融合信息反馈共享与蜉蝣搜索机制的樽海鞘群算法

Salp swarm algorithm combining information feedback sharing and mayfly search mechanism

针对樽海鞘群算法(SSA)收敛速度慢和易陷入局部最优的问题,提出了一种融合信息反馈共享与蜉蝣搜索机制的改进樽海鞘群算法。使用Piecewise映射的方法进行种群初始化,使初始樽海鞘种群更均匀的覆盖可行域空间;采用信息共享机制,提出辅助领导者策略,改进领导者位置更新公式,增强全局搜索能力;利用进化学说以及正负反馈调节的思想,通过变异操作和自然选择原则选取更优领导者,从而提高搜索精度;最后,提出蜉蝣搜索机制,选取蜉蝣算法的交配公式,优化追随者位置迭代公式,使算法在后期更快收敛。通过在12个基准测试函数的多个维度以及17个CEC测试函数的实验,证明了改进樽海鞘群算法的综合性能,并通过消融实验验证了改进策略的有效性,实验结果表明,改进算法在收敛速度以及搜索精度上具有明显的优势。

Aiming at the problems of slow convergence speed and easy to fall into local optimum of SSA, this paper proposed an improved salp swarm algorithm combing information feedback sharing and mayfly search mechanism. Firstly, it used piecewise chaos to initialize the population to make the initial salps more evenly cover the feasible space. It adopted the information sharing mechanism to propose an auxiliary leader strategy, in order to improve the leader position update formula, and enhance the global search ability. It used the evolutionary theory and the idea of positive and negative feedback regulation, selecting better leaders through mutation operation and natural selection, so as to improve the search ability precision. Finally, it put forward the mayfly search mechanism, selected the mating formula of the mayfly algorithm to optimize the iterative formula of the follower position and make the algorithm converge faster in the later stage. Experiments on 12 benchmark functions and 17 CEC test functions prove the comprehensive performance of the improved salp population algorithm, and the ablation experiments verifiy the effectiveness of the improved strategy. The experimental results show that the improved algorithm has obvious advantages in convergence speed and search accuracy.