swarm optimization

pso algorithm

benchmark functions

simulated annealing

optimization algorithms

differential evolution

harmony search

Knowledge-based systems

In this paper, a novel swarm-based metaheuristic algorithm inspired from the behaviors of beluga whales, called beluga whale optimization (BWO), is presented to solve optimization problem. Three phases of exploration, exploitation and whale fall are established in BWO, corresponding to the behaviors of pair swim, prey, and whale fall, respectively. The balance factor and probability of whale fall in BWO are self-adaptive which play significant roles to control the ability of exploration and exploitation. Besides, the Levy flight is introduced to enhance the global convergence in the exploitation phase. The effectiveness of the proposed BWO is tested using 30 benchmark functions, with qualitative, quantitative and scalability analysis, and the statistical results are compared with 15 other metaheuristic algorithms. According to the results and discussion, BWO is a competitive algorithm in solving unimodal and multimodal optimization problems, and the overall rank of BWO is the first in the scalability analysis of benchmark functions among compared metaheuristic algorithms through the Friedman ranking test. Finally, four engineering problems demonstrate the merits and potential of BWO in solving complex real-world optimization problems. The source code of BWO is currently available to public: https://ww2.mathworks.cn/matlabcentral/fileexchange/112830-beluga-whale-optimization-bwo/ . • A novel metaheuristic algorithm named as Beluga Whale Optimization (BWO) is proposed. • The behaviors of swim, prey and whale fall are designed on BWO algorithm. • The BWO is tested on 30 well-known benchmark functions and 4 engineering problems. • The BWO is compared with 15 well-known metaheuristic algorithms. • The BWO outperforms comparing algorithms in benchmark functions, especially for scalability of dimension.

Beluga whale optimization: A novel nature-inspired metaheuristic algorithm