A Condition-based Grey Wolf Optimizer Algorithm for Global Optimization Problems

Many real-world optimization problems are complex and high-dimensional problems. In the problems, the search space grows exponentially as the problem dimension increases. Therefore, exact algorithms are not able to find the best solution in a reasonable time. As a result, approximate algorithms are applied to solve these problems. Among these algorithms, meta-heuristic algorithms have been shown a good performance in solving these problems. The Grey Wolf Optimizer (GWO) algorithm is one of the meta-heuristic algorithms. However, the structure of the algorithm limits its exploration capability and it may fall in local optima. In this case, the diversity of the population gradually decreases and sometimes, the algorithm is not able to escape from the local optima. To enhance the performance of GWO, an improved GWO algorithm called Condition-based Gray Wolf Optimization (Cb-GWO) algorithm is proposed in this study. In Cb-GWO, the exploration phase has been separated from the exploitation one and also some mechanisms have been considered to achieve better positions per iteration. Moreover, the balance between exploration and exploitation has been improved. The performance of proposed algorithm has been compared with several improved GWO algorithms, as well as Particle Swarm Optimization (PSO), Spotted Hyena Optimizer (SHO), Harris Hawk Optimization (HHO), Wild Horse Optimizer (WHO), Aquila Optimizer (AO), African Vultures Optimization Algorithm (AVOA), which are among the newest meta-heuristic algorithms. These algorithms have been evaluated by CEC2018 benchmark optimization functions and the pressure vessel design to find the best results. The experimental results showed the significant improvement of efficiency of the proposed algorithm compared with other competitor algorithms.