Abstract: Focused on the issue that the Particle Swarm Optimization（PSO） is easy to fall into local optimal and accuracy computation, this paper puts forward a Hybrid PSO Algorithm of Dynamic Topology with Levy mutation（DLPSO）. It introduces dynamic topology Dbest mechanism into the PSO algorithm to adapt particle swarm evolution, the particles are divided into global optimal particle, exploring particles and non target particle based on the solution at each iteration. Levy mutation is introduced in the adaptive mutation of offspring in the immune PSO to ensure the diversity. The DLPSO algorithm is tested by seven benchmark test functions, and the numerical experimental results show that the new algorithm has good convergence efficiency, high accuracy, and good stability.

Key words: particle swarm optimization, immune mechanism, dynamic topology, Levy mutation

摘要： 针对标准粒子群优化算法易陷入局部最优、收敛精度不高的问题，提出一种嵌入列维变异的混合动态粒子群算法（DLPSO）。算法在进化过程中采用动态拓扑Dbest策略以降低粒子趋同性，每次迭代时根据解的好坏将粒子分为全局最优粒子、探索粒子及无目标粒子，并对探索粒子进行分簇，簇内粒子的更新受到全局最优粒子及簇内最优粒子的共同影响；为确保粒子多样性，平衡局部搜索与全局搜索，采用免疫机制与自适应列维变异相结合的方式对粒子进行变异。利用7个测试函数对算法进行性能评价，数值仿真结果表明该算法搜索精度高且稳定性好，具有良好的收敛性能。

关键词: 粒子群优化, 免疫机制, 动态拓扑, 列维变异