Xinanjiang model parameter automatic calibration method based on dual-objective optimization genetic algorithm

Abstract

The invention discloses a Xinanjiang model parameter automatic calibration method based on a dual-objective optimization genetic algorithm, relates to the technical field of genetic algorithms and Xinanjiang models, and aims to solve the problems that a Xinanjiang three-water-source model has many parameters, many influence factors and large solution space, and a traditional method cannot well solve the problems. M points are randomly selected from a search space of Xinanjiang model parameters, the m points are coded, and a population is initialized; establishing a target function based on the error index, and obtaining the fitness of population individuals through the target function; carrying out evolution operation on the population through selection, recombination and variation to obtain breeding offspring; and performing reinsertion operation on the obtained breeding offspring to complete automatic calibration of the parameters of the Xinanjiang model, and completing the automatic calibration of the parameters of the Xinanjiang model. Individuals most suitable for the environment are obtained, and flood forecasting and water resource management are carried out.

Description

Technical Field

[0001] The present invention relates to the technical field of genetic algorithms and Xin'anjiang models, in particular to a Xin'anjiang model parameter automatic calibration method based on a dual-objective optimization genetic algorithm. Background Art

[0002] The Xin'an River Model is a distributed model suitable for use in humid and semi-humid regions during the wet season. When the watershed area is small, the Xin'an River Model uses a lumped model, while when the area is larger, a block model is used. However, the Xin'an River Three-Source Model has a large number of parameters, involves numerous influencing factors, and has a large solution space. Traditional methods are not well suited to solving this type of problem.

[0003] A genetic algorithm is a randomized global search optimization method that simulates phenomena such as replication, crossover, and mutation that occur in natural selection and inheritance. Starting from any initial population, it generates a group of individuals that are more suitable for the environment through random selection, crossover, and mutation operations, allowing the group to evolve to better and better areas in the search space. In this way, generation after generation of continuous reproduction and evolution will eventually converge to a group of individuals that are most adapted to the environment, thereby obtaining a high-quality solution to the problem. Summary of the Invention

[0004] The purpose of the present invention is to overcome the deficiencies of the prior art and provide a method for automatically calibrating the parameters of the Xin'anjiang model based on a dual-objective optimization genetic algorithm, comprising the following steps:

[0005] Step S1, randomly select m points from the search space of Xinanjiang model parameters, encode the m points, and initialize the population;

[0006] Step S2, establishing an objective function based on the error index, and obtaining the fitness of the individuals in the population through the objective function;

[0007] Step S3, performing evolution operations on the population through selection, recombination and mutation to obtain breeding offspring;

[0008] Step S4, performs reinsertion operation on the obtained breeding offspring to form the next generation population, and then loops the calculation until the set termination condition is reached, completing the automatic calibration of the Xin'anjiang model parameters. By completing the automatic calibration of the parameters of the Xin'anjiang model, flood forecasting and water resources management are carried out according to the collected data.

[0009] Furthermore, the randomly selecting m points from the search space of the Xin'anjiang model parameters, encoding them, and initializing the population includes: encoding the m points using real-valued encoding.

[0010] Furthermore, the objective function is established based on the error index, including:

[0011] The error indicators include absolute error, relative error and coefficient of certainty;

[0012] The absolute error is the difference between the predicted value of the hydrological element and the measured value of the hydrological element. The average of the absolute values of multiple absolute errors represents the average error level of multiple forecasts.

[0013] The relative error is the ratio of the absolute error to the measured value of the hydrological element, and the average value of multiple relative error absolute values represents the average relative error level of multiple forecasts;

[0014] The coefficient of certainty is the degree of agreement between the flood forecast process and the measured process;

[0015] The objective function is:

[0016]

[0017] Where DC is the deterministic coefficient, y0(i) is the measured value, and y c (i) is the forecast value, is the mean of the measured values, and n is the length of the data sequence.

[0018] Furthermore, the fitness of individuals in the population is obtained through the objective function, including:

[0019] Calculate the fitness of the individuals in the population, that is, calculate the corresponding objective function value, that is, compare the model calculation results of a given flood event calculated by the Xin'anjiang model with the actual observation results; if multiple events and multiple comparison indicators are involved, then find the average value of the comparison results; for the objective function, set different objective functions for calculation, and after setting the objective function, set the criteria for terminating the optimization, which include setting the maximum number of iterations or setting the convergence tolerance of the objective function value; when the difference between the two optimization results is less than the difference threshold, terminate the optimization.

