HornSAT Solver Using Agent-Based Modelling in Hopfield Network

Abstract

Horn Satisfiability (HornSAT) problem has been undoubtedly used as a relevant logical rule in much research works of recent, this is not only because it has a special formation in logic programming but also poses interesting interpretations in artificial intelligence. So many attempts have been employed to solve HornSAT and the results are amazingly good, one area that has not seen intense exploration is solving HornSAt using the Agent-based modelling approach, an area we consider very important because of massively parallel computation leverage offered by the ABM. In this research, we implement HornSAT in an Agent-based model using NETLOGO as our platform and embedded horn clause as a logical rule in Hopfield neural network. We investigated the performance of the integration by global minima, local minima, hamming distance, and CPU time. The summary of our findings showed high proficiency in the performance of HornSAT in the Agent-based model, our model provided an additional option to the number of available algorithms that solve HornSAT with an additional user-friendly environment and interface that reduces the bulkiness of the program.