Propositional Satisfiability Logic via Ant Colony Optimization in Hopfield Neural Network

Abstract

Minimizing the cost function that corresponds to propositional logic is vital to ensure the learning phase of HNN can occur optimally. In that regard, optimal and non-biased algorithm is required to ensure HNN will always converge to global solution. Ant Colony Optimization (ACO) is a population-based and nature-inspired algorithm to solve various combinatorial optimization problems. ACO simulates the behaviour of the real ants that forage for food and communication of ants through pheromone density. In this work, ACO will be used to minimize the cost function that corresponds to the logical rule in Hopfield Neural Network. ACO will utilize pheromone density to find the optimal path that leads to zero cost function without consuming more learning iteration. Performance for all learning models will be evaluated based on various performance metrics. Results collected from computer simulation implies that ACO outperformed conventional learning model in minimizing the logical cost function.