中文摘要：

为了使粒子群优化（PSO）适于求解更多类问题,提出一种由动力空间和制导空间共同进化的改进粒子群优化算法—具有双重进化空间的扩展粒子群优化算法（简记EPSO）.在EPSO中,在演化转换映射的作用下,首先将动力空间中对粒子辅助位置的进化转换为制导空间中对主导位置的进化,然后基于对主导位置的择优选择操作实现算法的进化过程.EPSO克服了PSO仅适于求解连续域最优化问题的缺陷,也非常适于求解离散组合优化问题.对于随机3-SAT问题、背包问题和TSP问题,通过与PSO、ACO和GA等算法的计算对比表明：EPSO是一种继承了PSO优点的高效、扩展演化算法.

英文摘要：

Advanced an Expand Particle Swarm Optimization with Double Evolution Spaces （EPSO） from co-evolution of drive space and control-guide space, which is in order to make PSO apply to solve various optimization problems. In EPSO, evolution process is proceeding through evolution transformation mapping under coordination of drive space and control-guide space, the particle auxiliary position transforms into leading position, and accomplish evolutionary search by selection operator. EPSO has overcome disadvantages of PSO that is only applied to solve optimization problem over continuous field, and fit into combi- nation optimization over discrete field. For 3-SAT problem, Knapsack problem, TSP problem, the performance of EPSO is compared with PSO, ACO and GA etc. The results show that EPSO is a high efficient evolutionary algorithm that has inherited advantages of PSO.