Application of Independent-minded Particle Swarm Optimization for design of class-E amplifiers

摘要

The class-E amplifier is one of the switching amplifiers, which satisfies the class-E switching conditions. It is, however, difficult to determine the values of the passive elements included in the circuit for achieving the class-E switching conditions. Recently, the Newton type algorithm is proposed to determine the passive elements. However, this method requires a good initial estimation. In this paper, an algorithm using Independent-minded Particle Swarm Optimization (IPSO) is introduced to estimate the initial conditions. To find it efficiently, the MOSFET in the class-E amplifier is replaced with an ideal switch and the objective function for the optimization is efficiently evaluated. Unfortunately, the objective function has multimodal characteristics and a robust optimization method is required. Then, the optimum solution is found by using IPSO which shows good performances for multimodal cases. Therefore, the initial estimation for the Newton type algorithm is easily obtained and a good design of the class-E amplifier becomes available.