【摘要】 针对沙猫群优化算法存在全局探索能力弱、收敛速度慢等问题，提出一种混合多策略改进的沙猫群优化算法。采用佳点集策略增强种群分布的均匀性，设计一种非线性函数自适应调节机制以提升寻优精度，融入三角形游走和Circle混沌映射加强局部开发能力，依据折射反向学习动态优化解集质量。构建ISCSO的Markov链模型，运用鞅理论验证其依概率收敛到全局最优。通过测试函数评估改进算法的优化性能，并将其应用在机器人路径规划问题中。实验结果表明ISCSO得到的路径平滑度更好，相较于SCSO算法，搜索耗时减少了5.47%，最优路径长度缩短了17.49%。更多还原

【Abstract】 To address the problems of the sand cat swarm optimization algorithm, including weak global exploration capability and slow convergence rate, a hybrid multi-strategy improved sand cat swarm optimization algorithm(ISCSO) was proposed. The good point set strategy was employed to enhance the uniformity of population distribution. A novel adaptive adjustment mechanism for nonlinear functions was designed to improve the optimization accuracy of the algorithm. Triangle walk and Circle chaotic mapping were integrated to enhance local exploitation capability. Refraction opposition-based learning was utilized to dynamically optimize the quality of the solution set. A Markov chain model of ISCSO was constructed, and its convergence in probability to the global optimum was proven using martingale theory. The optimization performance of the improved algorithm was evaluated through benchmark functions and applied to the robot path planning problem. Experimental results demonstrate that ISCSO generates significantly smoother paths. Compared with the SCSO algorithm, the search time is reduced by 5. 47%, and the optimal path length is decreased by 17. 49%.更多还原