from the database 4 to a predetermined RAM area in the computer by processing in the computer in response to an instruction from the control means 6 at the start of the simulation. For example, as shown in FIGS. 21, 25, and 30, the analysis space data is specific data of a space (basic state in the absence of a pedestrian) made up of analysis target rooms, hallways, stairs, and the like. In addition, for example, as shown in FIG. 2, the analysis space data includes cell information (cells that can be entered (floor cells, stairs cells, entrance cells, exit cells, destination cells, etc.)), cells that cannot enter ( Column, wall cell, fixture cell, etc.) and the number of floors, etc. These analysis space data are stored in advance in the database 4 by the operator.

  The simulation condition input means 3 is a means for inputting simulation conditions, and a keyboard, a mouse or the like connected to a computer is used. This input is performed by the operator before the simulation starts, and the input simulation conditions are sent to the control means 6. As the simulation conditions, for example, as shown in FIG. 3, there are pedestrian information (initial number of people (initial density), initial position, evacuation route, evacuation start time, etc.), and simulation end conditions (all evacuation completed). .

  The database 4 is a database configured in a predetermined area of a storage device provided inside or outside the computer. The database 4 stores, for example, analysis space data used in the evacuation behavior simulation system 1 and simulation results when the evacuation behavior simulation system 1 performs a simulation. The simulation result includes, for example, drawing data indicating the movement of the pedestrian in the analysis space to be evacuated at each elapsed time (each discrete time step), and numerical data such as the total number of people leaving the passage, the flow coefficient, and the speed. is there.

  The simulation result output means 5 is a means for sequentially outputting the simulation results (or outputting the results of the simulation performed in the past) when the simulation is performed by the evacuation behavior simulation system 1 and connected to the computer. A display or a printer is used. As a simulation result to be output, for example, as shown in FIGS. 24, 28, 29, and 37 to 39, a screen drawing showing a pedestrian's movement in the analysis space of the evacuation target at each elapsed time, FIG. 23, FIG. 26, FIG. 27, and FIGS. 31 to 36, there are numerical output (graph output) of the cumulative number of people leaving and the flow coefficient with respect to the elapsed time.

  The control means 6 performs a main process of simulation in the evacuation behavior simulation system 1. The process by the control means 6 is demonstrated along the flowchart of FIG. The operator activates the evacuation behavior simulation system 1 and inputs simulation conditions through the simulation condition input means 3. In addition, the operator designates an analysis space (for example, a space only for a living room, a one-floor (or one-story) space consisting of a living room or a corridor, a space consisting of a plurality of multi-story floors) to be analyzed. To do. Then, the operator causes the evacuation behavior simulation system 1 to start simulation.

  First, the control means 6 issues a read command to the analysis space data reading means 2 to read the analysis space data from the database 4 (S10). And the control means 6 arrange | positions a pedestrian to each cell in analysis space according to the initial density of simulation conditions according to random or a position (S11). Basically, pedestrians are arranged in each room, but pedestrians may also be arranged in hallways and stairs. Each pedestrian is assigned a pedestrian number (1 to n). Further, the control means 6 determines the maximum walking speed (free walking speed) at which each pedestrian can walk (S12). The upper limit speed that can be set is 1.3 m / s. Then, the following processing is performed every execution time of the discrete time step = 0.23 s.

  The control means 6 sequentially extracts cells from the cells in the analysis space (cell number (1, 1) to cell number (m, n)) (S13). And the control means 6 determines whether the cell is a pedestrian cell (S14).

  When it determines with a pedestrian cell in S14, the control means 6 determines whether the pedestrian of the pedestrian cell is an evacuation completion state (S15). In this determination, for example, in the case of evacuation in the room under the simulation end condition, if the pedestrian has left the room, the evacuation has been completed, and in the case of floor evacuation, the pedestrian has exited from the moving space etc. The evacuation is complete when the vehicle enters or leaves the building, and the evacuation is complete when the pedestrian has left the building in the case of evacuation of the entire building. When it is determined in S15 that the pedestrian is not in the evacuation completion state, the control means 6 performs a pedestrian evacuation action process by the simulation execution means 7 for the pedestrian cell, and moves when the pedestrian cell and the pedestrian move. The scheduled state quantity in the next step for the previous cell is determined (S16). The pedestrian evacuation action process in the simulation execution means 7 will be described in detail later.

  When it determines with it not being a pedestrian cell in S14, the control means 6 determines whether the plan state quantity of the next step about the cell is determined (S17). If it is determined in S17 that the scheduled state quantity of the next step has not been determined, the control means 6 sets the current state quantity as the next scheduled state quantity (S18).

  When the processing for the extracted cells is completed, the control means 6 determines whether or not the scheduled state quantities in the next step of all the cells (cell number (1, 1) to cell number (m, n)) have been determined. Determine (S19). If it is determined in S19 that the planned state quantity in the next step of all the cells has not been determined, the control means 6 returns to the process of S13, performs the process for the cell of the next cell number, and in the analysis space The processing is sequentially performed on all the cells.

  When it is determined in S19 that the planned state quantity in the next step of all the cells has been determined, the control means 6 transitions the state quantity of all the cells with the planned state quantity and adds 1 to the number of steps (S20). . Then, the control means 6 issues a drawing command to the simulation result output means 5 and draws a screen showing the movement of the pedestrian in the analysis space at the current step (S21). Further, the control means 6 issues a file output command to the simulation result output means 5 and outputs data such as the total number of people leaving the current step and the flow coefficient in a file (S22). This output data is output by the operator.

  Then, the control means 6 determines whether or not there is no pedestrian cell from the analysis space (whether evacuation is completed) (S23). If it is determined in S23 that there is no pedestrian cell from the analysis space, the control means 6 returns to S13 and performs the process in the next step. When it is determined in S23 that there are no pedestrian cells from the analysis space, the control means 6 ends the simulation. In this determination, for example, in the case of evacuation in the room under the simulation end condition, the simulation ends (evacuation completion) when all pedestrians have left the room in the analysis space, and in the case of floor evacuation, the floor of the analysis space. When all the pedestrians have exited, entered the stairs or exited outdoors, the simulation is complete (evacuation completed). In the case of evacuation of the entire building, all pedestrians have exited from the building in the analysis space. The simulation is complete (evacuation completed). The time from the start to the end of the simulation is the predicted time for the evacuation action.

