队伍演化算法及其在微波电路设计中的应用

发布时间：2018-08-22 16:00

【摘要】：随着信息技术的发展,优化问题日趋复杂,而传统的优化方法在寻优时往往会有很大的局限性。以进化优化算法为主的现代优化算法是从生物智能或物理现象上获得启发建立起来的,它们具有适于高度并行、全局性与自适应等特点,用目的性和系统性的过程代替传统的手工实验方法,为解决复杂问题提供了一种新的途径。在微波电路设计中,许多应用问题都要求元件优化时要适应于各类型的设计变量、指标约束及单元组合结构的线性、非线性等。HFSS作为微波电路设计中的一款重要的电磁仿真软件,为用户提供了简明易用的建模界面和功能强大的处理器,能够计算分析各种复杂3D无源元件的电磁特性。粒子群优化算法(Particle Swarm Optimization,PSO)作为一种经典的群智能进化算法,由于其原理简单,较易实现等特点而被广泛研究和应用。本文对该算法深入研究,为了加快优化速度,提高收敛精度,提出了一种基于PSO的队伍演化算法(Team Evolutionary Algorithm,TeamEA),并将其与HFSS软件的接口脚本相结合,完成微波元件的建模、仿真与优化。主要工作概括如下:1.对PSO算法的基本概念、实现方法及其存在的问题进行分析描述,进一步阐述了该算法的研究现状和研究趋势。2.对优化算法中一些常用的测试函数深入研究,将原有测试函数进行旋转平移操作,新得到的测试函数更加复杂多变,对算法的性能提出了更高的要求,用于检验算法对于现实存在的许多未知结构或空间的适应能力与寻优能力。3.对于PSO算法存在的一些不足,引入多种群策略,提出了队伍演化算法。该算法将优化过程分为两个阶段:第一阶段为保持多样性,把队员分成若干个初级队伍并行优化,形成高级队伍;后一阶段为提高收敛速度,仅对高级队伍进行优化。在整个优化过程中,根据评估队员所取得的成绩,动态控制队员的调整步长和最大调整空间,同时产生教练组,为队员的进步方向提供指导。通过高维多峰测试函数及其旋转平移函数进行测试比较,验证了算法的优越性和有效性。4.研究通过宏命令控制HFSS软件实现自动建模,使用VBscript脚本语言实现Matlab与HFSS的联合仿真,利用队伍演化算法的寻优能力和HFSS强大的电磁分析能力,实现一种基于队伍演化算法和电磁仿真软件HFSS的优化方案。5.将该优化方案应用于微波电路设计中,通过对1.2~1.5 GHz发卡型滤波器、WLAN天线的成功设计优化,实验结果表明了基于队伍演化算法和HFSS软件的优化方案的有效性和应用价值。
[Abstract]:With the development of information technology, the optimization problem is becoming more and more complex, but the traditional optimization methods often have great limitations. The modern optimization algorithms, which are mainly evolutionary optimization algorithms, are constructed from the inspiration of biological intelligence or physical phenomena. They are suitable for highly parallel, global and adaptive, and so on. It provides a new way to solve complex problems by replacing the traditional manual experiments with purposeful and systematic processes. In microwave circuit design, many application problems require element optimization to adapt to various types of design variables, index constraints and the linearity of unit combination structures. As an important electromagnetic simulation software in microwave circuit design, HFSS provides users with a simple and easy to use modeling interface and powerful processor, which can calculate and analyze the electromagnetic characteristics of various complex 3D passive components. As a classical swarm intelligence evolutionary algorithm, particle swarm optimization (Particle Swarm) has been widely studied and applied due to its simple principle and easy implementation. In this paper, in order to accelerate the optimization speed and improve the convergence accuracy, a team evolution algorithm (Team Evolutionary algorithm based on PSO is proposed, which is combined with the interface script of HFSS software to complete the modeling, simulation and optimization of microwave components. The main work is summarized as follows: 1. The basic concept, implementation method and existing problems of PSO algorithm are analyzed and described. Some commonly used test functions in the optimization algorithm are deeply studied, and the original test functions are rotated and translated. The new test functions are more complex and changeable, and the performance of the algorithm is required higher. It is used to test the adaptability and optimization ability of the algorithm to many unknown structures or spaces in reality. For the shortcomings of PSO algorithm, a team evolution algorithm is proposed by introducing a variety of group strategies. The algorithm divides the optimization process into two stages: in the first stage, the players are divided into several junior teams to form a senior team in parallel, and in the latter stage, only the senior team is optimized to improve the convergence speed. In the whole optimization process, according to the evaluation of the achievements of the players, the dynamic control of the players to adjust the step size and the maximum adjustment space, at the same time, the production of coaching staff, to provide guidance for the direction of improvement of the players. The advantages and effectiveness of the algorithm are verified by comparing the high dimensional multimodal test function and its rotation translation function. The automatic modeling is realized by controlling HFSS software by macro command, the joint simulation of Matlab and HFSS is realized by using VBscript script language, and the optimization ability of team evolution algorithm and the powerful electromagnetic analysis ability of HFSS are used. To realize an optimization scheme based on team evolution algorithm and electromagnetic simulation software HFSS. 5. The optimization scheme is applied to microwave circuit design. The successful design of 1.2U 1.5 GHz filter antenna shows the effectiveness and application value of the optimized scheme based on team evolution algorithm and HFSS software.
【学位授予单位】：杭州电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN454;TP18

【相似文献】