10:00 AM - 10:15 AM
[HDS09-05] Consideration of Evacuation Rate Change by Presenting Evacuation Destination Information
-Application of Multi-Agent Simulation in the Central Area of Toyama City-
Keywords:Multi-Agent Simulation, Evacuation Rate, Evacuation Destination Information, Toyama city
Toyama Prefecture in Japan has seen an increase in tourists since the opening of the Hokuriku Shinkansen in 2015. However, the reality is that the number of tourists has decreased due to the Corona disaster. As a follow-up, the Noto Peninsula Earthquake occurred in 2024, and the maximum intensity of 5 upper was observed in Toyama City for the first time in its history. Toyama City also suffered damage to houses and liquefaction, and there are fears that the number of tourists may decrease in the future.
In addition, inbound demand is increasing in Toyama Prefecture, and the number of foreign tourists visiting the prefecture is on the rise. However, foreign tourists may not be able to take effective and efficient evacuation actions because they cannot know where to evacuate to in the event of a disaster in an unknown place. Considering that posting evacuation information will lead to reliable evacuation, it is necessary to mention how the information needs to be posted.
Therefore, in this study, we decided to pursue the possibility of changing the evacuation rate by posting information in the central area of Toyama City by applying a multi-agent simulation. For the multi-agent simulation, S4, a commercial software, was used. OpenStreetMap was used for the road network. This was done in order to be able to apply the same problem to any region in the country when it occurs.
In conducting the multi-agent simulation, the agents were positioned as foreign guests in Toyama, and agents were generated according to the actual number of guests in recent years. Four evacuation sites were selected: Inari Park, Joshi Park, Kencho-mae Park, and Ushijima Park. Agents were generated from random locations within Toyama City. However, since there is an extremely high possibility that they will not reach their destinations if they are allowed to move aimlessly, it was assumed that they are aware of evacuation to the extent that they are away from the sea. To reflect this in the simulation, the central area of Toyama City was divided into 32 blocks of 4 rows and 8 columns, and the probability of movement from each block to the adjacent block was set. The probability of moving to the mountain side (inland side) was set to be high. In addition, the condition was set that one information bulletin board be installed per block. Under these conditions, in addition to a model with no bulletin board, a model was set up with up to 8 blocks installed, for a total of 9 models. Ten cases of simulation were run for each model, and the average of the evacuation achievement rate after a certain time was used as the result of each model.
Comparing the simulation results, it was confirmed that the evacuation rate also improves with an increase in the number of information boards. However, the increase in the evacuation rate with an increase of one information board was not uniform, and the installation of three information boards was confirmed to be the most efficient. Conversely, the installation of five or six information boards was inefficient.
This study includes some conditions that cannot be said to be sufficient, such as the conditions for setting up information bulletin boards in blocks and the random selection of installation sites. We believe that further refinement of these conditions will improve the reproducibility of events. In addition, we believe that it is desirable to conduct simulations after selecting actual installation sites through field surveys. We believe that this will enable us to find a more realistic solution.
In addition, inbound demand is increasing in Toyama Prefecture, and the number of foreign tourists visiting the prefecture is on the rise. However, foreign tourists may not be able to take effective and efficient evacuation actions because they cannot know where to evacuate to in the event of a disaster in an unknown place. Considering that posting evacuation information will lead to reliable evacuation, it is necessary to mention how the information needs to be posted.
Therefore, in this study, we decided to pursue the possibility of changing the evacuation rate by posting information in the central area of Toyama City by applying a multi-agent simulation. For the multi-agent simulation, S4, a commercial software, was used. OpenStreetMap was used for the road network. This was done in order to be able to apply the same problem to any region in the country when it occurs.
In conducting the multi-agent simulation, the agents were positioned as foreign guests in Toyama, and agents were generated according to the actual number of guests in recent years. Four evacuation sites were selected: Inari Park, Joshi Park, Kencho-mae Park, and Ushijima Park. Agents were generated from random locations within Toyama City. However, since there is an extremely high possibility that they will not reach their destinations if they are allowed to move aimlessly, it was assumed that they are aware of evacuation to the extent that they are away from the sea. To reflect this in the simulation, the central area of Toyama City was divided into 32 blocks of 4 rows and 8 columns, and the probability of movement from each block to the adjacent block was set. The probability of moving to the mountain side (inland side) was set to be high. In addition, the condition was set that one information bulletin board be installed per block. Under these conditions, in addition to a model with no bulletin board, a model was set up with up to 8 blocks installed, for a total of 9 models. Ten cases of simulation were run for each model, and the average of the evacuation achievement rate after a certain time was used as the result of each model.
Comparing the simulation results, it was confirmed that the evacuation rate also improves with an increase in the number of information boards. However, the increase in the evacuation rate with an increase of one information board was not uniform, and the installation of three information boards was confirmed to be the most efficient. Conversely, the installation of five or six information boards was inefficient.
This study includes some conditions that cannot be said to be sufficient, such as the conditions for setting up information bulletin boards in blocks and the random selection of installation sites. We believe that further refinement of these conditions will improve the reproducibility of events. In addition, we believe that it is desirable to conduct simulations after selecting actual installation sites through field surveys. We believe that this will enable us to find a more realistic solution.