总结性自适应变异的粒子群算法

Self-Conclusion and Self-Adaptive Variation Particle Swarm Optimization

针对粒子群优化(particle swarm optimization,PSO)算法在迭代期间易陷入局部最优及寻优精度不高的缺点,提出一种总结性自适应变异的粒子群算法SCVPSO(self-conclusion and self-adaptive variation particle swarm optimization)。采用非线性转折上升再递减惯性权重动态更新每个粒子的位置,有效避免早熟;对筛选的局部粒子作反向搜索处理,提高种群寻优效率;引入新的参数scr(self-conclusion rate)以总结各个粒子近期求解情况,并通过概率单向变异引导粒子指向全局最优,增加粒子多样性。借助15个测试函数与其他变种粒子群优化算法对比,结果显示,改进之后的算法在求解性能上明显优于其他算法,验证了该策略的有效性。

Aiming at the shortcomings of particle swarm optimization(PSO)algorithm, which is easy to fall into local optimum and low precision during iteration, this paper proposes a self-conclusion and self-adaptive variation particle swarm optimization(SCVPSO). Firstly, the position of each particle is dynamically updated by nonlinear turning up and then decreasing inertia weight to avoid premature. Secondly, the local particles are searched backward to improve the efficiency of population optimization. Finally, a new parameter scr(self-conclusion rate)is introduced to summarize the recent solution situation of each particle, and the probability directed variation is used to guide the particles to the global optimum to increase the diversity of particles. With the help of 15 test functions, compared with other variant particle swarm optimization algorithm, the results show that the improved algorithm is significantly better than other algorithms in solving performance, which verifies the effectiveness of the strategy.