A New Model for International Relationship based on Multi-Agent

摘要

Analysis and simulate international relationship by using high-tech cluster of information technology is important and meaningful to the modern research of strategic work and the strategic decision-making. This paper use Multi-Agents system to model the international relationship with macrocosmic level and microcosmic level. For the microcosmic level shows activities and the rules of correlative activities of countries, propose a reinforcement learning algorithm to improve cooperative efficiency among agents, which based on common goal and joint rewards to insure agents to learn cooperative behavior. A prototype and an experiment were given.