CN110990761A - Hydrological model parameter calibration method and device, computer equipment and storage medium - Google Patents
Hydrological model parameter calibration method and device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN110990761A CN110990761A CN201911336959.1A CN201911336959A CN110990761A CN 110990761 A CN110990761 A CN 110990761A CN 201911336959 A CN201911336959 A CN 201911336959A CN 110990761 A CN110990761 A CN 110990761A
- Authority
- CN
- China
- Prior art keywords
- sequence
- sequence set
- parameter
- error value
- hydrological model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
- Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Algebra (AREA)
- Feedback Control In General (AREA)
Abstract
The application relates to a hydrological model parameter calibration method, a hydrological model parameter calibration device, computer equipment and a storage medium, wherein the method comprises the following steps: acquiring a numbering sequence which is randomly and initially generated according to a hydrological model hyper-parameter sequence, constructing a sequence set, acquiring a hyper-parameter value and a hydrological model real input parameter of each numbering sequence in the sequence set, acquiring a predicted flow output by the hydrological model, acquiring an error value corresponding to each numbering sequence according to an error value between the predicted flow and the real output flow, reconstructing the sequence set according to the error value corresponding to each numbering sequence, and taking the reconstructed sequence set as the sequence set again until the latest minimum error value meets a preset condition; and reducing the latest minimum error number sequence into a calibration parameter value to obtain the calibration parameter of the hydrological model. In the whole process, the hydrological model parameters are automatically calibrated, and the calibration process is based on a strict data processing process, so that the accuracy and the high efficiency of parameter calibration are ensured.
Description
Technical Field
The application relates to the technical field of hydrology and water conservancy, in particular to a hydrology model parameter calibration method and device, computer equipment and a storage medium.
Background
Hydrology refers to various phenomena in nature such as water change and movement. Now, the discipline of studying the space-time distribution and change rule of water in nature is commonly referred to. Hydrology is the science of the earth's surface, water in the soil, rock and atmosphere, the science of the occurrence, circulation, content, distribution, physicochemical properties, effects and relationships with all organisms.
In order to efficiently and accurately perform hydrological research, a plurality of hydrological models such as a new anjiang model, a zhangcunyi model, a topmodel and the like are proposed, and although the efficiency and accuracy of hydrological research can be improved based on the hydrological models, the parameter calibration of the hydrological models becomes a difficult problem.
The traditional hydrological parameter calibration generally adopts a manual trial and error method, the manual trial and error method is a parameter calibration method for manually adjusting parameters by comparing the fitting degree of an analog value and an actual value, and the method depends on personal experience too much, has strong subjectivity and poor calibration effect and is not beneficial to popularization and application of models.
Disclosure of Invention
In view of the above, it is necessary to provide an automatic and accurate hydrological model parameter calibration method, apparatus, computer device and storage medium.
A hydrological model parameter calibration method, the method comprising:
acquiring a numbering sequence which is randomly and initially generated according to a hyper-parameter sequence of a hydrological model, and constructing a sequence set;
acquiring a super parameter value, a real input parameter and a real output flow of a hydrological model of each serial number sequence in the sequence set, inputting the super parameter value and the real input parameter into the hydrological model, and acquiring a predicted flow output by the hydrological model;
obtaining an error value corresponding to each number sequence according to an error value between the predicted flow and the real output flow, and recording a minimum error value and a corresponding minimum error number sequence;
reconstructing a sequence set according to the error value corresponding to each numbering sequence, taking the reconstructed sequence set as the sequence set again, and returning to the step of obtaining the over-parameter value of each numbering sequence in the sequence set until the latest minimum error value meets the preset condition;
and reducing the latest minimum error number sequence into a calibration parameter value to obtain the calibration parameter of the hydrological model.
In one embodiment, the obtaining a numbering sequence randomly and initially generated according to a hyper-parameter sequence of a hydrological model, and the constructing a sequence set includes:
acquiring the number of over-parameters, the over-parameter range and the parameter calibration precision of the hydrological model;
generating a hyper-parameter sequence according to the hyper-parameter number, the hyper-parameter range and the parameter calibration precision;
based on the hyper-parametric sequence, randomly generating initially 2nAnd the number sequences form a sequence set, wherein n is parameter calibration precision.
