基于协同进化的大规模优化算法

发布时间：2017-12-28 08:06

本文关键词：基于协同进化的大规模优化算法　出处：《郑州大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着工程技术的发展与优化问题数学模型的逐渐完善,很多优化问题从最初的低维优化发展成为现在的高维优化问题。因此具有高维特性的大规模问题成为了当前研究的热点问题,并且在工程实践中也有广泛应用。虽然最近几年进化优化在许多实值和组合优化问题上取得了很大的成功,但是大多数的随机优化算法,包括粒子群优化算法、差分进化算法和遗传算法都会遭受“维数灾难”,随着搜索空间维数的增加,算法的性能急剧退化。因此,与拓扑结构简单的低维问题相比,大规模问题的全局最优解是很难找到的。综上所述,本文在动态多种群粒子群优化算法中引入协同进化策略来求解大规模优化问题。首先,本文介绍了大规模优化问题的研究背景及意义,介绍了进化算法在大规模优化领域的研究现状。重点讲述了粒子群优化算法在大规模问题中的应用,并介绍了粒子群优化算法的特点。随后又简述了近些年来粒子群优化算法的发展历程和研究方向。同时介绍了一些大规模优化问题的解决方案,为算法对比提供了基础。其次,介绍了大规模优化算法的基准测试函数,说明这些函数的性质,表现出大规模问题的特点。接着,提出了高维空间中的种群初始化方法。主要通过与基本的随机数生成器的初始化结果进行对比,说明在大规模优化问题中,不同初始化方法的不同作用。使算法的种群在决策空间中分布更加均匀,使算法避免陷入局部最优解。最后,针对大规模优化问题的特点,提出多种群动态的协同进化算法,重点讨论了该算法的特点、以及算法中涉及到的策略调整等。使用动态粒子群优化算法来解决具有大量决策变量的优化问题。通过与近些年提出的几个大规模进化算法的对比试验,验证算法的有效性。
[Abstract]:With the development and optimization of engineering technology and the gradual improvement of mathematical models, many optimization problems develop from the initial low dimensional optimization to the current high-dimensional optimization problem. Therefore, the large-scale problem with high dimensional characteristics has become a hot issue in the current research and has been widely used in engineering practice. Although in recent years evolutionary optimization has achieved great success in many real and combinatorial optimization problems, but most of the stochastic optimization algorithms, such as particle swarm optimization algorithm, differential evolution algorithm and genetic algorithm will suffer from the "Curse of dimensionality", with the increase of the search space dimension, the algorithm performance degradation. Therefore, it is difficult to find the global optimal solution of the large-scale problem compared with the low dimensional problem with simple topology. To sum up, this paper introduces co evolution strategy in dynamic multigroup particle swarm optimization algorithm to solve large-scale optimization problems. First, this paper introduces the research background and significance of the large-scale optimization problem, and introduces the research status of the evolutionary algorithm in the large-scale optimization field. This paper focuses on the application of particle swarm optimization in large-scale problems, and introduces the characteristics of particle swarm optimization (PSO) algorithm. Then the development and research direction of particle swarm optimization (PSO) algorithm in recent years are briefly described. At the same time, some solutions for large-scale optimization problems are introduced, which provides a basis for algorithm comparison. Secondly, the reference test function of large-scale optimization algorithm is introduced, and the properties of these functions are illustrated, and the characteristics of large-scale problems are shown. Then, the method of population initialization in high dimensional space is proposed. By comparing with the initialization results of the basic random number generator, it is shown that the different initialization methods are different in the large-scale optimization problem. The population of the algorithm is more evenly distributed in the decision space, so that the algorithm can avoid falling into the local optimal solution. Finally, aiming at the characteristics of large-scale optimization problems, we propose a multigroup dynamic co evolutionary algorithm, focusing on the characteristics of the algorithm and the strategy adjustment involved in the algorithm. The dynamic particle swarm optimization (PSO) algorithm is used to solve the optimization problem with a large number of decision variables. The effectiveness of the algorithm is verified by comparison with several large-scale evolutionary algorithms proposed in recent years.
【学位授予单位】：郑州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18

【相似文献】