Abstract: Three kinds of programming operators involving combination operator, harmony operator and abruption operator have been defined to perform community dynamic distribution in particle swarm optimization. Thus a new random optimization tool has been developed, i.e. community dynamic assignation-based particle swarm optimization (CDAPSO). Via changing organization structure and distribution character of particle swarm to increase varieties of eliciting information during optimization, this algorithm notably improves global searching abilities and effectively avoid problems of premature and convergence.

摘要： 通过定义三类群落规划算子：合并算子、融合算子和裂分算子，实现了粒子群优化算法进程中的群落动态分配思想，从而构造了一种新的随机优化技术：基于群落动态分配的粒子群优化算法(Community Dynamic Assignation-based Particle Swarm Optimization, CDAPSO)。新算法通过动态改变粒子群体的组织结构和分配特征来维持寻优过程中启发信息的多样性，从而使其全局收搜索能力得到了显著提高，并且能够有效避免早熟收敛问题。