Ms. Liu Shan won the competition in PRIANAC
Category：Award｜Publishing : November 5, 2018
Liu Shan, Department of Computer Science
Winner of Pacific Rim International Automated Negotiation Agents Competition (PRIANAC) (Social Welfare Category)
PRIANAC is a competition of standard automated negotiation (bilateral negotiation with alternative proposals). Ms. Liu submitted an agent called Agent33, which won the competition in the social welfare category. By dividing the bids offered by the opponent agents into issue units and analyzing them with a heuristic method, Agent33 can reconstruct bids near the Nash bargaining solution. By choosing bids that have high joint utilities from those reconstructed bids and offering them to the opponent agents, Agent33 got the highest joint utility in the PRIANAC.
Award winner's comments:
I feel very honor to get this award. I am deeply grateful for the guidance from Prof. Ito, Assoc. Prof Moustafa, and the advice from some members in the laboratory. Agent33 focused on bid searching strategy that can get the higher joint utility. However, it performed weakly on individual utility. In order to improve its individual utility, two directions will be considered. First, use machine learning to classify the bids list offered by the opponent agents in order to get better bids that are nearer to the Nash Bargaining solution. Second, choose different bid acceptance strategies for different opponent agents.