Modified election algorithm in hopfield neural network for optimal random k satisfiability representation

Abstract

Election algorithm (EA) is a novel metaheuristics optimization model motivated by phenomena of the socio-political mechanism of presidential election conducted in many countries. The capability and robustness EA in finding an optimal solution to optimization has been proven by various researchers. In this paper, modified version of EA has been utilized in accelerating the searching capacity of Hopfield neural network (HNN) learning phase for optimal random-kSAT logical representation (HNN-R2SATEA). The utility of the proposed approach has been contrasted with the current standard exhaustive search algorithm (HNN-R2SATES) and the newly developed algorithm HNN-R2SATICA. From the analysis obtained, it has been clearly shown that the proposed hybrid computational model HNN-R2SATEA outperformed other existing model in terms of global minima ratio (Zm), mean absolute error (MAE), Bayesian information criterion (BIC) and execution time (ET). The finding portrays that the MEA algorithm surpassed the other two algorithms for optimal random-kSAT logical representation.