Research on Optimal Water Level of Lakes Based on PSO Algorithm

The issue of water resource management in the Great Lakes has become a complex system due to the interaction of many factors. The water level regulation problem and the conflict of interest between stakeholders are outstanding and need to be solved urgently. First, we constructed a network flow model supported by flow balance differential equations. Based on the consideration of stakeholders' interests and constraints, we selected six representative stakeholders and transformed the problem of finding the optimal water level into a multi-objective planning problem. The upper and lower limits of the water level were determined by the historical extremes, and the intermediate value of the water level was determined by the KDE. Then, we calculated the optimal water level using the PSO algorithm. On this basis, a multi-objective planning model for dynamically predicting the optimal water level was established after comprehensively considering the basic model, environmental parameters, and dam control principles. We used ARIMA-LSTM hybrid model to realize the long-term prediction. The multi-objective planning model and trust-constr algorithm are used to realize the dynamic planning and finally the dynamic real-time control of the optimal water level.