The whale optimization algorithm (WOA) is one of the most powerful swarm-based, nature-inspired metaheuristic algorithms, developed by mimicking the bubble-net hunting maneuver technique of humpback whales to solve complex optimization problems. The WOA has been widely adopted in various real-world optimization fields due to its simple structure, minimal parameter requirements, and fast convergence rate. This article proposes a novel modified version of the original WOA, named the random-flight whale optimization algorithm (RFWOA), which incorporates three types of random-flight mechanisms, i.e., uniform distribution, Rayleigh flights, and Lévy flights. These mechanisms are employed to generate elite solutions and enhance the search performance. The proposed RFWOA helps obtaining a better tradeoff between the exploration and exploitation properties of the original WOA. To evaluate its effectiveness, the RFWOA is tested against ten benchmark functions and applied to solve ten multi-depot multi-vehicle routing problems (MDMVRP) for global optimization. The results obtained by RFWOA are compared with those obtained by other well-known nature-inspired algorithms, including the original WOA, particle swarm optimization (PSO), and genetic algorithm (GA). The experimental results demonstrated that the proposed RFWOA significantly outperformed the competing algorithms in solving both the benchmark functions and the MDMVRP optimization tasks. These results confirm the superiority of the RFWOA over the original WOA, PSO, and GA.
Supaporn Suwannarongsri (Tue,) studied this question.