  The simulation execution means 7 performs a pedestrian evacuation action process for each pedestrian cell, and determines a planned state quantity (moving or stopped at the destination cell) in the next step of the pedestrian. The process by the simulation execution means 7 is demonstrated along the flowchart of FIG.

  In the simulation execution means 7, first, the pedestrian number of the pedestrian existing in the pedestrian cell is acquired (S30). Then, the simulation execution means 7 determines whether or not the pedestrian (pedestrian cell) is in the living room (S31). If it is determined in S31 that the user is not in the room, the simulation execution means 7 determines whether a selection exit has been determined (S32). If it is determined in S32 that the selection exit has not been determined, the simulation execution means 7 determines the selection exit (S33). For example, in the case of a pedestrian in the corridor, the exit of the corridor is the selected exit (however, in the case of a pedestrian in the first floor corridor, the exit of the building is the selected exit), and in the case of a pedestrian on the stairs The exit of the stairs will be the selected exit. When the simulation end condition is room evacuation, the determination in S31 is only YES, and the processes in S32 and S33 are not performed.

  When it is determined in S31 that the user is in the room, the simulation execution means 7 determines whether or not the evacuation start time of the room in which the pedestrian is present has elapsed (S34). When it is determined in S34 that the evacuation start time of the room has not elapsed, the simulation execution means 7 updates the pedestrian information without setting each process, and sets the scheduled state quantity (stop) for the next step. (S45). For example, when evacuating due to a fire, evacuation is started first from the fire room, and evacuation from other rooms starts after information on the fire is transmitted. The start time is set to a time when a predetermined time has elapsed from the reference time. Moreover, even if it is a pedestrian in the same living room, you may change evacuation start time for every pedestrian. When the simulation end condition is a room evacuation, when the same evacuation start time is set for all pedestrians in the room, the determination in S34 is only YES, and a different evacuation start time for pedestrians in the room. When it is set, the determination in S34 is YES / NO.

  When it is determined in S34 that the evacuation start time of the room has elapsed, the simulation execution means 7 determines whether or not the priority order of selection exits in the room is determined (S35). If it is determined in S35 that the priority order of selection exits has not been determined, the simulation execution means 7 determines the priority order of each selection exit in the room (S36). Here, the priority order is determined according to the movement rule between the pedestrians staying in the room. Then, the simulation execution means 7 determines whether or not evacuation is possible in the shortest time from the living room at the first selection exit (S37). Here, the determination is made by comparing the predicted evacuation end time for each exit. If it is determined that the evacuation is impossible in the shortest time from the room at the first selection exit at S37, the simulation execution means 7 changes the selection exit according to the priority order (S38). In addition, when there is one exit of a living room, each determination of S35 and S37 is only YES.

  When the selection exit is determined, the simulation execution means 7 determines the direction vector of the first candidate and the direction vector of the second candidate from the positional relationship between the selection exit and the pedestrian, and the direction vector of the first candidate and the second candidate. A main direction axis is determined from the direction vector (S39). Then, the simulation execution means 7 performs an action determination based on the local neighborhood rule using the determined main direction axis (S40). Here, when the main direction axis is an orthogonal direction vector, an orthogonal type personal space is set, and when the main direction axis is a 45 ° direction vector, a 45 ° direction personal space is set and moved in the personal space. Decide the destination cell (stop or change main direction axis if there is no destination cell), position of other pedestrians in the personal space, number of people and direction of movement (in the case of stairs, consider up and down) Accordingly, the movement probability is calculated, and whether to move to the movement destination cell in the next step or to stop in the current cell in the next step is determined according to the movement probability.

  In the simulation execution means 7, it is determined whether or not the destination cell in the next step has been determined by action determination based on the local neighborhood rule (S41). If it is determined in S41 that the destination cell in the next step has been determined, the simulation execution means 7 updates the pedestrian information and sets the scheduled state quantity (moving to the determined destination cell) in the next step ( S45).

  If it is determined in S41 that the destination cell in the next step has not been determined, the simulation execution means 7 determines whether or not this pedestrian has stopped in the previous step (S42). If it is determined in S42 that the operation has stopped in the previous step, the simulation execution means 7 changes the remaining direction vector as the main direction axis (S43). Then, the simulation execution means 7 performs an action determination based on the local neighborhood rule using the changed main direction axis (S44). And in the simulation execution means 7, when the pedestrian information is updated, and the behavior determination based on the local neighborhood rule is performed in S44, the scheduled state quantity of the next step is set based on the result of the behavior determination, If the action determination based on the local neighborhood rule is not performed in S44, the scheduled state quantity of the next step is set based on the result of the action determination in S40 (S45).

  Next, an example in which a simulation is actually performed by the evacuation behavior simulation system 1 according to the present embodiment will be described. Here, a simulation using a simple passage model, a simulation using a simple staircase model, a simulation of evacuation using a model of any room in a building, a room using a model of any floor in the building + Example of simulation of corridor evacuation Each example of simulation of living room + corridor + stair evacuation using models of the second and third floors of a three-story building is shown.

  First, an example of simulation in a passage will be described with reference to FIGS. FIG. 14 shows passage analysis space data (passage model). FIG. 15 shows the result of simulation (relationship between density and speed) in each passage model of FIG. FIG. 16 shows the result of simulation (relationship between density and flow coefficient) in each passage model of FIG. FIG. 17 is a screen drawing of the simulation in the orthogonal path model of FIG.

  When a simulation is performed using analysis space data for an actual building such as a living room, living room + passage, living room + passage + stairs, etc., it is difficult to maintain a fluid state at a uniform density at each exit and each confluence. It is difficult to stably show the reproduction of the relationship between density and flow coefficient. Therefore, a simple passage is simulated by the evacuation behavior simulation system 1 according to the present embodiment, and a speed and a flow coefficient (particularly, crowds) that are targets for reproduction in horizontal walking for each density of pedestrians. It shows that 1.5 person / m · s) is obtained in the fluidized state.

