Abstract

In the open and dynamic market environment, the factors affecting the negotiation results are changing rapidly. A single predetermined negotiation strategy may fail to make a response to them. In this work, we explore an adaptive negotiation process using the Market-Driven Agents (MDAs) to adopt adaptive conceding strategies responding to changes in the environment. Meanwhile, a simulated annealing (SA) algorithm is given to distribute concession into each negotiating issue, which applies the similarity criteria to generate the counter-offer most similar to counterpart's bid. In order to improve the performance of similarity criteria policy, we try to learn the negotiation issue's weight of the counterparts using the Bayesian Learning method. Our experiments show that this model has a better adaptability to the changing environment. And our model can get a better result and create a win-win situation especially when the preferences of the two sides are quite different.

Paper Details