A Global Optimization Algorithm Based on Support Vector Machines for Electromagnetic Inverse Problem

摘要

The problems of the lower convergence speedsand the long time for solving that exist in the global optimization algorithm of the inverse electromagnetic problem solution are given. The main reasons for these problems are analyzed. A global optimization algorithm based on Support Vector Machines of the inverse electromagnetic problem solution is presented. The numerical comparison shows that, comparing with the self-learning Simulated Annealing Algorithm, the times of solving forward electromagnetic problem is decreased greatly, so the speed of solving the inverse electromagnetic problem is improved noticeably.