Sunday, February 19, 2017
Exhibit Hall (Hynes Convention Center)
Masashi Miura, Tottori University, Tottori, Japan
Agent-Based Modeling (ABM) and Multi-Agent Simulation (MAS) are effective methods for modeling complex social phenomena as well as understanding and predicting such phenomena through simulation. Using ABM, phenomena are described as collections of “agents” that act autonomously, make decisions, and learn according to the given rules. MAS is a method of reproducing the behavior of multiple interacting agents described using ABM on a computer, including investigation of system behavior under various conditions. This approach has been successfully applied in social sciences since the spread of computers in the 1990s, and numerous simulation models have been constructed. However, their actual social applications have thus far been limited to certain fields such as transportation, distribution, and evacuation, although their extension to a wider range of fields is expected in future. Furthermore, as it promises to be capable of responding to more finely segmented requirements, a method to efficiently model social phenomena for specific regions, organizations, and environments is required. In this study, we propose an interactive modeling method called Game-Based Situation Prototyping (GBSP), which uses the process of creating a game to perform ABM through collaboration with specialists and non-specialists in different fields. Both games and agent-models are characterized by rules, decision making, and interaction. So ABM and game design are said to be essentially similar actions. Furthermore, with games, the method of processing the overall process and specific situations are described in an easy-to-understand manner with rules. This description is an algorithm for programming MAS. Even people who do not have specialized ABM or MAS knowledge can perform social modeling through interactions using the process of creating a game as an intermediary. In GBSP, the paper prototyping method in game design is used to visualize situations and scenes and the social phenomenon is modeled as a game through trial and error. Thus far, in this study, we have practiced several forms of social modeling using GBSP and arranged the basic procedure and necessary tools. In this poster presentation, we primarily introduce an image of GBSP conducted by students in a classroom and its results and effects. This class has practical sessions for constructing ABM and MAS of social phenomena based on themes set by the students themselves,. Even students without any prior knowledge of modeling or programming could perform social modeling and simulation using GBSP.