Compromising Strategy Using Weighted Counting in Multi-times Negotiations

Abstract

Bilateral multi-issue closed negotiation is an important class of real-life negotiations. Usually, negotiation problems have constraints, such as a complex and unknown opponent's utility in real time or time discounting. In the class of negotiation with constraints, effective automated negotiation agents can estimate their opponent's model depending on the proposals of their opponents and the negotiation scenarios. Recently, the attention of this study has focused on interleaving learning with negotiation strategies from past negotiation sessions. By analyzing such previous sessions, agents can estimate their opponent's utility function based on exchanging bids. In this paper, we propose an automated agent that estimates its opponent's strategies based on past negotiation sessions. Our agent decides the estimated values of its opponent using effective weighted functions based on the negotiation time. By using the estimated values of each issue, our agent can calculate its opponent's utility. In addition, we employ the estimated method proposed in this paper to the compromise strategy, which is the agent of the basic strategy of our proposed agent. In our experiments, we compared seven different weighted functions to determine the most effective one. In addition, we demonstrated that our proposed agent has better outcomes and a greater search technique for the Pareto frontier than existing ANAC2013 agents. We also compared our proposed agent and the basic compromising strategy.