An improved multi-objective learning automata and its application in VLSI circuit design

In this paper, an improved multi-objective optimization method, based on learning automata (called IMOLA), is proposed and its performance on the design of a variety of functional circuits is investigated. The most important feature of the proposed method is to provide a suitable schedule for effective compromise between exploration and exploitation during the search process. To evaluate the capability of the proposed method on multi-objective problems, digital and analog circuits have been selected. The results show the superiority in comparison with new and common algorithms called non-dominated sorting genetic algorithm III, multi-objective multi verse optimization, adaptive multi-objective black hole algorithm, multi-objective modified inclined planes system optimization, and multi-objective grasshopper optimization algorithm. Evaluation of the results was reported in terms of power-delay-product, power-area-product, success rate, Pareto-front, multi-objective criteria, circuit variables, design constraints, runtime, and performance analysis.