Modelling Agents Behaviour in Automated Negotiation

This paper presents a learning mechanism that applies nonlinear regression analysis to model a negotiation agents behaviour based only on the opponent's previous offers. The behaviour of negotiation agents in this study is determined by their tactics in the form of decision functions. Heuristics based on estimates of an agents tactics are drawn from a series of experiments. By applying the nonlinear regression and the obtained heuristic knowledge, an agent can improve their overall performance by predicting the opponents deadline and reservation value, terminating pointless negotiation, and avoiding negotiation breakdown. The findings of this study show that this approach can be used to obtain better deals than previously proposed tactics. The learning mechanism can be used online, without any prior knowledge about the other agents and is therefore, very useful in open systems where agents have little or no information about each other.