在P2P邻接选择模型的基础上对针锋相对策略提出一种适应值函数选取的方法,给出使用离散粒子群算法的求解方法.定义了约束算子使即将越界的粒子随机跳回某一个边界值,既解决了约束问题,又利用了候选解之间的联系.实验结果表明,对于大规模的邻接选择问题,本方法在收敛速度和结果方面均好于遗传算法.
Based on P2P neighbor selection model,a new evaluation function is proposed for tit-for-tat strategy,and a discrete PSO method is applied to solve it.Aiming at solving constrained problems and utilizing the relations between backup particles,the constrained operator is designed to bring the particles that are crossing the restriction boundary back to border.The results show that compared with GA,the proposed DPSO usually has better outcome in shorter iterations for large scale problems.