Hiroyuki Iizuka, Hokkaido University, Hokkaido, Japan
Masahito Yamamoto, Hokkaido University, Hokkaido, Japan
Keiji Suzuki, Hokkaido University, Hokkaido, Japan
Azuma Ohuchi, Hokkaido University, Hokkaido, Japan

Abstract: The dynamics of positions of political parties have been extensively studied by both numerical and agent-based approaches. We focus on the dynamics of formation and division of parties, particularly on how game rules to determine a winning party influences the agents’ decision-making and process to reaching a consensus, and we perform an agent-based simulation using a game model called Spatial Voting Game. In this paper, we briefly describe the game as an environment of our agent-based simulation and learning architecture of agents for adaptive players in a multiagent system. Then we present simulation results showing that the players’ strategies acquired by evolutionary operator and the process to form parties in a negotiation are different according to the game rules.

Keywords: spatial voting model, agent-based simulation, coalition formation, casting vote