群智能优化算法反演概率积分参数的性能比较与分析  

Comparison and Analysis of the Performance of Swarm Intelligence Optimization Algorithms for Inversion of Probability Integral Parameters

作　　者：黄金中 王磊[1] 李靖宇 蒋创 滕超群 李忠[1] 李世保 HUANG Jinzhong;WANG Lei;LI Jingyu;JIANG Chuang;TENG Chaoqun;LI Zhong;LI Shibao(School of Spatial Information and Surveying Engineering,Anhui University of Science and Technology,Huainan 232001,China)

机构地区：[1]安徽理工大学空间信息与测绘工程学院,安徽淮南232001

出　　处：《金属矿山》2022年第8期173-181,共9页Metal Mine

基　　金：国家自然科学基金项目(编号:52074010);安徽省自然科学基金优青项目(编号:2108085Y20);安徽理工大学研究生创新基金项目(编号:2021CX2142)。

摘　　要：目前,各种群智能优化算法涌现且各有特色、性能各异、普适性不强、在开采沉陷领域应用较少,更为重要的是,众多学者面对该类算法,无法有效选择最优算法进行开采沉陷研究。常见的群智能优化算法中狮群算法(Lion Swarm Optimization,LSO)、蝙蝠算法(Bat Algorithm,BA)、人工鱼群算法(Artificial Fish School Algorithm,AFSA)具有不同的特征,且在概率积分参数反演中鲜有应用。为此,将上述3种群智能优化算法引入概率积分参数反演中,并从抗随机误差性能、抗粗差干扰性能、观测点缺失的抗干扰性能、参数波动性和全局搜索性能等几个角度对群智能优化算法进行研究分析。模拟试验及工程实例分析表明,上述3种群智能优化算法均满足应用精度要求。LSO算法在抗随机误差干扰影响、观测点缺失的抗干扰能力方面以及参数结果总体波动性方面相对于BA、AFSA算法有一定优势;BA算法在抗粗差干扰能力方面优于LSO、AFSA算法;在全局搜索性能方面,随着反演参数解空间范围扩大为原来的2倍后,用AFSA算法反演概率积分参数的精度优于LSO、BA算法。通过详细比较分析,总结了上述3种算法在开采沉陷中的性能表现,可为有效选择合适的群智能优化算法进行概率积分参数反演提供参考。At present,various swarm intelligence optimization algorithms have emerged with their own characteristics,different performances,not strong universality,and few applications in the field of mining subsidence,more importantly,many scholars are unable to effectively choose the optimal algorithm for mining subsidence research in the face of this type of algorithm.Among the common swarm intelligence optimization algorithms,lion swarm algorithm(LSO),bat algorithm(BA),and artificial fish swarm algorithm(AFSA)have different characteristics,and are rarely used in probability integral parameter inversion.Therefore,the above three population intelligent optimization algorithms are introduced into the probability integral parameter inversion,and the swarm intelligent optimization algorithm is studied and analyzed from the aspects of anti random error performance,anti gross error interference performance,anti-interference performance of missing observation points,parameter fluctuation and global search performance.The results of simulation experiments and engineering examples show that the above three population intelligence optimization algorithms all meet the requirements of application accuracy.The LSO algorithm has certain advantages over the BA and AFSA algorithms in terms of anti-interference ability against random error interference,missing observation points,and overall volatility of parameter results;The BA algorithm is better than the LSO and AFSA algorithms in the ability to resist gross error interference;In terms of global search performance,as the range of the inversion parameter solution space is expanded to twice the original,the accuracy of the inversion of probability integral parameters using the AFSA algorithm is higher than that of the LSO and BA algorithms.Through detailed comparison and analysis,the performance of the above three algorithms in mining subsidence is summarized,which provides a reference for effectively selecting the appropriate swarm intelligence optimization algorithm for probability in

关 键 词：开采沉陷预计 概率积分法 参数反演 群智能优化算法 

分 类 号：TD325[矿业工程—矿井建设]