Learning single-issue negotiation strategies using hierarchical clustering method

This research proposes an off-line learning method targeted for systematically constructing single-issue negotiation strategies in electronic commerce. Our research is motivated by the following fact: evidence from both theoretical analysis and observations of human interaction shows that if decision makers have a prior knowledge on the behaviors of opponents, the overall payoffs would increase. Given past negotiation data set, a competitive learning and a variant of hierarchical clustering model are applied to extract the negotiation strategies. A negotiation strategy is a chain of the pairs consisting of (buyer's offer, seller's counteroffer). An agent-based simulation convinced us that the proposed method is more effective than human negotiation in terms of the ratio of negotiation agreement and resulting payoffs.