In one embodiment, the reconstructing the sequence set according to the error value corresponding to each number sequence includes:
screening the number sequences according to the error value corresponding to each number sequence to obtain a screening sequence set;
randomly combining the number sequences in the screening sequence set in pairs, respectively and randomly selecting a non-first starting point aiming at the two number sequences combined in a single pair, and mutually exchanging the subsequent number values of the starting points in the two number sequences to obtain an exchange sequence set;
and reconstructing a sequence set according to the exchange sequence set.
In one embodiment, the reconstructing the sequence set according to the exchange sequence set includes:
carrying out duplication removal processing on the exchange sequence set to obtain a duplication-removed exchange sequence set;
and randomly generating a supplementary numbering sequence, and supplementing the supplementary numbering sequence to the de-duplicated exchange sequence set to obtain a reconstructed sequence set, wherein the reconstructed sequence set and the sequence set contain the same number of subsets.
In one embodiment, the filtering the number sequences according to the error value corresponding to each number sequence to obtain a filtered sequence set includes:
acquiring the weight of the error value corresponding to each serial number sequence;
and randomly selecting the number sequence by adopting a roulette algorithm according to the weight to obtain a screening sequence set.
In one embodiment, the reconstructing the sequence set according to the error value corresponding to each numbering sequence, taking the reconstructed sequence set as the sequence set again, and returning to the step of obtaining the hyper-parameter value of each numbering sequence in the sequence set until the latest minimum error value satisfies the preset condition includes:
and reconstructing a sequence set according to the error value corresponding to each numbering sequence, taking the reconstructed sequence set as the sequence set again, and returning to the step of obtaining the super-parameter value of each numbering sequence in the sequence set until the minimum error value obtained in a preset time is returned and is not updated.
In one embodiment, before obtaining an error value corresponding to each number sequence according to an error value between the predicted flow and the real output flow and recording a minimum error value and a corresponding minimum error number sequence, the method further includes:
acquiring a preset error function;
and calculating an error value between the predicted flow and the real output flow through the preset error function.
A hydrological model parameter calibration apparatus, the apparatus comprising:
the sequence set construction module is used for acquiring a numbering sequence which is randomly and initially generated according to the hyper-parameter sequence of the hydrological model and constructing a sequence set;
the calculation module is used for acquiring a super-parameter value, a real input parameter and a real output flow of a hydrological model of each serial number sequence in the sequence set, inputting the super-parameter value and the real input parameter into the hydrological model and acquiring a predicted flow output by the hydrological model;
the recording module is used for obtaining an error value corresponding to each number sequence according to the error value between the predicted flow and the real output flow, and recording a minimum error value and a corresponding minimum error number sequence;
the cyclic updating module is used for reconstructing a sequence set according to the error value corresponding to each serial number sequence, taking the reconstructed sequence set as the sequence set again, and returning to the step of acquiring the super-parameter value of each serial number sequence in the sequence set until the latest minimum error value meets the preset condition;
and the calibration module is used for reducing the latest minimum error serial number into a calibration parameter value to obtain the calibration parameter of the hydrological model.
A computer device comprising a memory storing a computer program and a processor implementing the steps of the method as described above when executing the computer program.
A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method as described above.
The hydrological model parameter calibration method, the device, the computer equipment and the storage medium acquire the number sequence which is randomly and initially generated according to the hydrological model hyper-parameter sequence, construct the sequence set, acquire the hyper-parameter value of each number sequence in the sequence set, the real input parameter and the real output flow of the hydrological model, input the hyper-parameter value and the real input parameter into the hydrological model, and acquire the predicted flow output by the hydrological model, obtaining an error value corresponding to each serial number sequence according to an error value between the predicted flow and the real output flow, reconstructing a sequence set according to the error value corresponding to each serial number sequence, taking the reconstructed sequence set as the sequence set again, and returning to the step of acquiring the super-parameter value of each serial number sequence in the sequence set until the latest minimum error value meets the preset condition; and reducing the latest minimum error number sequence into a calibration parameter value to obtain the calibration parameter of the hydrological model. In the whole process, the hydrological model parameters are automatically calibrated, and the calibration process is based on a strict data processing process, so that the accuracy and the high efficiency of parameter calibration are ensured.
