【作者】 杨建宾；

【导师】 须文波； 孙力；

【作者基本信息】 江南大学 ， 计算机软件与理论， 2012， 硕士

【摘要】 随着社会不断进步和科学技术的飞速发展,社会化生产对准确度和精度的要求越来越高,优化设计在这些领域发挥着越来越重要的作用,各种智能算法应运而生,并起到不可替代的作用,粒子群优化算法就是其中一类,算法简单易用,对多种不同类型的工程问题具有较广泛的适应性和适用性。因此,很快有效地被应用到了多个领域。物联网是新一代信息技术的重要组成部分,随着物联网的深入研究和在逐渐各行各业的应用推进,无线传感器网络行业在其中起到关键的推动作用。无线传感器网络是信息感知和采集的一场革命,发挥着越来越重要的作用,采集的数据不仅需要准确的数值,而且更需要精确的位置信息。在一些特殊应用中,准确的节点定位尤为重要。本文首先研究量子粒子群优化算法(Quantum-behaved Particle Swarm Optimization, QPSO)与和声搜索算法(Harmony search,HS),针对两种算法中求解高维优化效果差的问题,提出了一种和声搜索和量子粒子群混合算法。在新算法中,在QPSO进化过程中每代产生的最优个体以新陈代谢方式进入和声记忆库中并进行和声搜索,利用和声搜索算法局部搜索能力强的优点。在算法迭代过程中不断调整参数,而且加入新的变异元素,维持了种群的多样性,避免算法陷入局部最优解,提高了算法全局搜索能力。通过对5个典型的Benchmark标准测试函数的测试结果表明,该算法收敛精度有较好的提高。证明了改进算法的有效性和合理性,适合求解高维复杂的全局优化问题。针对无线传感器网络节点定位精度不高,能量消耗较大的问题,利用智能算法在处理优化问题上的优势,提出了基于群体智能算法的节点定位中的优化算法,采用智能算法求解优化模型,求解节点的精确位置。提出了一种基于量子粒子群优化和和声搜索混合算法的无线传感器网络定位算法。在节点定位过程中,提高了定位精度,算法收敛速度快。同等情况下,可以减少定位过程中消耗的能量,一定程度上可以提高效率。实验证明是一种非常有效的方法。更多还原

【Abstract】 With the improvement of society and development of science and technology at full speed,the social production process in much field needs more and more accuracy and precision. Optimization design has played an increasingly important role in these field, various intelligent algorithms have emerged, and have played an irreplaceable role. Particle Swarm Optimization (PSO) algorithm is one of this kind, algorithm is simple to use, and has the widespread adaptivity and applicability in a variety of different types of engineering problems. Therefore, it has been applied effectively in many fields.Internet of Things is an important part of new generation information technology, with in-depth study of Internet of Things and it gradually promote the application of all walks of life, wireless sensor networks plays a key role in promoting. Wireless sensor network is a revolution in collection and perception of information, and has been playing an increasingly important role. Collecting data not only need accurate values, but also need precise location information. In some special applications, accurate location information of nodes is very important.The purpose of this paper is to research Quantum-behaved Particle Swarm Optimization algorithm (QPSO for short). For solving the problems of standard Harmony Search (HS) and quantum particle swarm optimization (QPSO)algorithm badly for solving high-dimensional optimization. A hybrid algorithm of harmony search and quantum particle swarm optimization algorithm is presented. In the new optimization algorithm, the best individual produced in each generation of QPSO evolution process into the harmony memory with the metabolic manner and using the advantages of Harmony Search algorithm’s strong local search ability. Moreover, an adjustable parameter is regulated and new element is added during the iteration process to maintain the diversity of the whole swarm. Therefore the algorithm can avoid falling into local optimal solution and increase the global search ability. Simulation tests of five typical functions shows that the proposed algorithm can efficiently improve accuracy of converge. And demonstrates that efficiency and rationality of the improved algorithm for solving high-dimensional and complex global optimization problem.For the issue of that accuracy is not high and energy consumption is large in wireless sensor network node-positioning, then a new algorithm for optimizing the localization of nodes is presented, which is based on the hybrid optimization algorithm of QPSO and HS model, and the final exact location of the nodes can be obtained by the optimization results of the proposed algorithm. In the node localization process, algorithm can improve the positioning accuracy, fast Converge to the optimal solution. Under the same circumstances, it can reduce the energy consumed in the process, can improve efficiency to some extent. Experiments proved it to be a very effective method.更多还原