DIFFERENTIAL EVOLUTION OPTIMIZATION APPLIED TO THE PERFORMANCE ANALYSIS OF A MICROWAVE TRANSISTOR

Abstract

Differential Evolution (DE) is arguably one of the most powerful stochastic real-parameter optimization algorithms in current use. Although the methodology of the DEA is similar to the genetic algorithm, DEA is simpler and has a better convergence rate than the other counterpart Meta-heuristic optimization algorithms. Herein, DE optimization is applied to determining the Feasible Design Target Space of a microwave transistor for use in Low Noise Amplifier (LNA) designs. Thus, a multiobjective cost function including all the performance measure functions of an LNA transistor which are the transducer gain (G(T)), Noise Figure (F), input and output Voltage Standing Wave Ratio (V-in, V-out) is built to determine the source (Z(S)) and load (Z(L)) terminations to meet the required (F >= F-min, GT, V-in >= 1, V-out >= 1) quadruple within the potential operation bandwidth of the device. A study case is also presented for an LNA transistor NE350184C by applying DE optimization in the determination of their typical performance quadruples together with the source (Z(S)) and load (Z(L)) terminations.