An Evolutionary State-Guided Exponential Distribution Optimizer with Adaptive Search Dynamics for Combined Heat and Power Economic Dispatch Problems

Combined heat and power economic dispatch (CHPED) is a challenging nonlinear, nonconvex optimization problem characterized by the simultaneous coordination of electrical and thermal power generation under complex operational constraints. The presence of valve-point loading effects, transmission power losses, practical operating regions, and the strong coupling between heat and power production significantly increases the complexity of obtaining globally optimal dispatch solutions. To address these challenges, this paper proposes an Evolutionary State-Guided Exponential Distribution Optimizer (ESG-EDO), an enhanced version of the recently developed Exponential Distribution Optimizer (EDO). The proposed ESG-EDO incorporates different search enhancement mechanisms to improve both convergence behavior and solution quality. First, a differential searching strategy is integrated to strengthen exploration and improve the capability of escaping local optima. Second, a perturbation module performs selective variable-wise perturbations instead of full-dimensional updates, thereby maintaining useful search directions while preserving population diversity. Third, an evolutionary state factor is introduced to adaptively regulate the activation of the perturbation mechanism according to the evolutionary status of the population, enabling a dynamic balance between exploration and exploitation throughout the optimization process. Finally, a regeneration mechanism periodically rejuvenates a portion of the population to prevent premature convergence and mitigate stagnation in high-dimensional search spaces. The effectiveness of the proposed ESG-EDO is comprehensively evaluated on benchmark CHPED systems of increasing complexity. The first case study considers a small-scale 7-unit CHPED system under two operating scenarios, namely with transmission losses and without transmission losses, to assess the algorithm’s ability to accurately solve conventional nonlinear dispatch problems. The second case study investigates a large-scale 48-unit CHPED system under the same two scenarios to evaluate scalability and robustness under substantially increased dimensionality and constraint interactions. Finally, the proposed algorithm is validated on a very large-scale 192-unit CHPED system comprising one highly complex operating scenario, demonstrating its capability for solving practical large-dimensional economic dispatch problems. The obtained results consistently demonstrate that the proposed ESG-EDO achieves superior optimization accuracy, faster convergence, enhanced population diversity, and improved robustness across all studied systems. For the very large 192-unit system, ESG-EDO reduced the cost, representing an improvement of approximately 15.1% compared with the original EDO. These findings confirm the effectiveness of the proposed ESG-EDO approach in enhancing the search capability of EDO and achieving superior economic costs reduction.