Abstract: To improve the grid connection capability and the coupling optimization of gearbox,a particle swarm optimization algorithm based on food chain conduction response was proposed,which integrated the class enrichment conduction with mutation particle swarm algorithm according to biological energy conduction in food chain.To make up for the deficiency of algorithm's diversity preservation and optimization ability,the algorithm population was mutated and the mutation sequence was taken as the initial search solution to adjust the search location.Gauss difference quotient variation operator and inertia weight were integrated into search mechanism for updating operator adaptively.Penalty variation was applied to the particle with lower fitness for maintaining the search ability and variety of the algorithm.The effectiveness of the algorithm could be verified by solving the gearbox optimization problems which were occupied with multivariable coupling.

Key words: food chain, response optimization, Gauss difference variation, inertia weight variation, particle swarm algorithm

摘要： 为改善齿轮箱的耦合优化性能,依据食物链中生物能层层传导的机制,拟将层级富集传导优化与变异的粒子群算法相融合,提出一种基于食物链传导响应粒子群算法。为弥补算法多样性保持和寻优能力的不足,对算法种群进行变异化处理,获取变异序列作为搜索初始解,调整搜索位置。在搜索机制中融合高斯差商变异算子、惯性权重变异算子对粒子进行自适应更新,对适应度较低的粒子施加惩罚变异,以保持算法的搜索能力和多样性。通过对风力机齿轮箱优化问题的研究,验证了该方法解决多变量耦合优化问题的有效性。

关键词: 食物链, 响应优化, 高斯差商变异, 惯性权重变异, 粒子群算法