基于传感器阵列多目标优化的SP-QPSO算法研究  

Research on SP-QPSO Algorithm Based on Multi-Objective Optimization of Sensor Array

作　　者：孔宇航 陶洋[1] 梁志芳 KONG Yu-hang;TAO Yang;Liang Zhi-fang(School of Communication and Information Engineering,Chongqing University of Posts and Telecommunications,Chongqing 400065,P.R.China)

机构地区：[1]重庆邮电大学通信与信息工程学院,重庆400065

出　　处：《新一代信息技术》2021年第24期10-17,共8页New Generation of Information Technology

基　　金：重庆市基础研究与前沿探索项目(项目编号:cstc2018jcyjAX0549);重庆市教育委员会科学技术研究项目(项目编号:KJQN201800617)。

摘　　要：针对高维空间中搜索算法面临“种群退化”现象,且通常涉及约束优化问题。提出了一种基于改组复合体演化量子行为粒子群优化算法(SP-QPSO)用于电子鼻传感器阵列多目标优化研究方法。其中基于改组复合体演化算法(Shuffled complex evolution with PCA,SP)用于构建复合体并监测种群维数的变化。量子行为粒子群优化算法(Quantum behaved particle swarm optimization,QPSO)用于每个复合体在搜索空间的演化。同时引入自适应惩罚函数计算搜索空间的违反度,用于指导搜索空间可行区域的求解。实验结果表明SP-QPSO算法明显优于其它对比算法,达到了90.1%。而且该算法将传感器阵列的数量降至6个以下,最优阵列的整体规模更小。The search algorithm in high-dimensional space is faced with the phenomenon of"population deg-radation",and usually involves the problem of constrained optimization.In this paper,an evolutionary quan-tum behavior particle swarm optimization(SP-QPSO)algorithm based on reorganization complex is proposed for multi-objective optimization of electronic nose sensor array.Among them,the shuffled complex evolution with PCA(SP)is used to build complex and monitor the change of population dimension.Quantum behaved particle swarm optimization(QPSO)is used for the evolution of each complex in search space.At the same time,the adaptive penalty function is introduced to calculate the violation degree of search space,which is used to guide the solution of the feasible region in the search space.The experimental results show that SP-QPSO algorithm is superior to other comparison algorithms,reaching 90.1%.Moreover,the number of sensor arrays is reduced to less than 6,and the overall scale of the optimal array is smaller.

关 键 词：电子鼻 传感器阵列 多目标优化 量子行为粒子群优化算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]