Social Simulations Using Multi-agent Systems
摘要
In recent years, “multi-agent simulation” as social simulation has been the focus of much attention, including countermeasures against COVID-19. Multi-agent simulation is a model that can represent complex social phenomena, predict future events, and investigate their causes. Therefore, the realization of various multi-agent simulations has potential for various applications. However, there are also several challenges in achieving realistic simulations. In particular, it is essential to understand the relationship between agents, such as humans or robots, and the environment in which they act. As a study focusing on the environment of simulation targets, this chapter presents research on map quantification and evaluation of a multi-agent system using a disaster relief simulation, “RoboCupRescue Simulation.” In addition, this chapter introduces several studies analyzing the relationships between traffic accidents and the environment in which they occur for realizing multi-agent simulations of traffic accidents.