Drawings
FIG. 1 is a diagram of an exemplary embodiment of a method for scaling parameters of a hydrological model;
FIG. 2 is a schematic flow chart illustrating a method for parameter calibration of a hydrological model in one embodiment;
FIG. 3 is a schematic flow chart of a method for calibrating parameters of a hydrological model in another embodiment;
FIG. 4 is a block diagram of an exemplary hydrological model parameter calibration apparatus;
FIG. 5 is a diagram illustrating an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The hydrological model parameter calibration method provided by the application can be applied to the application environment shown in fig. 1. The user operates the terminal 102, sends out a hydrological model parameter calibration operation instruction, the terminal 102 executes the hydrological model parameter calibration method, specifically, a numbering sequence randomly and initially generated according to a hydrological model hyper-parameter sequence is obtained, a sequence set is constructed, a hyper-parameter value of each numbering sequence in the sequence set, a hydrological model real input parameter and a hydrological model real output flow are obtained, the hyper-parameter value and the real input parameter are input into the hydrological model, a prediction flow output by the hydrological model is obtained, obtaining an error value corresponding to each serial number sequence according to an error value between the predicted flow and the real output flow, reconstructing a sequence set according to the error value corresponding to each serial number sequence, taking the reconstructed sequence set as the sequence set again, and returning to the step of acquiring the super-parameter value of each serial number sequence in the sequence set until the latest minimum error value meets the preset condition; and reducing the latest minimum error number sequence into a calibration parameter value to obtain the calibration parameter of the hydrological model. Optionally, the terminal 102 may further store the hydrological model after parameter calibration for further subsequent hydrological data processing. The terminal 102 may be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers, and portable wearable devices.
In one embodiment, as shown in fig. 2, a method for calibrating parameters of a hydrological model is provided, which is described by taking the method as an example applied to the terminal 102 in fig. 1, and includes the following steps:
s100: and acquiring a numbering sequence which is randomly and initially generated according to the hyper-parameter sequence of the hydrological model, and constructing a sequence set.
The number and the range of the hyper-parameters corresponding to each hydrological model are fixed, generally speaking, when it is clear which hydrological model needs to be subjected to parameter rate timing, the number and the range of the hyper-parameters can be directly determined, a hyper-parameter sequence can be generated in advance based on the number and the range of the hyper-parameters, then the hyper-parameter sequence is randomly initialized to obtain a number sequence, and all the number sequences are collected to obtain a sequence set.
S200: and acquiring the super parameter value, the real input parameter and the real output flow of the hydrological model of each serial number sequence in the sequence set, inputting the super parameter value and the real input parameter into the hydrological model, and acquiring the predicted flow output by the hydrological model.
And acquiring the super parameter value of each serial number sequence, the real input parameter and the real output flow of the hydrological model, substituting the real input parameter and the super parameter value into the hydrological model, and acquiring the predicted flow output by the hydrological model, namely performing primary flow prediction processing by introducing the hydrological model with the super parameter value, and acquiring the predicted flow of the hydrological model aiming at the input real input parameter. Specifically, the super-parameter value of each number sequence can be calculated according to the calculation range and the parameter calibration precision of each super-parameter. For example, if the parameter calibration accuracy n is 2 and there are 3 superparameters, their values range from 1 to 2, from 2 to 3, and from 3 to 4, and their number sequences are 10, 11, and 00, respectively, the corresponding superparameters are 1.67, 3, and 3. Furthermore, based on the fact that the number sequences corresponding to the hyper-parameters with the parameter calibration precision of 2 are respectively 00, 01, 10 and 11 in sequence, the first hyper-parameter is taken as an example, 4 values can be taken within the range of 1-2, 1-2 are divided into 4 values in total by taking an average value at equal intervals, 1, 1.33, 1.67 and 2 are taken as an example, and the third value corresponding to the number sequence of 10 is taken, so that the value of the hyper-parameter of the number sequence 10 is 1.67. And substituting the super parameter values and the real input parameters into the hydrological model, and predicting the flow according to the hydrological model after the super parameter values are substituted to obtain the predicted flow output by the hydrological model. Assuming that a hydrological model is y (a (X1+ X2) + b + c, X1 and X2 are true hydrological model real input parameters, a, b and c are the above-determined three excessive values, y is a hydrological model flow predicted value, a, b and c have 4 values in the value range, assuming that X1 is 1 and X2 is 2, and the real flow value is 7, the 4 values corresponding to a, b and c need to be randomly selected so that the final y is as close to 7 as possible. It should be noted that the above is only used to simplify the schematic step S200, and the actual corresponding equations of the hydrological model are much more complicated than the above.
