摘要
为了解决麻雀搜索算法在迭代后期种群多样性降低且容易陷入局部最优的问题,提出一种采用危险度预警的进取麻雀搜索算法。首先制定保持进取策略,选取种群中适应度值优秀的个体与适应度值变化幅度大的个体共同培育生成学习样本,加快寻优速度。然后将种群进化过程分为稳定阶段和进取阶段,并在每个阶段执行不同策略:稳定阶段策略利用全局最优解和学习样本引导种群进化,增强局部开发能力;进取阶段策略利用个体历史最优解和学习样本引导种群进化,增加种群多样性并提高全局探索能力。此外,设计危险度预警策略用于检测种群是否陷入局部最优,并通过保存的优秀样本动态更新较差个体,帮助种群跳出局部最优。与其他几种有代表性的改进麻雀搜索算法一起针对CEC2017测试集进行对比实验,其结果证明了本文算法在收敛速度和收敛精度上有明显改进,能有效避免陷入局部最优。
A new aggressive sparrow search algorithm(SSA)with danger warning is proposed to solve the problem of reduced population diversity and great chance of falling into local optimum in the late iteration of the standard SSA.Firstly,a keeping aggressive strategy is formulated,and the individuals with excellent fitness value and those with large variation of fitness value in the population are selected to generate learning samples together to speed up the optimization.Then,the population evolution process is divided into stable period and aggressive period,and different strategies are implemented in each period.The stable period strategy uses the global optimal solution and learning samples to guide the evolution,which is beneficial for local exploitation.The aggressive period strategy uses the individual optimal experience and learning samples to guide the evolution,which can increase the population diversity and improve the global exploration ability.In addition,a danger warning strategy is developed to detect whether the population has been trapped in local optimum,and individuals with low fitness are dynamically updated by the saved excellent samples to help the population jump out of local optimum.Through comparative experiments with several other representative improved SSAs on CEC2017 test suite,it is proved that the proposed algorithm has significant improvement on convergence speed and solution accuracy,and can escape from the local optimum effectively.
作者
万琪
李俊
Wan Qi;Li Jun(College of Computer Science and Technology,Wuhan University of Science and Technology,Wuhan 430065,China;Hubei Province Key Laboratory of Intelligent Information Processing and Real-time Industrial System,Wuhan University of Science and Technology,Wuhan 430065,China)
出处
《武汉科技大学学报》
CAS
北大核心
2023年第2期153-160,共8页
Journal of Wuhan University of Science and Technology
基金
国家自然科学基金资助项目(61572381)
武汉科技大学智能信息处理与实时工业系统湖北省重点实验室基金资助项目(znxx2018QN06).
