【摘要】 提出一种基于排异竞争机制的粒子群优化算法。算法取消传统PSO算法中的全局最优值"gbest",通过设定竞争区域,使得当前种群中所有粒子和上一代种群中的精英粒子,一同参与竞争。并采取适应值竞争策略、适应度选择策略和粒子间的排异策略,来保证种群的多样性,避免了算法初期陷入局部极值的可能;并通过对排异策略的动态调整,提高了算法后期的收敛速度和精度。通过对几类典型函数的仿真测试表明,算法具有较好的全局搜索能力和收敛速度。更多还原

【Abstract】 In order to enrich the population of particle swarm optimization algorithm diversity,a competitive mechanism was proposed based on the rejection of the particle swarm optimization algorithm.Modified algorithm cancelled the global optimal value of "gbest" in traditional PSO algorithm,and made the current population of all the particles and the previous generation of elite population particles to compete together by setting the competition area.Through adopting competitive strategy of fitness and selection strategy of fitness,it ensures the population diversity and avoids the possibility of initial algorithm into local minima.At the same time,through the competition between particle radiuses of the dynamic adjustment of rejection,it makes the late algorithm convergence speed and accuracy improved.By comparing the simulation tests of several types of typical function,it shows that the improved algorithm has good global searching ability and convergence speed,much better than the traditional algorithms.更多还原