Parallel Optimization for Economic Load Dispatch in Time-Varying Load Scenarios with Noise Mitigation Using Auto-Encoder

Navigating the complexities of Economic Load Dispatch (ELD) within power systems is notably challenging, particularly in scenarios with variable loads and noise disturbances. Traditional methodologies often fall short, necessitating more robust and innovative strategies. In this paper, we introduce a novel approach aimed at enhancing the reliability and efficiency of the ELD process in the face of these challenges. Our strategy hinges on the integration of an autoencoder model, specifically designed to mitigate the effects of Additive White Gaussian Noise (AWGN) on the variable load demand signal. This enhances the system’s resilience against noise disturbances, ensuring a reliable solution. Furthermore, we implement a Multi-Algorithm Approach, combining the strengths of Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) in a parallel processing framework. This not only accelerates the convergence rate but also optimizes costs more effectively compared to traditional sequential methods. The synergistic sharing mechanism between GA and PSO ensures a seamless and efficient optimization process, showcasing our framework’s capability to handle fluctuating load scenarios with enhanced precision and speed. Our approach marks a significant advancement in the field, redefining the paradigms of power system optimization and setting a new benchmark for performance and resilience in ELD tasks.