Load Optimization Distribution Among Cascade Hydropower Stations Based on the A* Algorithm

Under high-intensity peak shaving and frequency regulation, the output of hydropower plants fluctuates frequently, significantly increasing the complexity of load distribution among cascade hydropower plants (LDCHP). Currently, commonly used solution methods include traditional dynamic programming and intelligent optimization algorithms. The former suffers from the "curse of dimensionality" during computation, whereas the latter typically relies on population-based evolutionary mechanisms, often resulting in low convergence efficiency and suboptimal computational performance. To address the LDCHP challenge, this study develops a model for LDCHP and applies the A* algorithm to solve the related dispatching problem. Using cascade hydropower plants in the middle reaches of the Dadu River Basin, China, as a case study, the proposed approach is empirically validated. The results indicate that, compared with the DP algorithm, the A* algorithm achieves an average computational efficiency improvement of 5,372.9 times, alongside a 1.56% decrease in energy storage utilization efficiency. Furthermore, the total energy storage loss calculated by the A* algorithm is 22.41% lower than actual operational data, demonstrating its capacity to significantly enhance solution efficiency without compromising scheduling benefits. Overall, this approach provides valuable technical support for solving complex cascade hydropower dispatching models.