基于近邻族群学习和自适应变异的黏菌算法OA

To address the limitations of the Slime Mould Algorithm(SMA),such as slow convergence speed and susceptibility to local optima,this paper proposes an improved SMA named HKTSMA based on neighborhood group learning and adaptive muta-tion strategies.HKTSMA uses a perturbed Halton sequence for population initialization to enhance the uniformity and coverage of the population in the search space,improving global exploration capabilities.The dynamic convergence mechanism of the oscilla-tion factor is restructured to establish a nonlinear step-size adjustment model,balancing global search and local exploitation.An adaptive neighborhood group learning strategy is introduced to enhance population information utilization through dynamic neigh-borhood interactions,improving convergence speed and accuracy.A t-distribution-based adaptive mutation operator is incorpo-rated,utilizing dynamic degree-of-freedom parameters to adjust mutation strength and effectively escape local optima.A com-plete algorithmic framework with parameter sensitivity analysis is constructed,forming an improved algorithm with multi-strategy collaborative optimization features.Simulation experiments are conducted using selected benchmark functions from CEC2014,CEC2017,and CEC2019 to validate the effectiveness of the proposed strategies.Tests on the CEC2021 benchmark suite demon-strate the superiority of HKTSMA over other algorithms in terms of convergence accuracy,convergence speed,and Wilcoxon rank-sum test results.Finally,HKTSMA is applied to the optimization design of an industrial refrigeration system,further verify-ing its potential in solving engineering optimization problems.