The analysis space data will be described. As shown in FIG. 14, the analysis space data includes an orthogonal path (total length 23.40 m, width 3.00 m), 1/3 oblique path (total length 24.66 m, width 3.03 m), 1 / This is a four-passage model consisting of two oblique passages (total length 22.80 m, width 3.02 m) and 45 ° direction passage (total length 22.06 m, width 2.97 m). In each passage model, the right end is spatially connected to the left end of the same passage model, and the pedestrian loops in the space at a constant density. The 1/3 skew direction indicates a tan ⁻¹ (1/3) direction (ie, 18.4 ° direction), and the 1/2 skew direction indicates a tan ⁻¹ (1/2) direction. (Ie, 26.6 ° direction).

The simulation conditions will be described. The initial density is in increments of 0.1 person / m ² in the range of 0.0 to 4.0 person / m ² , and the position of the pedestrian is random. One discrete time step is 0.23 s. The speed is calculated from the average time required for each pedestrian to move from the left end to the right end of the passage model. The flow coefficient is calculated by measuring the number of people who have passed through a fixed section (eg, the right end section) of the passage model per 260 time steps (about 60 seconds).

In the evacuation behavior simulation system 1, the simulation was performed five times for each path model for each path model. FIG. 15 shows the relationship between density and speed obtained from the average of these five simulations, and FIG. 16 shows the relationship between density and flow coefficient. In addition, with respect to the passage model in the orthogonal direction, the density is 0.5 person / m ² , 1.0 person / m ² , 2.0 person / m ² , 3.0 person / m ² , 4.0 person / m ^2. FIGS. 17A to 17E show the drawing results of the screen showing the movement of the pedestrian in the case of.

In FIG. 15, the graph indicated by the broken line _VP0 is the speed of the reproduction target, and corresponds to the relationship between the density and speed in the horizontal portion walking shown in FIG. Further, the graph indicated by the solid line V _P1 is the simulation result of the speed in the orthogonal passage model, the graph indicated by the solid line V _P2 is the simulation result of the speed in the oblique 1/3 direction passage model, and the solid line V _P3 The graph shown by is a simulation result of the speed in the diagonal ½ direction passage model, and the graph shown by the solid line _VP4 is the simulation result of the velocity in the 45 ° direction passage model. When the simulation results V _P1 , V _P2 , V _P3 , and V _P4 of the speed of each passage model are compared with the speed V _{P0 of the} reproduction target, the target speed can be reproduced with high accuracy.

In FIG. 16, the graph indicated by the broken line _FP0 is the flow coefficient of the reproduction target, and corresponds to the relationship between the density and the flow coefficient in the horizontal walking shown in FIG. The graph indicated by the solid line _FP1 is a simulation result of the flow coefficient in the orthogonal passage model, and the graph indicated by the solid line _FP2 is the simulation result of the flow coefficient in the oblique ３ direction passage model, which is indicated by the solid line. a graph indicated by F _P3 is the result of a simulation of the flow coefficient in the oblique 1/2 direction of passage models, the simulation results of the flow coefficient of a graph is 45 ° direction of the passage model shown by the solid line F _P4. Comparing the simulation results _{F P1, F P2, F P3} , F P4 and reproduction target flow coefficient _{F P0} flow coefficient of each passage models are able to reproduce the flow coefficient as a target with high precision. In particular, a flow coefficient of approximately 1.5 person / m · s can be reproduced in the crowd flow state.

  Next, with reference to FIGS. 18-20, the Example of the simulation in a staircase is described. FIG. 18 shows staircase analysis space data (stair model). FIG. 19 shows the result of simulation (relationship between density and speed) in the staircase model of FIG. FIG. 20 shows a simulation result (relationship between density and flow coefficient) in the staircase model of FIG.

  Here again, a simulation is performed by the evacuation behavior simulation system 1 according to the present embodiment for a simple staircase, and for each density of pedestrians, the speed to be reproduced in staircase walking (up and down) It shows that a flow coefficient (particularly 1.33 person / m · s in a crowd flow state) is obtained.

  The analysis space data will be described. As shown in FIG. 18 (a), the analysis space data is obtained from the 3rd floor to the 2nd floor (passage width 1.2m, staircase length 3.0m, landing length 1.2m), FIG. 18 (b). As shown, there are two staircase models from the second floor to the first floor (passage width 1.2m, staircase section length 3.0m, landing hall length 1.2m). Here, the 1st floor stairway exit SE and the 3rd floor stairway entrance SG are spatially connected, and a pedestrian loops in this space at a constant density.

The simulation conditions will be described. The initial density is in increments of 0.1 person / m ² in the range of 0.0 to 4.0 person / m ² , and the position of the pedestrian is random. One discrete time step is 0.23 s. The descending speed is calculated from the average time required for each pedestrian to move from the third-floor staircase entrance SG to the first-floor stairway exit SE of the staircase model. The ascending speed is calculated from the average time required for each pedestrian to move from the first floor stairway exit SE to the third floor stairway entrance SG of the staircase model. The flow coefficient is calculated by measuring the number of people who have passed through a fixed section of the staircase model (for example, a cross section at the first floor stairway exit SE) per 260 time steps (about 60 seconds).

  In the evacuation behavior simulation system 1, the simulation was performed five times in the upward direction and the downward direction using two staircase models. FIG. 19 shows the relationship between the density and speed determined from the average of five simulations in the upward and downward directions, and FIG. 20 shows the relationship between the density and the flow coefficient.

In FIG. 19, the graph indicated by the broken line _VSU0 is the speed of the reproduction target in the upward direction, and corresponds to the relationship between the density and speed in the staircase (upward) walking shown in FIG. Further, the graph indicated by the broken line _VSD0 is the speed of the reproduction target in the downward direction, and corresponds to the relationship between the density and speed in the staircase (downward) walking shown in FIG. Also, the speed of the simulation results for the uplink of the graph staircase model shown by the solid line V _SU1, a graph indicated by a solid line V _SD1 is the speed of the simulation results for the downlink stairs model. When the simulation result V _SU1 of the speed in the up direction of the stair model is compared with the speed V _{SU0 of the} reproduction target, the target speed in the up direction can be reproduced with high accuracy. Further, when the simulation result V _SD1 of the speed in the down direction of the staircase model is compared with the speed V _{SD0 of the} reproduction target, the target speed in the down direction can be reproduced with high accuracy.

