基于文化改进量子粒子群算法的无线传感器网络声源定位  

Source Localization Algorithm Based on Cultural Modified Quantum Particle Swarm Optimization Algorithm for Wireless Sensor Networks

作　　者：刘玉柱[1] 孙学梅[1] 武继刚[1] 

机构地区：[1]天津工业大学计算机科学与软件学院,天津300387

出　　处：《小型微型计算机系统》2013年第12期2786-2789,共4页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61173032)资助

摘　　要：无线传感器网络中基于声音能量的声源定位常采用最大似然估计法,该方法将定位问题转换为非线性函数的极值优化问题.本文提出一种文化-改进的量子粒子群优化算法(CMQPSO)解决这一非线性优化问题.首先,在量子粒子群(QPSO)的基础上,结合自适应变异思想和RSNTO算法,提出改进的量子粒子群算法(MQPSO).然后,为了进一步改善算法的全局搜索能力、提高计算精度,利用文化算法的双重演化机制,将改进的量子粒子群算法纳入文化算法框架形成本文提出的CMQPSO算法.大量仿真实验表明,CMQPSO算法在全局搜索能力和收敛性能上较PSO、混合PSO-SNTO算法都有很大的提高;在解决声源定位上,CMQPSO算法与其他优化算法相比,定位精度有了明显提高.Maximum Likelihood Estimation is often used in the acoustic source localization based on energy in wireless sensor network , which transform localization problem into global optimization problem of nonlinear function. This paper presents a Cultural Modi- fied quantum particle swarm optimization algorithm （CMQPSO） to solve this nonlinear optimization problem. First, on the basis of the quantum particle swarm （QPSO） , combined with adaptive the variability ideas and RSNTO algorithm, the modified quantum par- ticle swarm optimization （ MQPSO } was proposed. Then, in order to further improve the global search ability and calculation accura- cy of the algorithm and according to the dual evolutionary mechanisms of the cultural algorithm, we integrated MQPSO algorithm into cultural algorithm framework to form a CMQPSO algorithm proposed in this paper. Simulation results demonstrate that CMQPSO al- gorithm can achieve robust convergence performance and better localization accuracy.

关 键 词：量子粒子群优化 自适应变异 文化算法 声源定位 无线传感器网络 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]