A modified differential search algorithm with gradient-based repair is developed to obtain the solution to the optimal power flow problem with renewable energy resources, such as wind energy, photovoltaic energy, combined wind–tidal systems, and plug-in electric vehicles. Implemented modifications of the proposed algorithm include a random- as well as Lévy step-based hybrid population initialization technique for enhanced exploration capability and improved convergence speed, a self-adaptive affinity index-based global-best-guided mechanism to achieve a balanced exploration–exploitation trade-off, and a \(\beta\) -hill climbing technique to strengthen the exploitation phase. An efficient constraint-handling method, referred to as gradient-based repair, is integrated with the proposed evolutionary optimization algorithm to address explicit system constraints. Exhaustive test studies are carried out on two test systems considering both single- and multi-objective functions, and the obtained results are compared with those of seven sophisticated algorithms. The results indicate that the proposed algorithm outperforms the compared optimization techniques in terms of solution quality and convergence speed. Furthermore, the findings confirm that the proposed approach yields superior performance in terms of solution quality, accuracy, and convergence speed for both small-scale and large-scale test systems when compared with the seven sophisticated algorithms. The high dimensionality, strong nonlinearities, and probabilistic uncertainty modeling associated with the large-scale test system impose significant computational demands, thereby demonstrating that the proposed approach is well suited for high-performance computing platforms and parallel or distributed implementations.
