Sequential inference method of electromagnetic optimization in engineering

Sequential inference method was presented for engineering electromagnetic optimization by the combination of statistical approximate models with sequential sampling method. In the implementation, approximate models include radial basis model and compactly supported radial basis model; sequential sampling and optimization process is composed of coarse optimization process and fine optimization process. Compared with traditional direct optimization method, the optimization models constructed by the sequential sampling method only need a small sample data, and the overall computational effort needed is much less than that by direct optimization method. Meanwhile, sequential method can provide prior information for the optimization algorithm. Finally, to illustrate the performance of the new method, Monte Carlo simulation experiment and the IEEE TEAM Workshop Problem 22 are investigated. Analysis demonstrates that the computation cost of the new methods is less than 1/10 of direct optimization algorithm. The optimization results are also satisfied the design objectives.