基于加速遗传模拟退火算法的协商策略研究

ON NEGOTIATION POLICY BASED ON AGASA ALGORITHM

随着电子商务的不断发展,如何快速有效地进行自动协商是研究所面临的一个重要问题。根据Bazzar协商模型,提出了一种加速遗传模拟退火算法(AGASA),该算法将遗传算法和模拟退火算法结合,并且加入压缩搜索范围的算法加速机制,同时采用特殊的实数编码方式令算法能更加稳定地收敛。仿真试验表明,算法能快速稳定地解决协商模型所描述的协商问题。

Along with the growing development of the E-commerce,there is an important subject for our Institute that to solve the problem of automatic negotiation rapidly and efficiently. In the paper,according to the negotiation model of Bazzar,we propose an accelerating genetic and simulated annealing（AGASA） algorithm. It is based on the union of GA and SA and combines algorithm accelerating mechanism which compresses the searching area,at the same time the special real coding mode is adopted to make the algorithm converges more stably. Experiments of simulation indicate that this algorithm can solve the negotiation problem described by the negotiation model quickly and stably.