S300: and obtaining an error value corresponding to each number sequence according to the error value between the predicted flow and the real output flow, and recording the minimum error value and the corresponding minimum error number sequence.
And calculating an error value between the predicted flow and the real flow, determining an error value corresponding to each number sequence, namely determining an error value corresponding to each hyper-parameter, and recording a minimum error value and a corresponding minimum error number sequence. Specifically, a preset error function may be obtained first, and the predicted flow rate and the actual flow rate are introduced into the error function to obtain an error value.
S400: and reconstructing the sequence set according to the error value corresponding to each numbering sequence, taking the reconstructed sequence set as the sequence set again, and returning to the step of acquiring the hyper-parameter value of each numbering sequence in the sequence set until the latest minimum error value meets the preset condition.
The preset condition is a preset circulation stop condition, and specifically may be that the number of times of return reaches a certain number of times (preset number of times), or that the minimum error values obtained by a certain number of continuous circulation times are equal. Step S400 is specifically a loop step, where a number sequence with a larger error value is removed according to an error value corresponding to each number sequence, the sequence set is reconstructed, the reconstructed sequence set is used as the sequence set again, and the step of obtaining the hyper-parameter value of each number sequence in the sequence set is returned until the latest minimum error value satisfies the preset condition. Assuming that the number sequences in the currently obtained sequence set are respectively 01, 10, 11 and 00, removing 01 with a larger error value according to the corresponding error value, reconstructing the sequence set to be 10, 11 and 00, returning to the step of obtaining the over-parameter value of each number sequence in the sequence set, and obtaining a new minimum error value until the latest minimum error value is equal to the last corresponding minimum error value, or the number of times of returning is greater than a preset number of times, for example, 10 times, or the latest 5 times of returning corresponding minimum error values are equal (that is, the minimum error value is not updated), and stopping.
S500: and reducing the latest minimum error number sequence into a calibration parameter value to obtain the calibration parameter of the hydrological model.
And reducing the latest minimum error number sequence corresponding to the latest minimum error value into a calibration parameter value to obtain a calibration parameter of the hydrological model, and completing parameter calibration of the hydrological model.
The method for calibrating the parameters of the hydrological model comprises the steps of obtaining a numbering sequence which is randomly and initially generated according to a hydrological model hyper-parameter sequence, constructing a sequence set, obtaining a hyper-parameter value of each numbering sequence in the sequence set, a hydrological model real input parameter and a hydrological model real output flow, inputting the hyper-parameter value and the real input parameter into the hydrological model, obtaining a predicted flow output by the hydrological model, obtaining an error value corresponding to each numbering sequence according to an error value between the predicted flow and the real output flow, reconstructing the sequence set according to the error value corresponding to each numbering sequence, using the reconstructed sequence set as a sequence set again, returning to the step of obtaining the hyper-parameter value of each numbering sequence in the sequence set until the latest minimum error value meets a preset condition; and reducing the latest minimum error number sequence into a calibration parameter value to obtain the calibration parameter of the hydrological model. In the whole process, the hydrological model parameters are automatically calibrated, and the calibration process is based on a strict data processing process, so that the accuracy and the high efficiency of parameter calibration are ensured.
As shown in fig. 3, in one embodiment, step S100 includes:
s120: acquiring the number of over-parameters, the over-parameter range and the parameter calibration precision of the hydrological model;
s140: generating a hyper-parameter sequence according to the hyper-parameter number, the hyper-parameter range and the parameter calibration precision;
s160: based on the hyper-parametric sequence, random initial generation 2nAnd the number sequences form a sequence set, wherein n is parameter calibration precision.
