基于高斯概率分布采样学习的SSA算法探讨

Discussion on SSA Algorithm Based on Gaussian Probability Distribution Sampling Learning

针对基本的樽海鞘群算法(Salp Swarm Algorithm,SSA)收敛性能差等缺点,利用基于高斯概率分布采样学习的方法对樽海鞘群算法进行优化.文章用一种基于全局最优位置的高斯采样方式改变樽海鞘(追随者)的位置更新公式,在一定程度上避免陷入局部最优,同时提高收敛性能.通过8个不同特征的测试函数仿真,与其他算法的实验结果进行对比,验证了所提算法在收敛速度、收敛精度和局部搜索能力方面的良好性能.

Aiming at the shortcomings such as the poor convergence performance of the basic Salp Swarm Algorithm(SSA),the algorithm based on Gaussian probability distribution sampling learning is used to optimize the Salp Swarm algorithm.In this paper,a Gaussian sampling method based on global optimal position is used to change the position update formula of the Salp(follower),which avoids falling into local optimum to some extent and improves convergence performance.Through the test function simulation of eight different features,compared with the experimental results of other algorithms,the performance of the proposed algorithm in terms of convergence speed,convergence precision and local search ability is verified.