Hydrological ensemble interval forecast building method based on multi-model random linear combination

The invention discloses a hydrological ensemble interval forecasting construction method based on multi-model random linear combination, and the method comprises the following steps: constructing different types of hydrological forecasting models, determining forecasting objects and forecasting factors of each forecasting model, calibrating parameters of each forecasting model, and carrying out precision evaluation on each forecasting model; eliminating the forecasting models with obviously low forecasting effects, linearly combining the remaining forecasting models through random weights to obtain an ensemble forecasting model, and constructing a multi-objective optimization mathematical model; and finally, optimizing the upper limit and the lower limit of the random weight by adopting amulti-objective optimization algorithm to obtain a Pareto solution set, and screening the Pareto solution set by adopting a multi-attribute decision-making method to obtain an equilibrium solution; and generating an ensemble