基于分区和分层搜索的并行粒子群算法*

为提高粒子群优化算法在优化问题中的效率，提出了并行粒子群优化算法(SLPSO)。其基本思想是并行机制+解空间压缩+分层搜索。主要工作包括：搜索空间划分为n个区，由n个子群并行搜索，将搜索结果最好的作为指定的搜索空间，即将搜索空间缩小到原解空间的(1/n)；提出了粒子群两层划分模型，底层利于扩大搜索范围，上层利于全局精细搜索。在四个基准函数上的优化实验表明，新方法比经典的IPPSO并行粒子群算法在解的精度上提高了80.37%。

Parallel particle swarm optimization algorithm based on space-division and layered search

To improve the efficiency of particle swarm optimization, this paper proposed a novel parallel particle swarm optimization algorithm(SLPSO).The basic idea is parallel mechanism and space-division and layered search. The main contributions include, divided whole search space into n sub ares. For certain generations, let the best sub area be the search space. This shrinked the search space to the solution space. Proposed two layers partition of particles, the lower and topper work well for global and local search respectively. By experiments on four benchmark functions, shows that, the new algorithm increases precision by 80.37% compared with IPPSO.