[0020] The beneficial effects of the present invention are: in view of the fact that the traditional Xin'an River three-source model involves many parameters and influencing factors, the genetic algorithm uses phenomena such as replication, crossover and mutation that occur in natural selection and inheritance. Starting from any initial population, through random selection, crossover and mutation operations, it produces a group of individuals that are more suitable for the environment, so that the group evolves to better and better areas in the search space. In this way, it continues to reproduce and evolve from generation to generation, and finally converges to a group of individuals that are most adapted to the environment, thereby solving such problems well and realizing flood forecasting and water resources management based on the collected data. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1This is a flow chart of the automatic calibration method of Xinanjiang model parameters based on dual-objective optimization genetic algorithm;

[0022] Figure 2 Schematic diagram of the implementation of the automatic calibration method of the Xin'anjiang model parameters based on the dual-objective optimization genetic algorithm. DETAILED DESCRIPTION

[0023] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings, but the protection scope of the present invention is not limited to the following.

[0024] The features and performance of the present invention are further described in detail below with reference to the embodiments.

[0025] like Figure 1 As shown in FIG, the automatic calibration method of the Xinanjiang model parameters based on the dual-objective optimization genetic algorithm includes the following steps:

[0026] Step S1, randomly select m points from the search space of Xinanjiang model parameters, encode the m points, and initialize the population;

[0027] Step S2, establishing an objective function based on the error index, and obtaining the fitness of the individuals in the population through the objective function;

[0028] Step S3, performing evolution operations on the population through selection, recombination and mutation to obtain breeding offspring;

[0029] Step S4, performs reinsertion operation on the obtained breeding offspring to form the next generation population, and then loops the calculation until the set termination condition is reached, completing the automatic calibration of the Xin'anjiang model parameters. By completing the automatic calibration of the parameters of the Xin'anjiang model, flood forecasting and water resources management are carried out according to the collected data.

[0030] The method of randomly selecting m points from the search space of the Xin'anjiang model parameters, encoding them, and initializing the population includes: encoding the m points using real-valued encoding.

[0031] The objective function is established based on the error index, including:

[0032] The error indicators include absolute error, relative error and coefficient of certainty;

[0033] The absolute error is the difference between the predicted value of the hydrological element and the measured value of the hydrological element. The average of the absolute values of multiple absolute errors represents the average error level of multiple forecasts.

[0034] The relative error is the ratio of the absolute error to the measured value of the hydrological element, and the average value of multiple relative error absolute values represents the average relative error level of multiple forecasts;

[0035] The coefficient of certainty is the degree of agreement between the flood forecast process and the measured process;

[0036] The objective function is:

[0037]

[0038] Where DC is the deterministic coefficient, y0(i) is the measured value, and y c (i) is the forecast value, is the mean of the measured values, and n is the length of the data sequence.

[0039] The fitness of individuals in the population is obtained through the objective function, including:

[0040] Calculate the fitness of the individuals in the population, that is, calculate the corresponding objective function value, that is, compare the model calculation results of a given flood event calculated by the Xin'anjiang model with the actual observation results; if multiple events and multiple comparison indicators are involved, then find the average value of the comparison results; for the objective function, set different objective functions for calculation, and after setting the objective function, set the criteria for terminating the optimization, which include setting the maximum number of iterations or setting the convergence tolerance of the objective function value; when the difference between the two optimization results is less than the difference threshold, terminate the optimization.

[0041] Specifically, Example 1: See Figure 2 The present invention provides a method for automatically calibrating parameters of the Xin'anjiang model based on a dual-objective optimization genetic algorithm, which specifically includes the following steps:

[0042] Step S1: Initialize the population, randomly select m points from the parameter search space, and then encode them to initialize the population. There are many encoding methods. Since the model is all real number operations, a real-valued encoding method can be used;

[0043] Step S2: Calculate the fitness, calculate the fitness of the individuals in the evaluation population, that is, calculate the corresponding objective function value. In the model, the calculation result of the model calculation of a given flood is compared with the actual observation result, involving multiple floods and multiple comparison indicators. For multiple floods, the average is calculated. For the objective function, different objective functions are set for calculation. After setting the objective function, it is necessary to set the criteria for terminating the optimization. There are many cases where the maximum number of iterations is set. The convergence tolerance of the objective function value can also be set, that is, the operation is terminated when the difference between the two optimization results is very small. The parameter iteration step size convergence tolerance can also be set;

[0044] Step S3: Evolutionary operations, including selection, recombination, and mutation. These operations can take many forms, and commonly used methods are generally selected, such as the NSGAII algorithm commonly used in multi-objective genetic algorithms.

