Enhanced Economic Dispatch Optimization Using Momentum in the Orca Predation Algorithm

Abstract

The economic dispatch problem aims to minimize total generation costs while maintaining a balanced power allocation in the electrical power system. This paper introduces a new algorithm called the Self-Adaptive Momentum Orca Predation Algorithm (SAMOPA), the first variant of OPA that integrates an adaptive momentum mechanism to solve the economic dispatch problem. This integration enhances the balance between exploration and exploitation, accelerates the convergence process, and reduces the risk of getting stuck in local solutions. SAMOPA was tested on three economic dispatch scenarios with 6, 15, and 40 power plants. In the 6-unit system, SAMOPA achieved the best cost of 15275.9304 dollars with a standard deviation of 2.54E-06, lower than OPA's deviation of 0.0029. SAMOPA's computation time was also faster, at 0.27 seconds compared to OPA's 0.9751 seconds. For the 15-unit system, SAMOPA reduced total generation costs by 0.17 percent compared to OPA. For a 40-unit system, SAMOPA achieves the best cost of 119,369.57 dollars, which is 0.58 percent lower than the best result from OPA. The consistency of SAMOPA's performance is also demonstrated in the Economic Dispatch case with nonlinear valve-point effects and in the CEC 2020 benchmark function, which includes unimodal, hybrid, and composition categories. Statistical validation using the Wilcoxon signed-rank test indicates that SAMOPA exhibits significantly better convergence behavior and solution consistency compared to the comparison method. These results confirm that SAMOPA is an effective and reliable new contribution to solving complex and large-scale economic dispatch problems.