In FIG. 20, the graph indicated by the broken line _FSU0 is the flow coefficient of the upward reproduction target, which corresponds to the relationship between the density and the flow coefficient in the staircase (upward) walking shown in FIG. 6B. Further, the graph indicated by the broken line _FSD0 is the flow coefficient of the reproduction target in the downward direction, and corresponds to the relationship between the density and the flow coefficient in the staircase (downward) walking shown in FIG. Further, the result of a simulation of the flow coefficient in the up direction of the graph staircase model shown by the solid line F _SU1, a graph indicated by the solid line F _SD1 is a simulation result of the flow coefficient in the downstream direction of the stairs model. When the simulation result F _SU1 of the flow coefficient in the upward direction of the staircase model is compared with the flow coefficient F _{SU0 of the} reproduction target, the target flow coefficient in the upward direction can be reproduced with high accuracy. Further, when the simulation result F _SD1 of the flow coefficient in the downward direction of the staircase model is compared with the flow coefficient F _{SD0 of the} reproduction target, the target flow coefficient in the downward direction can be reproduced with high accuracy. In particular, a flow coefficient of approximately 1.33 person / m · s can be reproduced in the crowd flow state.

  Next, with reference to FIG. 21 to FIG. 24, an example of simulation in room evacuation will be described. FIG. 21 is analysis space data of a living room. FIG. 22 shows the result of simulation of evacuation (relationship between elapsed time and cumulative number of people leaving). FIG. 23 shows a simulation result of room evacuation (relationship between elapsed time and flow coefficient). FIG. 24 is a screen drawing of a room evacuation simulation. The simulation target is one living room on an arbitrary floor of the building, and evacuation is completed when all pedestrians leave the living room.

The analysis space data will be described. The analysis space data is a room R of 9.0 m × 12.0 m as shown in FIG. The room R, there is a right exit _{E R} of the left outlet _{E L} and width 1.8m width 0.9 m. Evacuation is complete when the exits E _L and E _R are exited.

The simulation conditions will be described. The initial density is 0.93 person / m ² (the number of people in the room is 100), and the position of the pedestrian is random. The evacuation route is the route with the selected exit as the shortest distance. The evacuation start time is the same for all members. The evacuation end condition is that all pedestrians who have been evacuated have left the room.

  In the evacuation behavior simulation system 1, evacuation simulation was performed using this room model. FIG. 22 shows the relationship between the elapsed time from the start of evacuation and the cumulative number of people leaving the room, and FIG. 23 shows the relationship between the elapsed time and the flow coefficient when passing through each exit of the room. In addition, the drawing result of the screen showing the movement of the pedestrian in the living room model at the elapsed time of 0.0 s, 4.3 s, 9.2 s, 16.1 s, 23.0 s, and 28.75 s is shown in FIG. ) To (f).

In Figure 22, a total number of simulation results of pedestrians graph has exited from the left outlet E _L shown by a solid line P _RL, simulation of the cumulative number of pedestrians graph shown by a solid line P _RR has exited from the right outlet E _R As a result, the elapsed time from the start of evacuation is 28.75 s, and all the pedestrians have completed the exit. In Figure 23, the result of a simulation of the flow coefficient graph shown by the solid line F _RL is in the left exit E _L, a graph indicated by the solid line F _RR is a simulation result of the flow coefficient in the right outlet E _R. From the simulation results F _RL and F _RR at the exits E _L and E _R , a flow coefficient of approximately 1.5 person / m · s can be reproduced when passing through the exit where the density increases.

As shown in FIG. 24A, pedestrians are randomly placed in the living room R at the start of evacuation. As shown in FIG. 24 (b), and after 4.3s from the evacuation starts, the outlet _E L towards the pedestrian is close, gathered in _{E R,} crowd flow state from the free walking condition each outlet _E L, with _{E R} The flow coefficient at the exits E _L and E _R is approaching 1.5 person / m · s. As shown in FIG. 24 (c), and after 9.2s from the evacuation starts, the outlet _E L, has become a crowd fluidized state _{E R,} approximately 1 flow coefficient at the exit _E L, _{E R.} 5 people / m · s. As shown in FIG. 24 (d), even after 16.1s from the start of evacuation, the exit E _L and E _{R are} in a crowded flow state, and the flow coefficient at each exit E _L and E _R is approximately 1. 5 people / m · s. As shown in FIG. 24 (e), and after 23.0s from the evacuation started immediately before evacuation completed, each exit _E L, has shifted to the free walking state from the crowd flow state _{E R,} each outlet _E L, E The flow coefficient at _R is smaller than 1.5 person / m · s. As shown in FIG. 24 (f), 28.75 s after the start of evacuation, all pedestrians leave the room R and evacuation is complete.

  Next, with reference to FIG. 25 to FIG. 29, an example of simulation in the room + corridor evacuation (floor evacuation) will be described. FIG. 25 is analysis space data of a living room + corridor (one floor). FIG. 26 shows the result of the simulation of the room + corridor evacuation (relationship between elapsed time and cumulative number of people leaving). FIG. 27 shows the result of simulation of living room + corridor evacuation (relationship between elapsed time and flow coefficient). 28 and 29 are screen drawings of a simulation of a living room + corridor evacuation. The simulation target is one floor of an arbitrary floor of the building, and the evacuation is completed when all pedestrians leave the hallway of the floor.

The analysis space data will be described. Analysis space data, as shown in FIG. 25, three room _{R A,} R B, is a one floor consisting of _{R C} and one corridor C. The A room _RA is 6.0 m × 9.0 m, and there is an exit _ERA with a width of 1.5 m connected to the hallway C. B room _{R B} is 6.0 m × 4.5 m, there is an outlet _{E RB} width 0.9m leading to the corridor C. C room _RC is 7.8 m × 4.5 m, and there is an exit E _RC with a width of 1.2 m that leads to corridor C. Corridor C is 1.8 m × 16.5 m, there is an outlet _{E C} width 0.9m leading to stairs. Is the evacuation is complete at the time of the exit the exit E _C.