[0045] Step S4: A new population is formed. After the evolution operation is completed, breeding offspring are obtained. In order to maintain the population size, a reinsertion operation can be performed. For example, the NSGAII algorithm has an elite retention operation, so the framework can be used to retain these new individuals to form the next generation population. The calculation is then cyclically calculated until the previously set termination condition is reached, that is, the optimization is completed.

[0046] Example 2: See Figure 2 , further, step S2, setting the objective function, specifically includes the following steps:

[0047] Step S21: selecting absolute error as the error index, that is, the predicted value of the hydrological element minus the measured value, and the average of multiple absolute error values represents the average error level of multiple forecasts;

[0048] Step S22: selecting relative error as the error index, that is, the absolute error divided by the measured value, expressed as a percentage, and the average of multiple relative error absolute values represents the average relative error level of multiple forecasts;

[0049] Step S23: Selecting the coefficient of certainty as the error index, the degree of agreement between the flood forecast process and the measured process can be measured using the coefficient of certainty as an index;

[0050] Step S24: Determine the objective function.

[0051] The foregoing description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein and should not be construed as excluding other embodiments. Rather, the present invention can be used in various other combinations, modifications, and environments and can be modified within the scope of the concept described herein through the above teachings or techniques or knowledge in the relevant field. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention are intended to be protected by the appended claims.

Claims

1. The automatic calibration method of Xinanjiang model parameters based on dual-objective optimization genetic algorithm is characterized by: The steps include: Step S1, randomly select m points from the search space of Xinanjiang model parameters, encode the m points, and initialize the population; Step S2, establishing an objective function based on the error index, and obtaining the fitness of the individuals in the population through the objective function; Step S3, performing evolution operations on the population through selection, recombination and mutation to obtain breeding offspring; Step S4, performs reinsertion operation on the obtained breeding offspring to form the next generation population, and then loops the calculation until the set termination condition is reached, completing the automatic calibration of the Xin'anjiang model parameters. By completing the automatic calibration of the parameters of the Xin'anjiang model, flood forecasting and water resources management are carried out according to the collected data.

2. The Xinanjiang model parameter automatic calibration method based on the dual-objective optimization genetic algorithm according to claim 1 is characterized in that: The method of randomly selecting m points from the search space of the Xin'anjiang model parameters, encoding them, and initializing the population includes: encoding the m points using real-valued encoding.

3. The Xinanjiang model parameter automatic calibration method based on the dual-objective optimization genetic algorithm according to claim 2 is characterized in that: The objective function is established based on the error index, including: The error indicators include absolute error, relative error and coefficient of certainty; The absolute error is the difference between the predicted value of the hydrological element and the measured value of the hydrological element. The average of the absolute values of multiple absolute errors represents the average error level of multiple forecasts. The relative error is the ratio of the absolute error to the measured value of the hydrological element, and the average value of multiple relative error absolute values represents the average relative error level of multiple forecasts; The coefficient of certainty is the degree of agreement between the flood forecast process and the measured process; The objective function is: Where DC is the deterministic coefficient, y0(i) is the measured value, and y c (i) is the forecast value, is the mean of the measured values, and n is the length of the data sequence.

4. The Xinanjiang model parameter automatic calibration method based on the dual-objective optimization genetic algorithm according to claim 3 is characterized in that: The fitness of individuals in the population is obtained through the objective function, including: Calculate the fitness of the individuals in the population, that is, calculate the corresponding objective function value, that is, compare the model calculation results of a given flood event calculated by the Xin'anjiang model with the actual observation results; if multiple events and multiple comparison indicators are involved, then find the average value of the comparison results; for the objective function, set different objective functions for calculation, and after setting the objective function, set the criteria for terminating the optimization, which include setting the maximum number of iterations or setting the convergence tolerance of the objective function value; when the difference between the two optimization results is less than the difference threshold, terminate the optimization.

Images