CAJ | 학술논문

针对标准粒子群优化算法在优化多极值函数时容易陷入局部最优的缺点,分析了其进化原理以及过早收敛的原因,并在此基础上提出了分阶段进化的改进算法,即将进化过程分成多个阶段,不同的进化阶段应用不同的迭代进化公式,以提高种群的多样性,进而有效避免过早收敛。仿真实验结果表明,对于复杂的多极值函数优化问题,改进后的方法比标准粒子群优化算法具有更强的全局寻优性能。
침대표준입자군우화산법재우화다겁치함수시용역함입국부최우적결점,분석료기진화원리이급과조수렴적원인,병재차기출상제출료분계단진화적개진산법,즉장진화과정분성다개계단,불동적진화계단응용불동적질대진화공식,이제고충군적다양성,진이유효피면과조수렴。방진실험결과표명,대우복잡적다겁치함수우화문제,개진후적방법비표준입자군우화산법구유경강적전국심우성능。
In view of the shortcomings of the standard particle swarm optimization algorithm （PSO） easily falling into local optimization when solving multi-extreme value function, a multi-stages optimi- zation algorithm is presented through analyzing the evolution principle and the reason of premature convergence in this paper. The improved algorithm divides the evolution process into multi-stages. In addition, a different fit iterative formula is adopted in each stage, so the population diversity can be increased and the premature convergence ean be effectively avoided accordingly. Simulation results show that the improved algorithm has better global extreme function problems optimization capability than standard PSO on multi-