CAREER: Electromagnetic Component Design Using Genetic Algorithms

职业：使用遗传算法进行电磁元件设计

• 批准号: 9502138
• 负责人: Eric Michielssen
• 金额: $ 31万
• 依托单位: University of Illinois at Urbana-Champaign
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1995
• 资助国家: 美国
• 起止时间: 1995-05-15 至 2001-04-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9502138&HistoricalAwards=false
• 关键词: CAREER; Electromagnetic; Component; Design; Using

9502138 Michielssen Research The objective functions that arise in the context of electromagnetic optimization and component design are often highly non-linear, stiff, multi-external, sometimes non-differentiable, and almost always computationally expensive to evaluate. the effectiveness of Genetic Algorithms (Gas) is investigated within the context of electromagnetic component design by developing global GA- based optimizers for a number of electromagnetic systems, e.g., reflect arrays, frequency scanned grating antennas, waveguide components, and broadband antennas systems. It is conjectured that the proposed research will engender novel solution techniques for a variety of electromagnetic optimization problems and bring about a number of powerful, highly efficient and systematic design techniques for a aforementioned electromagnetic components. This study will also permit the formulation of guidelines regarding the suitability of Gas for other electromagnetic optimization problems in terms of number and the character of those parameter that require optimization the computational complexity of the forward analysis problem, and the available computing platform. Education Teaching excellence is an important component of all first-class undergraduate and graduate programs. This young faculty member in electromagnetic will be committed to significantly revise a number of electromagnetics courses during the next for years in order to maintain and improve the quality of the undergraduate and graduate electromagnetic curricula at the University of Illinois at Urbans-Champaign (UIUC). The recently established College of Engineering's interdisciplinary program in Computational Science and Engineering (CSE) will instrumental in achieving the goal. Furthermore, he will foster discussions regarding the establishment and the promotion of an exchange program for undergraduate and graduate students between the Colleges of Engineering at the UIUC and Belgian universities.

在电磁优化和元件设计中出现的目标函数通常是高度非线性的，刚性的，多外部的，有时不可微的，并且几乎总是计算昂贵的。在电磁元件设计的背景下，通过开发基于全局遗传算法的优化器来研究遗传算法（Gas）的有效性，这些优化器适用于许多电磁系统，例如反射阵列、频率扫描光栅天线、波导组件和宽带天线系统。预计本研究将为各种电磁优化问题提供新颖的求解技术，并为上述电磁元件带来一系列强大、高效、系统的设计技术。这项研究还将允许根据需要优化的参数的数量和特征、正演分析问题的计算复杂性以及可用的计算平台，制定关于Gas对其他电磁优化问题的适用性的指导方针。优秀的教学是所有一流的本科和研究生课程的重要组成部分。为了保持和提高伊利诺伊大学香槟分校本科和研究生电磁学课程的质量，这位年轻的电磁学教员将致力于在未来几年内大幅修订一些电磁学课程。最近成立的工程学院计算科学与工程（CSE）跨学科项目将有助于实现这一目标。此外，他还将推动建立和促进UIUC工程学院与比利时大学之间本科生和研究生交流项目的讨论。