A simple and efficient honey bee mating optimization approach to performance characterization of a microwave transistor for the maximum power delivery and required noise

Abstract

In this work, a simple, efficient and multi objective Honey Bee Mating Optimization (HBMO) is presented for the performance characterization of a microwave transistor to deliver maximum power to the load with the required noise F-req. Thus all the possible compatible {FreqFmin, V-out=1, G(Tmax)} triplets and the corresponding source Z(S) and load Z(L)=Z(out)(*) (Z(S)) terminations can be obtained in the device operation domain of (V-DS, I-DS and f) without working analytically for the nonlinear performance and stability equations. HBMO is a recently emerging meta-heuristic algorithm that combines the powers of the simulated annealing and genetic algorithms. Here, a microwave transistor is chosen as a case study, effectiveness and efficiency of the HBMO are shown by comparing its performance to those of the standard meta-heuristics Genetic and Particle Swarm algorithms and the mean cost results for 10 runs are found to be 0.22, 1.65 and 1.85, respectively, for the comparable execution times. Furthermore, all the numerical results are found to agree with their analytical counterparts obtained using the microwave, linear circuit and noise theories. The Feasible Design Target Space FDTS can be built by the cross relations among all the possible compatible {FreqFmin, V-out=1, G(Tmax)} triplets together their terminations {Z(S), Z(L)=Z(out)(*) (Z(S))} covering all the possible amplifier designs where both noise figure and output power are at a premium. Copyright (c) 2015 John Wiley & Sons, Ltd.