本文關鍵詞： 風驅(qū)動優(yōu)化算法 變異算子 方向圖綜合 阻抗匹配網(wǎng)絡 濾波器　出處：《江蘇科技大學》2016年碩士論文　論文類型：學位論文

【摘要】：風驅(qū)動優(yōu)化(Wind Driven Optimization,WDO)算法是由Bayraktar Z等人在2010年提出的一種基于群體的全局優(yōu)化算法,該算法模擬空氣質(zhì)點在大氣中的受力運動。相比于其他智能優(yōu)化算法,該算法魯棒性強、概念清晰、易于實現(xiàn)且尋優(yōu)效率高,適用于解決多維和多模態(tài)問題,也可以處理連續(xù)和離散優(yōu)化問題,因此越來越引起研究學者們的關注。電磁綜合問題多為模型結(jié)構復雜、目標函數(shù)非線性的復雜數(shù)學問題,傳統(tǒng)的優(yōu)化算法已經(jīng)不再滿足電磁綜合問題的高求解精度和效率的要求。自2010年Bayraktar Z等人使用WDO算法優(yōu)化設計了雙面人工磁導體,WDO算法在電磁綜合問題中的應用便層出不窮。然而,每一種算法都有其強項和弱項,在解決部分多峰值問題時,WDO算法也存在收斂性差的情況。為了改善這種情況,本文對基本W(wǎng)DO算法進行了適當?shù)母倪M,并將改進后的算法應用于幾個電磁綜合問題,本文的主要研究成果可歸納如下:(1)提出了五種不同變異策略的WDO算法,進行了數(shù)值仿真實驗并分析,總結(jié)出最能均衡全局探索能力與局部開發(fā)能力的WDOWM算法。(2)在WDO算法中引入雙層學習機制,形成雙層變異策略WDO算法,應用經(jīng)典測試函數(shù)測試并分析。(3)介紹了均勻設計法確定參數(shù)組合,從而提高算法計算效率,進而實現(xiàn)對該算法不斷的發(fā)展和完善。(4)研究了WDO算法和WDOWM算法用于線陣方向圖綜合問題。WDOWM算法綜合出具有更低副瓣電平和零陷深度的結(jié)果,優(yōu)于現(xiàn)有文獻結(jié)論。(5)研究了WDO算法和WDOWM算法用于阻抗匹配網(wǎng)絡綜合問題。WDOWM算法仿真的11S曲線均滿足設計要求,相比于WDO算法優(yōu)化的結(jié)果更具有工程應用價值。(6)提出了基于WDOWM算法和HFSS軟件的電磁優(yōu)化方法。利用HFSS-MATLAB-API接口程序?qū)崿F(xiàn)高效電磁仿真,應用該電磁優(yōu)化方法建模優(yōu)化設計了一個雙層EBG結(jié)構濾波器。
[Abstract]:Wind driven Driven Optimization algorithm is a population-based global optimization algorithm proposed by Bayraktar Z et al in 2010. The algorithm simulates the force motion of air particles in the atmosphere. Compared with other intelligent optimization algorithms, this algorithm is robust. The concept is clear, easy to realize and the optimization efficiency is high. It is suitable for solving multi-dimensional and multi-modal problems, and can also deal with continuous and discrete optimization problems. Complex mathematical problems with nonlinear objective functions, The traditional optimization algorithm is no longer satisfied with the requirement of high precision and efficiency in solving electromagnetic synthesis problem. Since 2010, Bayraktar Z and others have used WDO algorithm to optimize the design of double-sided artificial magnetic conductor WDO algorithm in electromagnetic synthesis problem. But, Each algorithm has its own strengths and weaknesses, and the convergence of WDO algorithm is poor when solving partial multi-peak problem. In order to improve this situation, the basic WDO algorithm is improved properly. The improved algorithm is applied to several electromagnetic synthesis problems. The main research results of this paper can be summarized as follows: (1) five WDO algorithms with different mutation strategies are proposed, and the numerical simulation experiments are carried out and analyzed. This paper summarizes the WDOWM algorithm, which can balance the global exploration ability with the local development ability, and introduces a two-layer learning mechanism into the WDO algorithm to form a two-layer mutation strategy WDO algorithm. This paper introduces the uniform design method to determine the parameter combination by using the classical test function test and analysis, which improves the calculation efficiency of the algorithm. Then, the WDO algorithm and the WDOWM algorithm are studied to synthesize the linear array pattern synthesis problem. The results with lower sidelobe level and zero sink depth are obtained. WDO algorithm and WDOWM algorithm for impedance matching network synthesis problem. WDOWM algorithm simulation of 11s curve meet the design requirements. Compared with the result of WDO algorithm optimization, this paper presents a method of electromagnetic optimization based on WDOWM algorithm and HFSS software, which is more valuable in engineering application. The HFSS-MATLAB-API interface program is used to realize high efficiency electromagnetic simulation. The electromagnetic optimization method is used to model and design a double layer EBG filter.
【學位授予單位】：江蘇科技大學
【學位級別】：碩士
【學位授予年份】：2016
【分類號】：TP18
，

本文編號：1526385