Optimal design of high-power microwave source based on particle simulation and genetic algorithms

Abstract

Optimal design method of high-power microwave source based on particle simulation and parallel genetic algorithm is presented in this paper. The output power of the high-power microwave device, simulated by the fully electromagnetic particle simulation code UNIPIC, is given as a fitness function, and the float-encoding genetic algorithm is used to optimize the high-power microwave device. Using this method, we encode the position and height of the Bragg reflector of the relativistic backward wave oscillator (RBWO), and optimize the parameters for a massively parallel processor. Simulation results demonstrate that we can obtain the global optimal parameters of the Bragg reflector of the RBWO.