基于合作博弈的网格资源分配

针对网格资源分配中的竞争问题,提出了一种合作博弈的方法,通过建立用户的评估函数,使得竞争同一资源的多个用户能够结成联盟来增加整体效用,其中用户分得的资源份额取决于出价水平.着重研究了联盟中用户的评估函数并证明了其满足非联盟评估函数的性质,因此在资源竞争中联盟成为有效的参与者.分析表明,该方法不仅可产生惟一的纳什均衡,而且能够为联盟用户带来更多的收益,从而实现帕累托改进.

Grid resources allocation based on cooperation game

A cooperation game-based method is proposed to address the problem of competition in grid resources allocation. Through establishing valuation function of grid user, a coalition with muhiple users is formed to increase the utility, where the resources partition gained by user depend on the bidding level. The valuation function is studied in emphasis, and its characteristics are proved to be consistent with those of valuation function out of the coalition. Hence, coalition is an efficient player during the resources allocation competition. The analyses show that the proposed method can generate a unique Nash equilibrium, and a coalition of users performs better than a single user, and then can realize the Pareto improvement.