基于改进粒子群优化算法的网络化仿真任务共同体服务选择

Service Selection of Network Simulation Task Community Based on Improved Particle Swarm Optimization Algorithm

作为网络化仿真中新的应用需求,如何动态地把散布在网络上各种服务整合起来以形成新的、满足不同用户需求的仿真任务共同体(STC)成为了当前研究热点。提出了一种基于粒子群优化(PSO)算法的仿真服务选择方法,针对传统PSO易陷入局部最优和收敛速度慢等不足,设计了一种惯性权重动态变化策略和一种可选的变异操作方法。该算法不仅能提高服务选择收敛速度,还能避免算法陷入局部最优。通过实验,采用典型函数进行了测试,并详细介绍了算法在STC服务选择上的实际运用,说明了算法的可行性和有效性。

As one of new application requirements in network simulation, to dynamically integrate the distributed various services in network to form a new simulation task community （STC） which meets the needs of different users has become current research focus. This paper presents a simulation service selec- tion method based on the particle swarm optimization. The traditional particle swarm algorithm has some shortcomings that may easily fall into local optima and have slow convergence rate. We design a dynamic inertia weight strategy and a selectable method of mutation. The algorithm can improve the convergence speed not only, but also avoid falling into local optimum. Finally, some typical functions are chosen to test the algorithm. And the results show that the algorithm can select services feasibly and effectively for STC.