Atlas3: A Negotiating Agent Based on Expecting Lower Limit of Concession Function

Abstract

The sixth international Automated Negotiating Agents Competition (ANAC) was held in conjunction with the ninth international joint conference on Autonomous Agents and Multi-Agent Systems (AAMAS). We developed Atlas3 that is an automated negotiating agent for ANAC2015. In this paper, we explain about the searching methods and compromising strategy that is used by our agent. Our agent uses appropriate searching method based on relative utility for linear utility spaces. Moreover, our agent applies replacement method based on frequency of opponent’s bidding history. Our agent decides concession value according to the concession function presented by us in Mori and Ito (A compromising strategy based on expected utility of evolutionary stable strategy in bilateral closed bargaining problem, 2015, [2]). In Mori and Ito (A compromising strategy based on expected utility of evolutionary stable strategy in bilateral closed bargaining problem, 2015, [2]), we derived an estimated expected utility to estimate an appropriate lower limits of concession function. However, Mori and Ito (A compromising strategy based on expected utility of evolutionary stable strategy in bilateral closed bargaining problem, 2015, [2]) proposes a concession function for bilateral multi-issue closed bargaining games. Therefore, we extend the concession function for multi-lateral multi-issue closed bargaining games.