采用多搜索策略的改进贫富优化算法  

Improved Poor and Rich Optimization Algorithm Using Multiple Search Strategies

作　　者：刘棋锴 王勇[1,2] LIU Qikai;WANG Yong(College of Artificial Intelligence,Guangxi Minzu University,Nanning 530006,China;Guangxi Key Laboratory of Hybrid Computation&IC Design Analysis,Nanning 530006,China)

机构地区：[1]广西民族大学人工智能学院,广西南宁530006 [2]广西混杂计算与集成电路设计分析重点实验室,广西南宁530006

出　　处：《广西民族大学学报（自然科学版）》2022年第3期58-67,共10页Journal of Guangxi Minzu University ：Natural Science Edition

基　　金：国家自然科学基金项目(61662005);广西自然科学基金项目(2021JJA170094)。

摘　　要：针对标准贫富优化算法(PRO)存在的不足,提出采用多搜索策略的改进贫富优化算法(IPRO):第一,当前群体最富者(最优者)利用柯西变异对自身施加扰动变异,以增强自身的开发能力;第二,一般富人在搜索中既要考虑自身与最富者之间的差距,又要考虑其与最优穷人之间的阶级差距,利用Levy飞行增加了一般富人搜索步长的灵活性和多样性;第三,穷人群体在算法前期和后期均可动态调整其搜索策略,增加了穷人个体搜索的多样性,进而增强了算法的全局搜索能力。数值实验结果表明:改进算法的全局搜索速度和优化精度均得到了明显的提高,算法规避陷入局部最优的能力得到了增强。最后,利用活塞杆设计参数优化实例验证了IPRO具有较好的实用性。In view of the shortcomings of the standard wealth-poor optimization algorithm(PRO),an improved wealth-poor optimization algorithm(IPRO)using multiple search strategies is proposed:Firstly,the richest person(the best person)in the current group uses the Cauchy mutation to impose perturbation mutation on themselves to enhance their own development ability;Secondly,the average rich people need to consider not only the gap between themselves and the richest people,but also the class gap between them and the best poor people in their search. Using Levy flight increases the flexibility of the average rich person’s search step size and diversity;Thirdly,the poor groups can dynamically adjust their search strategies in the early and late stages of the algorithm,which increases the diversity of poor individuals’ searches,thereby enhancing the algorithm’s global search capabilities.The numerical experimental results show that the global search speed and optimization accuracy of the algorithm in this paper have been significantly improved,and the ability of the algorithm to avoid falling into local optimum has been enhanced.Finally,an example of piston rod design parameter optimization is used to verify that IPRO has good practicability.

关 键 词：贫富优化算法(PRO) 柯西扰动 列维飞行 动态调整机制 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]