The number and the range of the hyper-parameters corresponding to the hydrological model are fixed in advance, each hydrological model can clearly specify the number and the range of the hyper-parameters when being generated, the parameter calibration precision refers to the requirement of the hydrological model calibration precision at this time, specifically, the data quantity which needs to be processed by different calibration precisions and is input to a terminal by a user is different, generally, the higher the calibration precision is, the higher the data quantity needs to be processed, the higher the hydrological model prediction hydrological data precision after the final parameter calibration is, and vice versa. The precision influences the span of the value of the hyper-parameter, and 2 can be selected from the spannThe individual super parameter values are numbered in order from small to large and are counted in binary. For example, if the value range of one hyper-parameter is 2-3 and the precision is 2, there are 4 hyper-parameters of 2, 2.33, 2.67, and 3, the numbers are 00, 01, 10, and 11, respectively, and it can be seen that the number of bits of the number is n. Typically, the precision is 4 (to 10%) or 7 (to 1%), with a precision of 4 corresponding to 16 copies of data and a precision of at least 10%, and with a precision of 7 corresponding to 128 copies of data and a precision of at least 1%. For example, when the precision is 2, the hydrological model has 3 hyper-parameters, and each hyper-parameter corresponds to 2nOne hyper-parameter value, i.e. 4 hyper-parameter values, the hyper-parameter sequence being 011110, representing the firstThe number of each hyper-parameter is 01, the number of the second hyper-parameter is 11, the number of the third hyper-parameter is 10, and the number of digits of the hyper-parameter sequence is m x n, wherein m is the number of the hyper-parameters.
In one embodiment, reconstructing the sequence set according to the error value corresponding to each number sequence includes: screening the number sequences according to the error value corresponding to each number sequence to obtain a screening sequence set; randomly combining the number sequences in the screening sequence set in pairs, respectively and randomly selecting a non-first starting point aiming at the two number sequences of a single pair of combinations, and mutually exchanging the subsequent number values of the starting points in the two number sequences to obtain an exchange sequence set; and reconstructing the sequence set according to the exchange sequence set.
Screening the number sequences according to the error values corresponding to each number sequence, specifically, preferentially selecting the number sequences with smaller error values to form a screening sequence set again, so that the hydrological model after the parameter calibration can accurately predict hydrological data, combining the number sequences in the screening sequence set in pairs at random, respectively selecting non-first starting points for the two number sequences of each combination, mutually exchanging subsequent numbers of the starting points to obtain an exchange sequence set, and reconstructing the sequence set according to the exchange sequence set. For example, the number sequences in the currently screened sequence set {010101, 101111, 111000} with smaller error values are randomly combined pairwise, a non-first-order starting point is respectively randomly selected for the combination of 010101 and 101111, if the 4 th order is selected as the starting point to exchange the 4 th subsequent number value 101 in 010101 with the 4 th subsequent number value 111 in 101111 to obtain 010111 and 101101, the similar processing procedure is also adopted for other pairwise combinations, which is not described herein again, an exchange sequence set is obtained, and the sequence set is reconstructed on the basis of the exchange sequence set.
In one embodiment, reconstructing the set of sequences from the set of exchange sequences comprises: carrying out duplication removal processing on the exchange sequence set to obtain a duplication-removed exchange sequence set; and randomly generating a supplementary numbering sequence, and supplementing the supplementary numbering sequence to the de-duplicated exchange sequence set to obtain a reconstructed sequence set, wherein the reconstructed sequence set and the sequence set contain the same number of subsets.
After the exchange processing is performed in the previous step, repeated number sequences may exist in the exchange sequence set, and therefore, the exchange sequence set needs to be subjected to deduplication processing to remove the repeated number sequences, the number sequences included in the deduplication exchange sequence set are reduced, and the number sequences need to be supplemented, so that the number of subsets (number sequences) included in the sequence set of each cycle is constant, the supplementary number sequences are randomly generated, and the generated supplementary number sequences are supplemented to the repeated exchange sequence set to obtain a reconstructed sequence set. Specifically, the random generation means that a complementary number sequence is randomly generated in the same manner as the initially generated number sequence, that is, the complementary number sequence is also related to the hyper-parameter, the hyper-parameter value and the calibration accuracy of the hydrological model, and the specific correspondence is referred to the correspondence between the number sequence and the hyper-parameter, the hyper-parameter value and the calibration accuracy of the hydrological model.
