A106 人工社会における流行の発生と崩壊(マルチエージェント基礎理論,シミュレーションプラットホーム)

概要

論文の詳細を見る
This paper describes concentration phenomena of information could be happen in a peer-to-peer community using an agent-based computer simulation. The simulation consists of programmed agents. Agents communicate and trade with each other. There are no global information and any market. Agents trade their item with other agents. They accept not only item they want but also they consider as popular. Each agent has its popularity information. After they trade with another agent, they consider the item that the other wanted to be popular. And they share the information between the two agents. In this situation, agents can exchange their item and they consume item that they want. Furthermore most of agents consider one kind of item to be the most popular even though there is no global information such as mass media. This concentration of information corresponds to trend phenomena in the world. When the rule that the agents trust other agents that they have traded, the concentrations of popularity continues for a certain period. The frequency distribution shows the same results as the contact process's result. Then the behavior of the agents could be illustrated.
一般社団法人日本機械学会の論文
2001-11-14