一种消解协商僵局的多目标粒子群优化算法

解决协商僵局问题是协商优化中的重要研究课题.利用协商议题之间的相关性,提出了一种用于消解双边多议题协商僵局的多目标粒子群优化算法(MOPSO).MOPSO首先动态放宽僵局议题的保留值,然后将僵局议题相关的多个议题的保留值缩紧问题转化为一个多目标优化问题,通过粒子群搜索到Pareto最优解集,从而并行优化了这些相关议题的保留值,最后在不降低协商者整体利益条件下进行协商议题保留值向量等效置换.实验验证了MOPSO是有效的,其僵局解决能力明显比现有的其他方法强.

A Multi-Objective Particle Swarm Optimization Algorithm for Solving Negotiation Deadlock

It is one of the important study tasks for negotiation optimization to solve negotiation deadlocks.In order to get rid of such deadlocks in the time-limited bilateral and multi-issue autonomous negotiation,a multi-objective particle swarm optimization algorithm,called MOPSO,is put forward in this paper.MOPSO makes full use of the relationship among issues and first relaxes the reserved value of the issue dynamically which triggers the negotiation deadlock.Then the algorithm translates the problem of tightening the reserved values of the issues relevant to the deadlock issue into a multi-objective optimization one and turns up a Pareto-optimal set by a particle swarm.In this way,these reserved values are optimized in parallel and the algorithm lastly replaces the old reserved vector of the negotiation issues with a new one equivalently,which keeps the level of the integrated utility of the negotiant.The obtained results of experiments on E-commerce support the claim that MOPSO is valid and it is preferable to the existing method in solving the problem of the negotiation deadlocks.