中文摘要：

针对微粒群优化解决复杂优化问题时易陷入局部收敛、效率不高的缺点,提出一种基于动态邻域和自适应惯性权重的微粒群优化算法。通过定义动态邻域及其最优维值,提出种群个体的动态邻域最优维值学习策略,使微粒跟踪个体极值和邻域的最优维值进行搜索,以增加学习样本的多样性,避免局部收敛;提出一种基于个体适应度的惯性权重动态调整方法,提高算法的寻优效率。通过优化5个典型测试函数验证了本文所提方法的有效性。

英文摘要：

Aimed to the disadvantage that the particle swarm optimization is easy to fall into the local convergence, and has low efficiency, a particle swarm optimization based on dynamic neighborhood topology and self-adaptive inertia weight is proposed. Firstly, by defining the dynamic neighborhood and its optimal dimension value, a learning strategy on optimal dimension values of dynamic neighborhood is proposed to lead the particles track the optimal dimension values of personal best positions and neigh- borhoods, to increase the diversity of learning samples, for avoiding the local convergence. Secondly, a self-adaptive method based on individuals＇ fitness is proposed to adjust the inertia weight in order to improve the searching efficiency of the proposed algorithm. Finally, the results on five typical tests verifies the effectiveness of the proposed method.