减少速度更新频率的混沌粒子群算法

Computation method for chaos particle swarm optimization based on relaxation velocity update frequency

把速度更新策略和混沌优化相结合,提出了减少速度更新频率的混沌粒子群算法.该算法根据群体适应值的方差进行早熟收敛判断,从而使算法摆脱后期易于陷入局部最优点的束缚,同时又保持前期优秀的搜索速度的特性.通过几个基准函数测试,结果表明,新算法的性能较基本粒子群优化算法有明显的改善.

Combining relaxation velocity update frequency and chaos optimization, a chaos particle swarm optimization was proposed based on relaxation velocity update frequency （CRVUPSO）. The algorithm adapts the swarm fitness＇ variance to implement the prematurity judgment, thus enabling algorithm break away from the shackles of local optimization. Simultaneously, the previous velocity characteristics was saved. Through tests several benchmark function,it is indicated that the performance of the new algorithm has significant improvement than the basic PSO, the search speed is quick, the computational accuracy is high.