Strategic Human Skill Development by Agent-based Participatory Approach: Improving Negotiation Skill by Interacting with Bargaining Agent

摘要

This paper explores a method of developing human skill through the interaction between human player and computer agent and investigates its effectiveness through the experiments on the bargaining games where human player negotiates with computer agents. In detail, human players negotiate with the following three kinds of agents: (a) the strong/weak attitude agent that makes the aggressive/defensive proposals in the advantageous/disadvantageous situation; (b) the fairness agent that makes the fair proposals; and (c) the human-like behavior agent that makes the mutual agreeable proposals as the iteration (games) increases. The analysis of the subject experiments has revealed the following implications: (1) the human player who negotiates with the strong/weak attitude agent obtains the large amount of the profit overall; (2) the human player who negotiates with the human-like behavior agent wins a lot of games; and (3) no relationship is found between the profit maximization and a win of the games.