一种基于粒子群算法和育种算法的混合算法  

A hybrid algorithm based on PSO and breeding algorithm

作　　者：张楠[1] 邢志栋[1] 董建民[1] 王辛[1] 

机构地区：[1]西北大学数学系,陕西西安710069

出　　处：《西北大学学报（自然科学版）》2008年第1期19-21,26,共4页Journal of Northwest University（Natural Science Edition）

基　　金：陕西省教育厅专项基金(05JK303)

摘　　要：目的求解无约束优化问题。方法结合粒子群算法收敛速度快而育种算法不易陷入局部最优的特点,设计了一种新的算法。结果数值试验结果表明算法对于多峰值函数有很好的优化效果。结论当群体最优信息停滞时,新算法能够使粒子群算法中的粒子跳出局部最优解,最终求得全局最优解。Aim To solve unconstrained optimization problems. Methods On the basis of the properties of particle swarm optimization （PSO）, which always converges quickly, and the properties of breeding algorithm （BA）, which is not easy to immerse the region containing local solution, a new hybrid algorithm is proposed. Results The results of the experiment show that the hybrid algorithm has good effects on the functions which have many local solutions. Conclusion The advantage of this hybrid algorithm is that, when maximum solution comes to a standstill gradually, the particles of PSO are able to get the global optimal solution.

关 键 词：粒子群算法 育种算法 全局最优解 

分 类 号：O224[理学—运筹学与控制论]