Performance Enhancement of MOSFET Circuits Using Evolutionary and Nature-Inspired Optimization Algorithms

Abstract

A comparative study for improving performance of MOSFET circuits through employing Particle Swarm Optimization (PSO) algorithm, Genetic Algorithm (GA), and Secretary Bird Optimization Algorithm (SBOA) is presented in this paper. The mission of optimization is to be formulated as a constrained maximization problem in which improving the voltage gain through varying the circuit design variables, particularly the transistor width-to-length ratio (W/L), drain resistance (RD), and load resistance (RL). The suggested algorithms are executed with the same number of iterations and independent runs to achieve suitable comparison. The results show that the highest best gain of (1378.125) is achieved by PSO with extremely fast convergence and zero variation across the recorded runs, which indicates strong exploitation and high repeatability. A gradual improvement behavior is produced by GA and can be reached a gain of (1176.766589), which reflects evolutionary diversity with slower convergence. While SBOA is reached a best gain of (1226.267555); but its convergence curve still almost flat, witch indicates early stagnation under the suited parameter configuration. The analysis demonstrates PSO has strong effective method for the present MOSFET gain-optimization case, while GA remains useful for diversity-preserving search and SBOA requires further adaptation parameter or hybridization to improve exploration.