CAJ | 학술논문

针对微粒群优化算法的早熟停滞缺陷问题，提出了一种基于种群年龄模型的动态粒子数微粒群优化算法．该算法建立了生物种群年龄模型，将每个粒子划分为不同的年龄段，动态地依据种群环境和个体信息有效地控制种群的粒子数规模；设计了较优粒子的生殖策略和较差粒子的死亡策略，增加群体的多样性和减少冗余计算量，以保证算法获得最优性能．将此算法与其他改进算法进行比较，仿真测试结果表明，新算法具有较高的全局搜索成功率和效率，计算量显著降低，优化精度显著提高，能够有效地避免算法陷入局部停滞的缺点．
침대미립군우화산법적조숙정체결함문제，제출료일충기우충군년령모형적동태입자수미립군우화산법．해산법건립료생물충군년령모형，장매개입자화분위불동적년령단，동태지의거충군배경화개체신식유효지공제충군적입자수규모；설계료교우입자적생식책략화교차입자적사망책략，증가군체적다양성화감소용여계산량，이보증산법획득최우성능．장차산법여기타개진산법진행비교，방진측시결과표명，신산법구유교고적전국수색성공솔화효솔，계산량현저강저，우화정도현저제고，능구유효지피면산법함입국부정체적결점．
For the problem that particle swarm optimization （PSO） algorithm often suffers from being trapped in local optima so as to be premature convergence, a dynamic particle population PSO based on population age model is proposed to efficiently control premature stagnation. Firstly, life population age model is constructed, which divides diverse age group for a certain particle, and effectively regulates population size in accordance with population environment and individual information. Secondly, the particle reproduction strategy of the good particle for keeping the diversity of swarm and the particle vanishing strategy of the worst particle for decreasing optimal performance is guaranteed in this algorithm. excessive calculated amount are designed, so the Finally, the comparison experiments have been made with four benchmark functions between the proposed algorithm and other improved PSO. The simulation experimental results show that the proposed method not only greatly improves the global successful searching probability and searching efficiency, but also effectively avoids the local stagnation problem.