Quasi-Gradient Nonlinear Simplex Optimization Method in Electromagnetics

Particle swarm optimization (PSO), genetic algorithm (GA), and nonlinear simplex method (SOM) are some of the most prominent gradient-free algorithms in engineering. When it comes to a common group electromagnetic problems wherein less than 10 parameters present problem domain, SOM features faster convergence rate vs PSO GA. Nevertheless, GA still outperform by having more accuracy finding global minimum. To improve with few parameters, quasi-gradient (Q-G) search direction is added conventional algorithm. An extra decision made proposed move alongside reflection or during error-reduction operations. This modification will SOM, which otherwise fails examples presented this article, levels similar GA, while retaining approximately 33% speed relatively small number 20% larger parameters. Following standard benchmark test verification, successfully solves suite problems. Representative include absorber dimensions an anechoic chamber, estimation properties unknown embedded object scattered microwave signals.