Comparative Study on Application of Multi-Variable Linear Regression and Neural Network in Reservoir Operation

With applications of multi-variable linear regression and improved artificial neural network algorithm, multi-objective reservoir operation functions are established respectively and verified by case study. Through comparison and analyses, it is shown that the artificial neural network algorithm has a better non-linear reflection ability, and can describe the complex relationship between the independent variable and dependent variable with higher precision and better feasibility, and is easier to be used.