约束优化问题的修正选择粒子群优化算法

Revised selection particle swarm optimization algorithm for solving constrained optimization problems

提出一种求解约束优化问题的修正选择粒子群优化算法(RSPSO).在这个算法中,利用动态多阶段罚函数方法处理约束,并加入一种违反约束的修正选择策略,采用线性递减违反约束容忍度来引导粒子,即利用修正的可行基规则来更新个体极值和全局极值,指引粒子迅速飞向可行域;考虑到粒子群中每个粒子周围的局部信息对它未来飞行的影响,改进了基本粒子群优化的速度方程.数值结果表明,所提出的算法求解约束最优化问题具有较高的计算精度、较好的稳定性和较强的全局寻优能力.

A revised selection particle swarm optimization（RSPSO） algorithm was presented for solving constrained optimization problems.In this algorithm a method of dynamic multi-stage penalty function was used to process the constraint,a revised selection strategy was supplemented for constraint violation,and use a linear decreasing the tolerance degree of constraint violation to guide the individual particle,i.e.use modified feasibility-based rule to update the individual extremum and global extremum to guide the individual particle fly to the feasible region as soon as possible.Taking into consideration the influence of local information around each particle＇s impact on its future fly,the optimized velocity equation of basic particle swarm was modified.Numerical simulation showed that the algorithm presented exhibited a higher computation accuracy,better stability,stronger ability for global optimization.