Abstract
Firefly algorithm (FA) is a new random swarm search optimization algorithm that is modeled after movement and the mutual attraction of flashing fireflies. The number of fitness comparisons and attractions in the FA varies depending on the attraction model. A large number of attractions can induce search oscillations, while a small number of attractions can cause early convergence and a large number of fitness comparisons that can add to the computational time complexity. This study aims to offer H-GA-FA, a hybrid algorithm that combines two metaheuristic algorithms, the genetic algorithm (GA) and the FA, to overcome the flaws of the FA and combine the benefits of both algorithms to solve engineering design problems (EDPs). In this hybrid system, which blends the concepts of GA and FA, individuals are formed in the new generation not only by GA processes but also by FA mechanisms to prevent falling into local optima, introduce sufficient diversity of the solutions, and make equilibrium between exploration/exploitation trends. On the other hand, to deal with the violation of constraints, a chaotic process was utilized to keep the solutions feasible. The proposed hybrid algorithm H-GA-FA is tested by well-known test problems that contain a set of 17 unconstrained multimodal test functions and 7 constrained benchmark problems, where the results have confirmed the superiority of H-GA-FA overoptimization search methods. Finally, the performance of the H-GA-FA is also investigated on many EDPs. Computational results show that the H-GA-FA algorithm is competitive and better than other optimization algorithms that solve EDPs.