Optimization Of HBV Hydrological Model Parameters Based On Improved Genetic Algorithm

Genetic algorithm is a method of searching for optimal solutions based on Darwin's biological evolution theory.Genetic algorithms have advantages over traditional methods,but also have drawbacks.For the optimization process of complex problems,the standard genetic algorithm has the problems of being trapped in local optimal solution,low search efficiency and instability,the main research work of this paper is as follows:First,this paper redefines the concept of chromosomes and individuals in genetic algorithms.Based on this,a genetic algorithm that relates chromosomes to individual probabilities is proposed and tested using common test functions.The results of the improved genetic algorithm are compared with the results of standard genetic algorithm and adaptive genetic algorithm.The analysis shows that the improved genetic algorithm has better robustness and faster convergence speed.Secondly,the improved genetic algorithm proposed in this paper is used for parameter calibration of HBV hydrological models.In the calibration,the sum of the absolute values of the error of the simulated flow sequence and the recorded flow sequence is used as the criterion function to perform the optimization calculation.The calculation results are compared with the optimization results of the standard genetic algorithm and the adaptive genetic algorithm.The results show that the simulation results can reflect the measured flow sequence better,and the improved genetic algorithm performs better than the comparison algorithm.Thirdly,aiming at the defect that the traditional criterion function is not adaptable in high water period and low water period,a new criterion function model is proposed to strengthen the description of these runoff features in high water period and low water period.Finally,the uncertainty of the parameters found in the parameter calibration process is analyzed.The results show that the upper and lower boundaries of the flow process line obtained by the simulation cannot completely include the measured flow process lines.Some measured flow values are not within the 90% confidence interval.It shows that the HBV hydrological model is not complete enough to simulate the flow of the basin.The improvement of the genetic algorithm in this paper makes it closer to the nature of biological genetics and enriches the theory of genetic algorithms.The successful application of the improved algorithm in the parameter determination of the HBV hydrological model satisfies the demand for river flow forecasting in actual production practice.