In one embodiment, the filtering the number sequences according to the error value corresponding to each number sequence to obtain a filtered sequence set includes:
acquiring the weight of the error value corresponding to each serial number sequence; and randomly selecting the number sequence by adopting a roulette algorithm according to the weight to obtain a screening sequence set.
The weight of the error value may be the difference between the maximum error value and the error value corresponding to the current number sequence, and the weight of the error value corresponding to the maximum error is smaller as the error value is closer to the maximum error, or vice versa. It is understood that after the above processing, based on the weight, the number sequences with smaller error values can be selected to form the selected sequence set. For example, if the error values of the three numbered sequences are 0.1, 0.2, and 0.3, the weight is the maximum error-error, if the error values corresponding to the three numbered sequences are 0.2, 0.1, and 0, respectively, the total weight is 0.3, a random number of 0 to 0.3 is taken, if the sequence 1 is taken in 0 to 0.2, the sequence 2 is taken in 0.2 to 0.3, and the worst sequence with the weight of 0 is not taken, and finally, the numbered sequences with smaller error values are obtained, and the screening sequence set is formed.
In one embodiment, reconstructing the sequence set according to the error value corresponding to each numbering sequence, re-using the reconstructed sequence set as the sequence set, and returning to the step of obtaining the hyper-parameter value of each numbering sequence in the sequence set until the latest minimum error value satisfies the preset condition, includes:
and reconstructing the sequence set according to the error value corresponding to each numbering sequence, taking the reconstructed sequence set as the sequence set again, and returning to the step of acquiring the super-parameter value of each numbering sequence in the sequence set until the minimum error value obtained in the preset time is returned and is not updated.
And if the minimum error values obtained in the preset times are not updated, namely the minimum error values obtained in the preset times are the same, for example, the minimum error values obtained in 10 times of return circulation are all 0.1, stopping circulation, recording the latest minimum error sequence, and reducing the minimum error sequence into a calibration parameter value to perform parameter calibration on the hydrological model.
In one embodiment, obtaining an error value corresponding to each number sequence according to an error value between the predicted flow rate and the real output flow rate, and before recording the minimum error value and the corresponding minimum error number sequence, the method further includes:
acquiring a preset error function; and calculating an error value between the predicted flow and the real output flow through a preset error function.
The preset error function is a preset function for error calculation, the error specifically comprises an absolute error, a relative error and a determination coefficient, namely the three error function formulas respectively correspond to the three error function formulas, the three error function formulas are general hydrologic general formulas, and the specific setting/selection can be according to the needs of an actual scene or the preference of a user.
In order to further explain the technical solution of the hydrological model parameter calibration method and the effect thereof in detail, the steps included in the whole solution will be described below by using specific examples. In one application example, the hydrological model parameter calibration method comprises the following steps:
1. setting the range of the hyper-parameters, generally, the number and the range of the hyper-parameters of each model are fixed, and recording the number of the hyper-parameters as m.
2. Setting an error function, wherein the error function is used for comparing the difference between a predicted value and a true value, and generally comprises three error function formulas of an absolute error, a relative error and a determination coefficient, the formula is a general hydrologic general formula, and the setting is carried out according to the preference of a user.
3. Setting a calibrated precision n, wherein the precision influences the span of the value of the hyper-parameter, and selecting 2 from the spannThe individual super parameter values are numbered in order from small to large and are counted in binary. For example, if the value range of one hyper-parameter is 2-3 and the precision is 2, there are 4 hyper-parameters of 2, 2.33, 2.67, and 3, the numbers are 00, 01, 10, and 11, respectively, and it can be seen that the number of bits of the number is n. Typically the precision is 4 (to 10%) or 7 (to 1%).
