一种基于差异演化变异的粒子群优化算法

为了保持粒子种群的多样性而避免发生“早熟”的问题,提出一种基于差异演化变异的粒子群优化算法（PSO）,该方法通过粒子聚集性判断如果粒子群中的粒子过于聚集,则使用差异演化算法对PSO算法中各个粒子的自身历史最佳位置进行变异,以实现保持粒子群种群多样性的目的。对4种常用函数的优化问题进行测试并进行比较,结果表明：所改进的粒子群优化算法比标准粒子群优化算法更容易找到全局最优解,优化效率和优化性能明显提高。

Particle swarm optimization algorithm based on differential evolution mutation

In order to preserve the varieties of the swarm and avoid to be in premature convergence,a Particle Swarm Optimization（PSO） algorithm based on the differential evolution mutation is proposed．The new algorithm uses the particle aggregation quality to judge that if the particles in the swarm are congregative so much,then apply the differential evolution to mutate the self prevenient best position of each particle,in order to realize the aim of preserving the varieties of the swarm.Then,this new PSO and the standard PSO are used to resolve four well—known and widely used test functions’ optimization problems．Results show that the new PSO has greater efficiency,better performance and more advantages than the standard PSO in many aspects.