The simulation conditions will be described. The initial density is 1 person / m ^{2 for} all three rooms R _A , R _B , and R _C (the number of people in the room is R _A : 54 people, R _B : 27 people, R _C : 35 people), walking The position of the person shall be random. The evacuation route is the route with the selected exit as the shortest distance. In addition, since there is one exit, the one exit is a selective exit. The evacuation start time is the same for all members (all rooms are simultaneously). The evacuation end condition is that all evacuated pedestrians exit the corridor.

  In the evacuation behavior simulation system 1, a simulation of the evacuation of the living room and the hallway was performed using this room and the hallway model. FIG. 26 shows the relationship between the elapsed time from the start of evacuation by this simulation and the cumulative number of people leaving the room and hallway. FIG. 27 shows the relationship between the elapsed time and the flow coefficient when passing through the exits of the room and hallway. Shown in Moreover, the drawing result of the screen which shows the mode of the movement of the pedestrian in the room + corridor model in elapsed time 0.0s, 9.2s, and 19.55s is shown to Fig.28 (a)-(c), and elapsed time 29 (a) to 29 (c) show the drawing results of the screen at 29.90s, 68.77s, and 98.21s.

In Figure 26, a is A room _R simulation cumulative number from the exit _{E RA} pedestrian who exit _A graph indicated by the solid line _{P RA,} a graph indicated by a solid line _{P RB} is exiting from the exit _{E RB} of B room _{R B} was a pedestrian simulation cumulative number result is the C room R _C outlet E cumulative number of simulations pedestrian left the _RC results of the graph indicated by the solid line P _RC, graph corridor C indicated by the solid line P _C it is a simulation result the cumulative number of pedestrian has exited from the exit E _C of. The elapsed time from the evacuation start is the exit completion of all of the pedestrian from C room R _C in 19.55s, the elapsed time is the exit completion of all of the pedestrian from the B-room R _B in 29.90s, the elapsed time There is a leaving completion of all pedestrians a room _{R a} in 68.77S, an elapsed time exit completion of all pedestrians corridor C in 98.21S (i.e., floors evacuation completed).

In Figure 27, the result of a simulation of the flow coefficient graph shown by the solid line _{F RA} is at the exit _{E RA} of A room _{R A,} in the simulation results of the flow coefficient graph shown by the solid line _{F RB} is at the exit _{E RB} of B room _{R B} There, the result of a simulation of the flow coefficient graph shown by the solid line F _RC is at the exit E _RC of C room R _C, the graph indicated by the solid line F _C is a simulation result of the flow coefficient at the outlet E _C corridor C. C room _{R C} and simulation in the corridor C result _F RC, from _{F C,} which can reproduce flow coefficient of approximately 1.5 people / m · s when the outlet passage density becomes high. Also, in the case of A room _{R A} and B room _{R B,} outlet _{E RA,} since merges with pedestrians evacuated from the back of the room as it exits the corridor C from _{E RB,} outlet _{E RA, E RB} (especially , Exit E _RA ), and the flow coefficient becomes smaller than 1.5 person / m · s.

As shown in FIG. 28 (a), at the start of evacuation, pedestrians are randomly arranged in the rooms R _A , R _B and _RC . As shown in FIG. 28 (b), and after 9.2s from the evacuation starts, each room _{R A,} R B, each pedestrian each exit _E RA of _{R C,} E _RB, gathered in _{E RC,} the corridor C I'm starting to leave. The flow coefficient at exit E _RC in room _{C RC} is approximately 1.5 persons / m · s, but at the exits E _RA and E _RB in room _{A RA} and room _{B RB} , evacuate from the back room. The flow coefficient is smaller than 1.5 people / m · s due to the influence of pedestrians. The exit E _{C of the} corridor C is still in a free walking state, and the flow coefficient is smaller than 1.5 person / m · s. As shown in FIG. 28 (c), after 19.55s from the start of evacuation, all pedestrians leave the C room _RC and evacuation is complete. In each of the exits E _RA and E _{RB of the} A room R _A and the B room R _B , the flow coefficient is smaller than 1.5 persons / m · s due to the influence of the pedestrian evacuating from the back room, and the A room R _A Is particularly small. The exit E _{C of the} corridor C is still in a free walking state, and the flow coefficient is smaller than 1.5 person / m · s.

As shown in FIG. 29 (a), and after 29.90s from the evacuation starts, the pedestrian everyone left the B room _{R B,} is a refuge completion. A room flow coefficient under the influence of pedestrians have been evacuated from the living room of the back in the exit E _RA of R _A is smaller than 1.5 people / m · s. The exit E _{C of the} corridor C is in a crowd flow state, and the flow coefficient is approximately 1.5 person / m · s. As shown in FIG. 29 (b), after 68.77s from the start of evacuation, all pedestrians leave the A room _RA and evacuation is complete. The exit E _{C of the} corridor C is in a crowd flow state, and the flow coefficient is approximately 1.5 person / m · s. As shown in FIG. 29 (c), 98.21 s after the start of evacuation, all the pedestrians have left the corridor C and the evacuation is completed.

  Next, with reference to FIG. 30 to FIG. 39, an example of simulation in the living room + corridor + stair evacuation (entire building evacuation) will be described. FIG. 30 is analysis space data of a room + a hallway + a staircase (a second floor part and a third floor part of the entire building). FIG. 31 shows the result of simulation of living room + corridor + stair evacuation (relationship between elapsed time on 3rd floor and cumulative number of people leaving). FIG. 32 shows the result of the simulation of the room + corridor + stair evacuation (relationship between the elapsed time on the second floor and the total number of people leaving). FIGS. 33 and 34 are the results of simulation of living room + corridor + stair evacuation (relationship between elapsed time and flow coefficient on the third floor). FIGS. 35 and 36 are the results of the simulation of the room + corridor + stair evacuation (relationship between elapsed time and flow coefficient on the second floor). 37 to 39 are screen drawings of a simulation of a living room + corridor + stair evacuation. The simulation target is a three-story building, and when all the pedestrians on the second and third floors have left the first floor stairs, the evacuation is completed and only the second and third floor parts are simulated.

