A modified whale optimization algorithm for large-scale global optimization problems

Abstract

As a new and competitive population-based optimization algorithm, the Whale Optimization Algorithm (WOA) outperforms some other biological-inspired algorithms from the perspective of simplicity and efficiency. However, WOA will get stuck into local optima and degrade accuracy for large-scale global optimization (LSGO) problems. To address the issue, a modified Whale Optimization Algorithm (MWOA) is proposed for solving LSGO problems. In order to balance the exploration and exploitation abilities, a nonlinear dynamic strategy based on a cosine function for updating the control parameter is given. A Lévy-flight strategy is adopted to make the algorithm jump out of local optima. Moreover, a quadratic interpolation method is applied to the leader of the population, which enhances the local exploitation ability and improves the solution accuracy. MWOA is tested over 25 well-known benchmark functions with dimensions ranging from 100 to 1000. The experimental results demonstrate the superior performance of MWOA on LSGO, in terms of solution accuracy, convergence speed, and stability compared with other state-of-the-art optimization algorithms.