[3O3-GS-5] Agents: multi-agents

[3O3-GS-5-04] A Negotiation Order Adjusting Method based on the Important Issue in Multiagent Automated Negotiations.

〇Shun Okuhara¹, Takayuki Ito² (1. Faculty of Engineering Graduate School of Engineering, Mie university, 2. Graduate School of Informatics, Kyoto University)

Keywords:Automated Negotiations, Multi-Agent Simulation, Negotiation protocol

Automated negotiation research investigates new ways of reaching an agreement among self-interested agents. In multi-player negotiations, outcomes differ greatly depending on the order in which the negotiations are conducted between the agents. Most studies, however, do not consider the negotiation order and its impact on the final outcomes. Many negotiation protocols, such as Stacked Alternating Offers Protocol (SAOP), negotiations are conducted in a predefined order. This setting does not always lead to maximal social welfare. In this paper, we propose to address this limitation by adding a pre-negotiation phase. We compared the values of social welfare and the negotiation durations between the proposed method and conventional protocols such as SAOP. Results show that the protocol yields higher social welfare.