The analysis space data will be described. As shown in FIG. 30, the analysis space data includes three rooms R _3A , R _3B , R _3C and one hallway C ₃ on the third floor, and three rooms R _2A , R _2B , There are corridor _{C 2} of R _2C and one, there is a stair _{S 21} from the third floor and stair _{S 32} to the second floor to the first floor from the second floor. The A-room R _3A on the third floor is 6.0 m × 9.0 m, and there is an exit E _3RA with a width of 1.5 m that leads to the hallway C ₃ . B room R _3B on the third floor is 6.0 m × 4.5 m, and there is an exit E _3RB with a width of 0.9 m that leads to hallway C ₃ . C room R _3C on the third floor is 7.8 m × 4.5 m, and there is an exit E _3RC with a width of 1.2 m that leads to hallway C ₃ . 3 floor corridor _{C 3} are 1.8 m × 16.5 m, there is an outlet _{E 3C} width 0.9m leading to stairs _{S 32.} The stair S _32, are connected to the space in a stepwise S ₂₁ as indicated by broken line arrows in FIG. The A-room R _2A on the second floor is 6.0 m × 9.0 m, and there is an exit E _2RA with a width of 1.5 m that leads to the hallway C ₂ . B room R _2B on the second floor is 6.0 m × 4.5 m, and there is an exit E _2RB with a width of 0.9 m that leads to corridor C ₂ . The second-floor C room R _2C is 7.8 m × 4.5 m and has an exit E _2RC with a width of 1.2 m that leads to the hallway C ₂ . Second floor corridor _{C 2} is 1.8 m × 16.5 m, there is an outlet _{E 2C} width 0.9m leading to stairs _{S 21.} The stair _{S 21,} there is an outlet _{E 1} width 1.2m in one archive. It is a leaving complete when exited the exit E _1.

The simulation conditions will be described. The initial density is 1 person / m ^{2 for} all three rooms R _3A , R _3B , R _3C , R _2A , R _2B , R _{2C on} each floor (the number of people in the room is R _3A : 54 people, R _3B : 27 people) R _3C : 35 people, R _2A : 54 people, R _2B : 27 people, R _2C : 35 people), and the position of the pedestrian is random. The evacuation route is the route with the selected exit as the shortest distance. In addition, since there is one exit, the one exit is a selective exit. The evacuation start time is the same for all members (all rooms are simultaneously). Evacuation termination condition, pedestrian all of refuge subject is in complete exit from the exit E ₁ of the stairs S _21. Note that the flow coefficient of the corridor _C 3, _{C 2} and stairs _{S 32, S 21} portion, and the way in providing a corridor passage section _{H 3C, H 2C} and stairs passage section _{H 32, H 21,} 20 discrete time step Calculated by measuring the number of people who passed each cross section.

  In the evacuation behavior simulation system 1, a simulation of evacuation of the living room + hallway + staircase was performed using this room + hallway + staircase model. And the relationship between the elapsed time from the start of evacuation by this simulation and the total number of people leaving each room and hallway and stairs on each floor is shown in FIGS. 31 and 32. The elapsed time and the passage time of each room on each floor and the hallway 33 to 36 show the relationship between the flow coefficient at the time of exit passage and during passage of the stairs, and at the time of exit passage of the stairs and during passage of the stairs. Moreover, the drawing result of the screen which shows the mode of the movement of the pedestrian in the room + corridor + staircase model in elapsed time 0.0s and 11.50s is shown to Fig.37 (a), (b), and elapsed time 57 38 (a) and (b) show the drawing results of the screen at .50s and 103.50s, and the drawing results of the screens at the elapsed time of 149.50s and 184.46s are shown in FIGS. ).

In Figure 31, a graph 3 floor A room _{R 3A} simulation results pedestrian cumulative number of which exits from the exit _{E 3RA} that indicated by the solid line _{P 3RA,} a graph indicated by a solid line _{P 3 RBs} is 3 floor B room _{R 3B} of a total number of simulation results of a pedestrian who left the exit E _{3 RBs,} a graph is three floors C room R _3C outlet E exited pedestrian simulation cumulative number results from _3Rc of indicated by the solid line P _3Rc, A graph indicated by a solid line P _3C is a simulation result of the cumulative number of pedestrians exiting from the exit E _{3C of the} third-floor corridor C ₃ . The elapsed time from the evacuation start is the exit completion of all of the pedestrian from the third floor of C living room _{R 3C} in 20.47s, the elapsed time is withdrawal of all of the pedestrian from the third floor of the B room _{R 3B} in 27.83s is completed, the elapsed a time exit completion of all persons from the third floor of a room _{R 3A} in 58.88S, elapsed time exit completion of all persons from the third floor hallway _{C 3} at 99.82s (That is, the third floor evacuation is completed).

In Figure 32, a graph is upstairs A room _{R 2A} simulation results pedestrian cumulative number of which exits from the exit _{E 2RA} that indicated by the solid line _{P 2RA,} a graph indicated by a solid line _{P 2 RB} are upstairs B room _{R 2B} of a total number of simulation results of a pedestrian who left the exit E _{2 RB,} a graph is second-order C room R _2C outlet E exited pedestrian simulation cumulative number results from _2RC of indicated by the solid line P _2RC, The graph indicated by the solid line P _2C is a simulation result of the cumulative number of pedestrians exiting from the exit E _{2C of the} second-floor corridor C ₂ , and the graph indicated by the solid line P ₁ is a pedestrian exiting from the exit E _{1 of the} stairs S ₂₁ It is a simulation result of the cumulative number of. Elapsed time from the evacuation start is all pedestrian exit completion from the second floor C room _{R 2C} in 21.16S, elapsed time exit all pedestrians from the second floor of the B room _{R 2B} in 27.83s is completed, the elapsed time is all pedestrian exit completion from the second floor of a room _{R 2A} in 75.90S, elapsed time second floor exit completion of all pedestrians corridor _{C 2} in 152.26s (i.e., the second floor evacuation completed), and the elapsed time exit completion of all pedestrians from the exit _{E 1} stairs _{S 21} in 184.46S (i.e., the entire evacuation completed).

