The Inverse Problem Of Electromagnetic Field Analysis Of The Fast Global Optimization Algorithm

Thanks to the advances in both the computer science and electromagnetics, the inverse electromagnetic problem has become a topical subject in Computational Electromagnetics in order to meet the ever-increasing requirements in design automations of electromagnetic devices.Based on a comprehensive analysis and integration of available global optimal algorithms, this dissertation mainly focuses on the study of fast global optimal methods applied to inverse electromagnetic problems.Firstly, the Genetic Algorithm (GA) and the Particle Swarm Optimization (PSO) methods are studied. Different improvements are proposed to develop robust optimizers in senses of both convergence speeds and global search abilities. Typical mathematic functions are solved to demonstrate the effectiveness and efficiency of the improved GA and PSO.Secondly, to develop computationally efficient global optimal optimizers with the main goal of releasing the requirement for tremendously heavy computation resources especially to reduce the number of function calls that involve computationally heavy procedures in the study of electromagnetic inverse problems, the response surface models (RSM) or methodologies are incorporated into the stochastic global optimal methods. For this purpose, the compactly supported radial basis function and the moving least squares approximation based ones are integrated into the aforementioned two global optimizers. As a result, two novel hybrid algorithms are introduced.Finally, the proposed hybrid algorithms are applied successfully to study benchmark inverse electromagnetic problems. The numerical results of TEAM Workshop Problem 22 and Problem 25 as reported validate the feasibility and the robustness of the corresponding algorithms.