[2I3-GS-5b] エージェント：マルチエージェントシミュレーション

[2I3-GS-5b-02] Fidelity Pedestrian Simulation Optimization for Large-Scale Crowd Event

〇Yusaku Kato¹, Shusuke Shigenaka¹, Ryo Nishida², Masaki Onishi³ (1. University of Tsukuba, 2. Tohoku University, 3. National Institute of Advanced Industrial Science and Technology(AIST))

Keywords:Multi-Agent Simulation, Pedestrian Simulation, Fidelity Optimization, Social Force Model

Pedestrian simulation is widely used for evacuation guidance in the event of natural disasters. This research fields are being actively conducted to understand congestion in advance and to alleviate congestion. In pedestrian simulation, the computation time increases with the number of pedestrians because the movement of each pedestrian is calculated to represent the movement of the crowd. Therefore, there is a limit in terms of computational resources when exhaustively simulating various experimental settings for a large number of pedestrians. The problem is to reduce the computation time of the simulation.we propose a fidelity optimization that can reduce the computation time while maintaining the accuracy of the simulation. The number of pedestrians in the simulation is reduced by 4,000 agents from 40,000 in the real world, and the computation time is significantly reduced.