In Figure 33, the result of a simulation of the flow coefficient at the outlet _{E 3RA} graphs third floor of A room _{R 3A} indicated by the solid line _{F 3RA,} the graph indicated by the solid line _{F 3 RBs} is at the exit _{E 3 RBs,} third floor B room _{R 3B} It is a simulation result of a flow coefficient, and the graph shown by the solid line F _3RC is the simulation result of the flow coefficient at the exit E _3RC of the C room R _3C on the third floor. In FIG. 34, the graph indicated by the solid line F _3C is the simulation result of the flow coefficient at the exit E _3C of the third floor corridor C ₃ , and the graph indicated by the solid line F _3H is the flow coefficient at the middle H _3C of the third floor corridor C _3. The graph shown by the solid line F ₃₂ is the simulation result of the flow coefficient in the middle H ₃₂ of the stairs S ₃₂ . In C room R _3C on the third floor, a flow coefficient of approximately 1.5 person / m · s can be reproduced at the exit where the density increases. Further, if the third floor of A room _{R 3A} and B room _{R 3B,} the outlet _{E 3RA,} since merges with pedestrians evacuated from the back of the room as it exits from the _{E 3 RBs} in the corridor _{C 3,} an outlet _{E 3RA,} E _{3 RBs} (in particular, exit _{E 3RA)} crowded with the flow coefficient is lower than 1.5 people / m · s. In addition, in the middle H _{3C of} the third-floor corridor C ₃ , the flow coefficient is substantially smaller than 1.5 person / m · s. In addition, the exit E _3C of the corridor C ₃ is smaller than the width of the staircase and is not affected by the flow coefficient in the staircase, and is not affected by merging with pedestrians from the upper floor. The flow coefficient of / m · s can be reproduced. In addition, in the middle H ₃₂ of the stairs S ₃₂ from the third floor to the second floor, crowded under the influence of the third floor of the confluence of the pedestrian exit from each room of the pedestrian and the second floor that was exit from each room, 1.33 It becomes smaller than the flow coefficient in the staircase portion of person / m · s.

In Figure 35, the result of a simulation of the flow coefficient at the outlet _{E 2RA} the graph shown by the solid line _{F 2RA} the second floor A room _{R 2A,} the graph indicated by the solid line _{F 2 RB} is at the exit _{E 2 RB} upstairs B room _{R 2B} It is a simulation result of a flow coefficient, and the graph shown by the solid line F _2RC is the simulation result of the flow coefficient at the exit E _2RC of the C room R _2C on the second floor. Also, in FIG. 36, the graph indicated by the solid line F _2C is the simulation result of the flow coefficient at the exit E _2C of the second floor corridor C ₂ , and the graph indicated by the solid line F _{2H is} at the middle H _2C of the second floor corridor C ₂ . It is a simulation result of the flow coefficient, and a graph indicated by a solid line F ₂₁ is a simulation result of the flow coefficient in the middle H ₂₁ of the stairs S ₂₁ . From the simulation result F _2RC in the C room R _2C on the second floor, a flow coefficient of approximately 1.5 person / m · s can be reproduced at the exit where the density increases. Also, in the case of the second floor of A room _{R 2A} and B room _{R 2B,} outlet _{E 2RA,} since merges with pedestrians evacuated from the back of the room as it exits from the _{E 2 RB} corridor _{C 2,} exit _{E 2RA,} E _{2 RB} (in particular, exit _{E 2RA)} crowded with the flow coefficient is less than 1.5 people / m · s. Further, in the second floor of the corridor C ₂ outlet E _2C, crowded pedestrian impact on the second floor of pedestrians and stairs has exited from the room, the flow coefficient is less than 1.5 people / m · s Become. Further, in the second floor corridor _{C 2} of the middle _{H 2C,} the flow coefficient is less than approximately 1.5 people / m · s. In addition, in the middle H ₂₁ of the stairs S ₂₁ from the second floor to the first floor, 3 Cuyahoga second floor will be crowded with pedestrians exited from each room, because there is no influence of the confluence of the pedestrian from the lower floor The flow coefficient at the staircase portion of approximately 1.33 person / m · s can be reproduced.

As shown in FIG. 37 (a), the evacuation starts, 3 floor and second floor of the room _{R 3A, R 3B, R 3C} , R 2A, R 2B, pedestrian _{R 2C} are randomly arranged. As shown in FIG. 37 (b), after 11.50s from the start of evacuation, the pedestrians in the third and second floor rooms R _3A , R _3B , R _3C , R _2A , R _2B , R _2C exit _{E 3RA, E 3RB, E 3RC} , E 2RA, E 2RB, gathered in _{E 2RC,} have started out on each floor of the corridor _C 3, _{C 2.} In each of the exits E _3RC and E _{2RC of the} C room R _3C and R _2C , the flow coefficient is approximately 1.5 person / m · s, but the exits of the A room R _3A and R _2A and the B room R _3B and R _{2B E 3RA, E 2RA, E 3RB} , flow coefficient under the influence of pedestrians have been evacuated from the living room of the back in the _{E 2RB} is less than 1.5 people / m · s. The exits E _3C and E _{2C of the} corridors C ₃ and C ₂ and the H _3C and H _2C on the way are still in a free walking state, and the flow coefficient is smaller than 1.5 person / m · s. Further, the middle _{H 32, H 21} in each step _{S 32, S 21} are still free walking state, the flow coefficient is less than 1.5 people / m · s.

As shown in FIG. 38 (a), and after 57.50s from the evacuation started, each floor B room _{R 3B, R 2B} and C room _{R 3C,} pedestrian everyone left the _{R 2C,} a refuge completion. In the A rooms R _3A and R _{2A on} each floor, the flow coefficient is smaller than 1.5 persons / m · s due to the influence of pedestrians evacuating from the back room. The exit E _3C of the third-floor corridor C ₃ is in a crowd flow state and has a flow coefficient of approximately 1.5 person / m · s. The exit E _2C of the corridor C _{2 on} the second floor has a flow coefficient smaller than 1.5 person / m · s due to the influence of pedestrians evacuating from the third and second floors (stair congestion). 3 floor midway _{H 3C} corridor _{C 3} has a flow coefficient is less than 1.5 people / m · s. 2 floor midway _{H 2C} corridor _{C 2} is the flow coefficient is less than 1.5 people / m · s. Further, the flow coefficient H _{32 in} the middle of the stairs S ₃₂ from the third floor to the second floor is smaller than 1.33 person / m · s in the stairs portion. The midway H ₂₁ of the stairs S ₂₁ from the second floor to the first floor has a flow coefficient of approximately 1.33 person / m · s at the stairs. As shown in FIG. 38 (b), 103.50 s after the start of evacuation, all pedestrians leave and escape from all the rooms R _3A , R _3B , R _3C , R _2A , R _2B , R _{2C on} each floor. It is complete. The third floor of the pedestrian all from the corridor C ₃ is to exit, is a three-floor evacuation is complete. The exit E _2C of the second floor corridor C ₂ has a flow coefficient slightly smaller than 1.5 person / m · s due to the influence of pedestrians evacuating from the third and second floors. 2 floor midway _{H 2C} corridor _{C 2} is the flow coefficient is less than 1.5 people / m · s. Further, the midway H ₃₂ of the stairs S ₃₂ from the third floor to the second floor has a flow coefficient smaller than 1.33 person / m · s in the stairs. The midway H ₂₁ of the stairs S ₂₁ from the second floor to the first floor has a flow coefficient of approximately 1.33 person / m · s at the stairs.

