Enhanced metaheuristic approach in pattern satisfiability problem
Mohd. Asyraf Mansor, Saratha Sathasivam, Mohd Shareduwan Mohd Kasihmuddin · AIP conference proceedings · 2018
Pattern satisfiability (Pattern-SAT) is another variant of the satisfiability problem that inspires researchers to develop model verification based on Boolean algebra. Since Boolean algebra is a solid representation of a classification problem, it was seen to work outstandingly in representing the pattern or image. In this research, enhanced artificial bee colony system (ABC) and modified genetic algorithm (GA) were deployed with the logic programming as a learning tool for pattern satisfiability problem. The aim of this study is to investigate the efficiency and performance of enhanced ABC and GA in doing Pattern-SAT. The comparison of ABC and GA is examined by using Microsoft Visual C++ 2015 and Microsoft Visual Basic C Sharp. The performance of the ABC and GA in doing Pattern-SAT will be appraised in terms of global Pattern-SAT, root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE) and computation time. Hence, the result obtained from computer simulation indicates the good features of ABC compared to GA in doing Pattern-SAT.