4. And setting input data of the hydrological model, such as rainfall and evaporation capacity, and setting real output data, such as flow. Specifically, the input data is also real data, that is, the input data is understood to be real data that is input to the hydrological model in practical application and real hydrological data of the hydrological model test object obtained by other methods, taking the output hydrological data as a flow rate as an example, and the input data is real data such as real rainfall, evaporation capacity and the like of a certain river region corresponding to the hydrological model.
5. The hyper-parameter sequences are formed in a fixed hyper-parameter order. For example, if the precision is 2 and there are three superparameters in total, the total superparameter sequence is 011110, which represents the first superparameter numbered 01, the second numbered 11, and the third numbered 10. The total sequence is seen to have the number of bits m x n.
6. Random initialization generation 2nThe number of sequences, the total number of sequences is denoted as N.
7. And calculating the super parameter value of each serial number sequence, substituting the super parameter value into the hydrological model, calculating the predicted flow of the hydrological model, and substituting the real flow and the predicted flow into an error function to obtain an error value. Finally, each number sequence corresponds to an error value, and the number sequence with the lowest error value and the corresponding error value are recorded separately.
8. The number sequence is randomly selected by a roulette algorithm according to the probability of the error value. This process is a simple process, for example, three sequences with an error of 0.1, 0.2, 0.3, the weight is the maximum error-error, 0.2, 0.1, 0, the sum is 0.3, a random number of 0 to 0.3 is taken, sequence 1 is taken in 0 to 0.2, sequence 2 is taken in 0.2 to 0.3, the worst sequence with a weight of 0 is not taken, and the process is repeated N times to generate a new set of repeated sequences.
9. In the new sequence, two are randomly combined, and a starting point is randomly selected, and values below the starting point are exchanged (the starting point cannot be the first point, but is not exchanged). E.g. two sequences 010101 and 101111, are swapped from the fourth point, resulting in 010111 and 101101.
10. The exchanged sequences are de-duplicated and new sequence complements are randomly generated to maintain a total number N.
11. And jumping to the step 7, and calculating an error value which is a loop. Comparing the lowest error value recorded separately, when the lowest error value in 10 cycles is not updated, the cycle is stopped.
12. And restoring the latest lowest error number sequence when the loop is stopped into a rating parameter value, namely finding the required automatic rating parameter.
It should be understood that although the various steps in the flow charts of fig. 2-3 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2-3 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performance of the sub-steps or stages is not necessarily sequential, but may be performed in turn or alternating with other steps or at least some of the sub-steps or stages of other steps.
As shown in fig. 4, a hydrological model parameter calibration apparatus includes:
the sequence set construction module 100 is used for acquiring a numbering sequence which is randomly and initially generated according to a hyper-parameter sequence of the hydrological model, and constructing a sequence set;
the calculation module 200 is configured to obtain a super-parameter value, a real input parameter and a real output flow of the hydrological model of each serial number in the sequence set, input the super-parameter value and the real input parameter into the hydrological model, and obtain a predicted flow output by the hydrological model;
the recording module 300 is configured to obtain an error value corresponding to each number sequence according to an error value between the predicted flow and the real output flow, and record a minimum error value and a corresponding minimum error number sequence;
the cyclic updating module 400 is configured to reconstruct a sequence set according to the error value corresponding to each numbering sequence, regard the reconstructed sequence set as a sequence set again, and return to the step of obtaining the hyper-parameter value of each numbering sequence in the sequence set until the latest minimum error value meets the preset condition;
and the calibration module 500 is configured to reduce the latest minimum error number sequence to a calibration parameter value to obtain a calibration parameter of the hydrological model.
The hydrological model parameter calibration device acquires a numbering sequence which is randomly and initially generated according to a hydrological model hyper-parameter sequence, constructs a sequence set, acquires a hyper-parameter value of each numbering sequence in the sequence set, a hydrological model real input parameter and a hydrological model real output flow, inputs the hyper-parameter value and the real input parameter into the hydrological model, acquires a predicted flow output by the hydrological model, acquires an error value corresponding to each numbering sequence according to an error value between the predicted flow and the real output flow, reconstructs the sequence set according to the error value corresponding to each numbering sequence, takes the reconstructed sequence set as the sequence set again, and returns to the step of acquiring the hyper-parameter value of each numbering sequence in the sequence set until the latest minimum error value meets a preset condition; and reducing the latest minimum error number sequence into a calibration parameter value to obtain the calibration parameter of the hydrological model. In the whole process, the hydrological model parameters are automatically calibrated, and the calibration process is based on a strict data processing process, so that the accuracy and the high efficiency of parameter calibration are ensured.
