Energy Flexibility Optimization in Industry: A Hybrid Approach with Synthetic Data Evaluation

The unpredictability of renewable energy sources, demand-response requirements, and grid constraints pose significant challenges for energy flexibility in industrial sectors. These challenges underscore the role of optimization in enhancing scheduling, reducing costs and minimizing environmental impact. This study proposes a novel AI-driven hybrid optimization for enhancing 48-hour scheduling in photovoltaic (PV) systems with battery storage, aiming to minimize energy costs and system degradation. The proposed algorithm integrates a heuristic search technique based on Depth-First Branch and Bound (DFBB) with Marine Predators Algorithm (MPA) metaheuristics to fine-tune the solutions. To ensure robustness, the methodology is validated using a synthetic dataset representing diverse industrial sectors in Spain, enabling an assessment of its performance under varied operational conditions. A comparative analysis evaluates the hybrid approach against standalone heuristic and metaheuristic methods, as well as other widely used metaheuristic techniques. Results indicate that incorporating heuristic search techniques significantly enhances performance compared to purely metaheuristic approaches in both solution quality and execution time. The hybrid approach consistently provides the best scheduling results across scenarios, balancing quality and computational efficiency. This research underscores the potential of integrating heuristic and metaheuristic techniques to address key operational challenges, enhance decision-making, and improve energy flexibility in industrial settings.