基于免疫进化的粒子群混洗蛙跳算法.docVIP

基于免疫进化的粒子群混洗蛙跳算法 　　?フ? 要:为了避免混洗蛙跳算法易于出现不成熟收敛，提高求解质量，提出了基于免疫进化的粒子群混洗蛙跳算法。该算法将粒子群算法中粒子追踪全局极值的思想融入混洗蛙跳算法中，对族群内的最差个体同时跟踪族群内和全局两个最优个体的信息，进行深度搜索；并引入免疫进化算法对群体中的最优个体进行免疫进化迭代计算，以达到充分利用最优个体的信息的目的。该算法不仅避免了陷入局部极值的局限，以更高的精度逼近全局最优解，而且能加速收敛。对多个典型测试函数的计算表明：基于免疫进化的粒子群混洗蛙跳算法比传统的混洗蛙跳算法具有更好的寻优能力、稳定效果和更快的收敛速度。 　　?ス丶?词:混洗蛙跳算法；免疫进化算法；粒子群算法；函数测试 　　 　　?ブ型挤掷嗪?: TP18 文献标志码:A 　　 　　Abstract: A new shuffled frog leaping algorithm based on immune evolutionary particle swarm optimization was proposed in order to avoid premature convergence and to improve the precision of solution by using basic Shuffled Frog Leaping Algorithm (SFLA). The proposed algorithm integrated the global searching idea in the Particle Swarm Optimization (PSO) into SFLA, to pursue the information of two optimal solutions in the sub??swarm and the whole??swarm simultaneously, so as to search thoroughly near by the space gap of the worst solution, and also integrated the immune evolutionary algorithm into SFLA making immune evolutionary iterative computation to the optimal solution in the whole??swarm, so as to use the information of optimal solution fully. This algorithm can not only get free from trapping into local optimum and be close to the global optimal solution with higher precision, but also speeds up the convergence. Calculation results show that the Immune Evolutionary Particle Swarm Optimization??Shuffled Frog Leaping Algorithm (IEPSO??SFLA) has better optimal searching ability and stability as well as faster convergence than those of basic SFLA. 　　??Key words: Shuffled Frog Leaping Algorithm (SFLA); Immune Evolutionary Algorithm (IEA); Particle Swarm Optimization (PSO); function test 　　?? 　　 　　 传统的优化算法，比如最小二乘法、最速下降法、线性规划、单纯形法和混合型算法等适用于较简单的问题求解，但对于具有多约束的非线性、多极值等复杂问题，不仅求解过程复杂、收敛速度慢，而且对一些问题很难求得全局最优解。近20年来，人们从自然界中一些生物比如蚂蚁、鸟类、鱼群和蜜蜂等群体的社会行为中获得启迪和思路，提出了多种用于求解复杂问题的仿生智能优化算法????［1-9］??。混洗蛙跳算法(Shuffled Frog Leaping Algorithm, SFLA)就是一种基于群体智能的具有全局协同的后启发式计算技术。由于它具有原理简单、概念清晰、运算简便、易于实现和全局收敛的特点，因而一经提出，已在若干领域得到应用????［10-