The Inverse Problem Of Electromagnetic Field Analysis Of The Evolutionary Algorithm For Computing Research

The inverse electromagnetic problem has become a topical area of computational electromagnetics, and thus attracted a significant amount of efforts form both academicians and engineers alike.This dissertation focuses on the study of scalar and vector optimal methods which are suitable for the numerical solutions of inverse electromagnetic problems.Firstly, the scalar stochastic optimal method is studied to find the global optimal solution of a multimodal function. For this purpose, an improved genetic (evolutionary) method and an improved tabu search algorithm are proposed.Secondly, based on the multiobjective nature of an electromagnetic design problem, an improved evolutionary algorithm is proposed. In the proposed algorithm, the ranking approach is improved and used for the assignment of fitness values of a solution. Also, improvements, such as fitness sharing and so on, are introduced to enhance the performances of the proposed algorithm.Finally, the proposed optimal methods, including both scalar and vector ones, are applied successfully to study both mathematical functions and prototype problems of engineering problems. The numerical results as reported validate the feasibility and the robustness of the corresponding algorithm.