IQSon: A Context-aware Negotiator Agent with Enhanced Utility and Decision Making Speed

In an automated negotiation, software agents try to gain the best possible utility on behalf of their owners. The main challenge in these negotiations is not knowing the opponents’ preferences. Although methods such as opponent modeling can be used to solve this problem, modeling generally has a computational overhead and requires time. In areas such as perishable goods, energy markets and fashion industry, we need to negotiate quickly which hinders the use of modeling. In this paper, an agent named IQSon is presented. Using a multi-policy strategy, this agent negotiates in a multilateral environment without creating a model but by considering a short history of opponents’ offers. Experimental results show that the agent gains good utility while negotiating with state of the art agents. Furthermore results show the effectiveness of the bidding policies. IQSon also participated in the International Automated Negotiating Agents Competition (ANAC2018) and was eventually ranked 4th out of 22 international participants in the individual utility rankings and 5th in the social welfare rankings.