University of Thi-Qar Journal for Engineering Sciences

The recent tendency of research is to hybridize two or more metaheuristics algorithms to find superior solutions in the field of feature selection problems. The Firefly Algorithm (FA) is a population-based optimization algorithm that attempts to replicate the normal behavior of firefly insects in seeking food. FA is broadly used in numerous engineering fields, but it endures from certain restrictions. This study emphasis emphasizes hybridizing FA with the Simulated Annealing algorithm (SA) as a strong localsearch algorithm to overcome FA limits and improve the overall performance in feature selection. In other words, a high-level relay hybrid (HRH) model is proposed in which self-contained optimization (i.e., FA and SA) are implemented in sequence. Obviously, metaheuristic algorithms (like FA) are not suitable for fine adjustment structures that are so near to optimal solutions, whereas local search algorithms (like SA) are the opposite. Accordingly, in the proposed FASA+FS model, the best regions are located by FA and then inputted to SA,respectively.