CAJ | 학술논문

针对群智能算法求解 Agent 联盟生成问题过程中易出现粒子过分聚集，导致多样性降低，甚至陷入局部最优现象提出一种基于改进量子粒子群的求解策略，在粒子过分聚集时借鉴实数编码遗传算法中的柯西变异使粒子聚集程度降低，进而维持了粒子的多样性。并采用多种群并行和最优粒子移民策略加快算法收敛。实验表明，该策略可以快速高效的求解Agent联盟，在运行效率上优于同类方法。
침대군지능산법구해 Agent 련맹생성문제과정중역출현입자과분취집，도치다양성강저，심지함입국부최우현상제출일충기우개진양자입자군적구해책략，재입자과분취집시차감실수편마유전산법중적가서변이사입자취집정도강저，진이유지료입자적다양성。병채용다충군병행화최우입자이민책략가쾌산법수렴。실험표명，해책략가이쾌속고효적구해Agent련맹，재운행효솔상우우동류방법。
For quantum particle swarm optimizationalgorithm to solve the Agent coalition generation problems particles which were liable to occur excessive accumulation, results in the decrease of diversity, which may be trapped in local optimal phenomenon, an improved strategy in particle gather too much reference to the Cauchy mutation in the real-coded genetic algorithm make the particles aggregation degree is reduced, thus maintaining the diversity of the particles.Using multiple populations in parallel and the optimal particle migration strategy to speed up the conver-gence.Experiments show that the algorithm can quickly efficient Agent union, on the operation efficiency is better than that of similar algorithms.