[4N3-GS-5] Agents: applications

[4N3-GS-5-02] Visual Evacuation Signage Assignment Considering Risks and Multi-Objectives

〇Akira Tsurushima¹ (1. SECOM CO., LTD.)

Keywords:Evaciatn Decision Making, Herd Behavior, Multi-agent Simulation, Stochastic Multi-Objective Optimization, Average Value at Risk

Visual evacuation signage is often employed to guide evacuees efficiently to the proper route or exit in the buildings in emergency situations. The proper arrangement of signs is required to obtain the most out of this attempt and avoid adverse effects. However, several properties in the problem make it challenging to solve. For example, herd behavior among the crowd transforms the crowd behaviors into a complex system; solutions were evaluated through simulations. Evacuation often involves multiple conflicting objectives. Solutions involve risk or uncertainty owing to the stochastic nature of the problem, which is required to consider in the solution process. We formulated this problem as a stochastic multi-objective optimization problem. The number of evacuees selecting the correct exit and the total evacuation time are two crucial but often conflicting objectives. Average value at risk (AVaR) is adopted as a risk measure of solutions. We employed NSGA-II algorithm, a typical evolutionary multi-objective optimization algorithm, to solve the problem and have obtained solutions satisfying Pareto-efficiency for both expected values and AVaRs.