Hybrid metaheuristic optimization based DSM approach towards effective energy recommender system

Abstract

The demand for electricity is rapidly increasing, along with its cost. To address consumer needs and alter electricity usage behavior, Demand Side Management (DSM) techniques, focusing on cost savings, are widely used globally. Various metaheuristic-based optimization algorithms, such as Genetic Algorithm (GA), Grey Wolf Optimization (GWO), Sperm Swarm Optimization (SSO), and Particle Swarm Optimization (PSO), have been applied in load-shifting strategies within DSM. This paper proposes a hybrid model, GAPSO, combining GA's global search with PSO's local search to optimize load shifting and curtailing, specifically for peak load reduction. The results show that GAPSO outperforms both GA and PSO, achieving a globally optimal solution and avoiding local minima. For peak load reduction, GAPSO achieves up to 2.26 % better results for aggregated Grid load and 0.8 % for AC load compared to the closest competitor. In terms of cost savings, GAPSO saves up to 155.60 cents for the Grid, while PSO leads in cost savings for AC load, outperforming both GA and GAPSO by 79.75 cents. Additionally, a GUI-based recommender system for air conditioning load is developed as a case study. The PSO-driven system helps reduce energy costs by providing recommendations on operating temperatures and time slots for AC load. The system operates in two modes: load curtailment when the AC is on and load shifting for a full 24 h period, presenting users with potential cost savings based on their choices.