Modified WFO: A Novel WFO Variant for Global Optimization Problems

Water Flow Optimization Algorithm (WFO) has been developed based on the observation of water flow patterns in nature. It utilizes a mathematical model that incorporates the laminar and turbulent hydraulic phenomena of water particles as they move from higher to lower ground. This algorithm aims to solve global optimization problems. Despite its strong competitiveness compared to other optimization methods, the WFO still encounters challenges related to escaping local extremum and maintaining population diversity. This article presents a modified water flow optimization algorithm that enhances the existing laminar flow operator by incorporating a double random particle influence factor. This factor is combined with the original common movement direction using normalized random coefficients. Consequently, the proposed algorithm introduces changes in the water particle position influenced by both the global optimal particle and three other water particles. As a result, the algorithm exhibits increased population diversity and improved capability to escape local minima. The algorithm underwent validation by employing the CEC2022 real parameter single objective optimization test set, along with an engineering design example, and was subsequently compared to the original WFO algorithm as well as five other recently published algorithms. The experimental findings indicated that the modified WFO algorithm exhibits competitiveness.