An Optimized Hybrid Dragonfly Algorithm Applied for Solving the Optimal Reactive Power Dispatch Problem in Smart Grids

Abstract

World’s growing population has resulted in an up-surged demand for electricity worldwide. The resulting pressure on the power systems has urged the search for measures to increase the performance of electricity distribution systems by minimizing loss, circumventing overload and reducing cost. The implementation of smart grid systems using artificial intelligence, and combinatorial optimization techniques is one of the ways to improve electricity distribution systems. Power grids including smart grids consist of a number of optimal power flow problems, one of which is the Optimal Reactive Power Dispatch (ORPD) problem. It involves determining the optimal configurations of the grid to curtail its cost. The ORPD problem may be solved by means of optimization algorithms including swarm intelligence algorithms. The Dragonfly Algorithm (DA), a high-performing swarm intelligence algorithm, has been successfully used for solving the ORPD problem. However, the performance of DA can still be amplified by overcoming its limitation of having a low exploitation phase. Previously, an optimized DA algorithm with an improved exploitation phase has been proposed. However, it has not been employed to solve the ORPD problem or to enhance the performance of energy distribution systems. In this paper, we propose a new algorithm by further intensifying the exploitation of the optimized DA. This is carried out by utilizing the steepest ascent hill climbing as a local search method instead of the stochastic hill climbing used in the optimized DA algorithm. The newly introduced algorithm is employed to solve the ORPD problem by making use of standard test cases and based on experimental results, it provides higher quality solutions in comparison to the original DA, the optimized DA, and a modified pathfinder algorithm.