As shown in FIG. 39 (a), and after 149.50s the evacuation started, and exit most of the pedestrian from the second floor of the corridor _{C 3,} the second floor of the flow coefficient at the exit _{E 2C} corridor _{C 2} is Less than 1.5 people / m · s. The midway H ₂₁ of the stairs S ₂₁ from the second floor to the first floor has a flow coefficient of approximately 1.33 person / m · s at the stairs. As shown in FIG. 39 (b), and after 184.46s the evacuation started, and exit pedestrians all from the exit _{E 1} stair _{S 21,} a Central evacuation completed.

  According to the evacuation behavior simulation system 1, when the density of the crowd flow state is reached, the behavior of the refugee evacuating from the building using the cellular automaton based on the local neighborhood rule that becomes the flow coefficient in the crowd flow state is illustrated. By simulating, the speed, moving direction (destination cell), etc. in each discrete time step can be reproduced with high accuracy for each pedestrian, and evacuation behavior in a free walking state and a crowd flow state can be simulated with high accuracy. In particular, according to the evacuation behavior simulation system 1, the real time of the discrete time step, the cell size, the setting of the personal space, the basic stop probability of the movement probability used for the action judgment based on the local neighborhood rule, the density coefficient, the movement direction probability By adjusting the stairs coefficient, the flow coefficient in the crowd flow state (especially 1.5 person / m · s for horizontal walking and 1.33 person / m · s for walking stairs) is accurately reproduced. it can.

  According to the evacuation behavior simulation system 1, the evacuation behavior in the free walking state and the crowd flow state can be simulated with high accuracy, so that the evacuation behavior in the passage, stairs, and living room can be reproduced with high accuracy, and from each room. Evacuation behavior can be reproduced with high accuracy even at junctions such as exit passages to corridors and exit passages from corridors on each floor to stairs. Moreover, according to the evacuation behavior simulation system 1, the evacuation behavior in the free walking state and the crowd flow state can be simulated with high accuracy. Can do.

  As mentioned above, although embodiment which concerns on this invention was described, this invention is implemented in various forms, without being limited to the said embodiment.

  For example, in the present embodiment, the case where the present invention is applied to the simulation of evacuation behavior in a living room, an arbitrary floor of the first floor, and a building of the third floor is shown, but various buildings such as a one-story building and a building having an underground floor are shown. Applicable to. It can also be applied when there are obstacles in the room. It can also be applied to buildings with spaces other than living rooms, stairs, and corridors. It can also be applied to outdoor spaces outside buildings such as underpasses, pedestrian decks, pedestrian bridges, and sidewalks (roads).

  In this embodiment, evacuee characteristics, evacuation space, evacuation behavior model, etc. are set based on the evacuation behavior prediction calculation method (the theory of evacuation prediction model), etc., in order to obtain a simulation approved by the Minister of Land, Infrastructure, Transport and Tourism. If it is not necessary to obtain authorization, the evacuee characteristics, the evacuation space, the evacuation behavior model, etc. may be appropriately set according to the required simulation.

  Also, in this embodiment, in order to make the simulation approved by the Minister of Land, Infrastructure, Transport and Tourism, the flow coefficient of the crowd flow state is 1.5 (1.33 for stairs) per person / m · s based on the Evacuation Safety Verification Act, etc. Although applied, if it is not necessary to obtain certification, another value may be set as the flow coefficient of the crowd flow state.

  In this embodiment, the actual time of discrete time step, cell size, setting of personal space, basic stop probability of movement probability used for action judgment by local neighborhood rule, density coefficient, moving direction probability, step coefficient Although an example of specific numerical values has been shown, in order to set the flow coefficient to 1.5 (1.33 for the staircase) / m · s, each numerical value may be set as appropriate. By setting each of these values to an optimum value, the flow coefficient can be made closer to 1.5 persons (1.33 for the staircase) / m · sec.

  DESCRIPTION OF SYMBOLS 1 ... Evacuation action simulation system, 2 ... Analysis space data reading means, 3 ... Simulation condition input means, 4 ... Database, 5 ... Simulation result output means, 6 ... Control means, 7 ... Simulation execution means.

Claims (4)

An evacuation behavior simulation system that simulates the behavior of a refugee evacuating from a building using a cellular automaton or evacuating in an outdoor space,
Divide the space inside or outside the building into cells, place all the evacuees to be simulated in each cell, and if the density of evacuees is the density of the crowd flow state at each discrete time step, An evacuation behavior simulation system characterized in that the next destination cell of all evacuees is determined based on a local neighborhood rule that becomes a flow coefficient.   2. The flow coefficient in the crowd flow state is 1.5 person / m · s for walking in a horizontal portion and 1.33 person / m · s for walking in a staircase portion. Evacuation behavior simulation system. An evacuation behavior simulation method for simulating the behavior of a refugee evacuating from a building or evacuating in an outdoor space using a cellular automaton,
Divide the space inside or outside the building into cells, place all the evacuees to be simulated in each cell, and if the density of evacuees is the density of the crowd flow state at each discrete time step, An evacuation behavior simulation method characterized in that the next destination cell of all evacuees is determined based on a local neighborhood rule that becomes a flow coefficient.   4. The flow coefficient in the crowd flow state is 1.5 person / m · s for walking in a horizontal portion and 1.33 person / m · s for walking in a staircase portion. 5. Evacuation behavior simulation method.