Fuzzy logic-based continuous Hopfield network for economic dispatch
K. Okba, Zakaria Bouhanch, Karim El Moutaouakil, Mohamed Benslimane · Mathematical Modeling and Computing · 2025
Continuous Hopfield networks (CHNs) have been extensively employed as neural models for constrained optimization problems due to their parallel computing capabilities and fast convergence properties. Nevertheless, given their reliance on rigid weight and bias parameters, their scalability in dynamic and volatile situations remains limited. To address this limitation, we introduce a CHN based on fuzzy logic (Fuzzy CHN), where fuzzy inference schemes actively tune weights and biases according to real-time feedback. This adaptive setting enhances flexibility, convergence speed, and scalability. As a practical example, we apply the proposed Fuzzy CHN to the economic dispatch (ED) problem in power systems, aimed at reducing production costs while meeting operational constraints. Simulation results demonstrate that the Fuzzy CHN outperforms the classical CHN in terms of solution accuracy, stability, and robustness against system fluctuations. Although the production costs slightly increase, the enhanced efficiency and scalability render the Fuzzy CHN especially beneficial in large-scale, dynamic scenarios. Beyond ED, the Fuzzy CHN approach is highly adaptable to various other constrained optimization problems in industrial engineering and intelligent systems. Moreover, the incorporation of a genetic algorithm (GA) to optimize fuzzy membership parameters further enhances cost minimization and mismatch reduction. The proposed method provides a higher degree of scalability and efficiency than traditional CHNs, delivering improved performance with fewer processing iterations and greater consistency.