In one embodiment, the sequence set building module 100 is further configured to obtain the number of hyper-parameters, the hyper-parameter range, and the parameter calibration precision of the hydrological model; generating a hyper-parameter sequence according to the hyper-parameter number, the hyper-parameter range and the parameter calibration precision; based on the hyper-parametric sequence, random initial generation 2nAnd the number sequences form a sequence set, wherein n is parameter calibration precision.
In one embodiment, the cyclic update module 400 is further configured to filter the number sequences according to the error value corresponding to each number sequence to obtain a filtered sequence set; randomly combining the number sequences in the screening sequence set in pairs, respectively and randomly selecting a non-first starting point aiming at the two number sequences of a single pair of combinations, and mutually exchanging the subsequent number values of the starting points in the two number sequences to obtain an exchange sequence set; and reconstructing the sequence set according to the exchange sequence set.
In one embodiment, the cyclic update module 400 is further configured to perform deduplication processing on the exchange sequence set to obtain a deduplicated exchange sequence set; and randomly generating a supplementary numbering sequence, and supplementing the supplementary numbering sequence to the de-duplicated exchange sequence set to obtain a reconstructed sequence set, wherein the reconstructed sequence set and the sequence set contain the same number of subsets.
In one embodiment, the loop update module 400 is further configured to obtain a weight of a magnitude of an error value corresponding to each number sequence; and randomly selecting the number sequence by adopting a roulette algorithm according to the weight to obtain a screening sequence set.
In one embodiment, the cyclic update module 400 is further configured to reconstruct the sequence set according to the error value corresponding to each numbering sequence, and use the reconstructed sequence set as the sequence set again, and return to the step of obtaining the super-parameter value of each numbering sequence in the sequence set until the minimum error value obtained within the preset time is returned and is not updated.
In one embodiment, the hydrological model parameter calibration device further includes an error value calculation module, configured to obtain a preset error function; and calculating an error value between the predicted flow and the real output flow through a preset error function.
For the specific definition of the hydrographic model parameter calibration device, reference may be made to the above definition of the hydrographic model parameter calibration method, which is not described herein again. The modules in the above-mentioned hydrological model parameter rate device can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a server, the internal structure of which may be as shown in fig. 5. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer equipment is used for storing data such as hydrological models and real hydrological model input and output. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method for parameter calibration of a hydrological model.
Those skilled in the art will appreciate that the architecture shown in fig. 5 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing the following steps when executing the computer program:
acquiring a numbering sequence which is randomly and initially generated according to a hyper-parameter sequence of a hydrological model, and constructing a sequence set;
acquiring a super parameter value, a real input parameter and a real output flow of the hydrological model of each serial number sequence in the sequence set, inputting the super parameter value and the real input parameter into the hydrological model, and acquiring a predicted flow output by the hydrological model;
obtaining an error value corresponding to each number sequence according to an error value between the predicted flow and the real output flow, and recording a minimum error value and a corresponding minimum error number sequence;
reconstructing a sequence set according to the error value corresponding to each serial number sequence, taking the reconstructed sequence set as the sequence set again, and returning to the step of acquiring the super-parameter value of each serial number sequence in the sequence set until the latest minimum error value meets the preset condition;
and reducing the latest minimum error number sequence into a calibration parameter value to obtain the calibration parameter of the hydrological model.
In one embodiment, the processor, when executing the computer program, further performs the steps of:
acquiring the number of over-parameters, the over-parameter range and the parameter calibration precision of the hydrological model; generating a hyper-parameter sequence according to the hyper-parameter number, the hyper-parameter range and the parameter calibration precision; based on the hyper-parametric sequence, random initial generation 2nAnd the number sequences form a sequence set, wherein n is parameter calibration precision.
In one embodiment, the processor, when executing the computer program, further performs the steps of:
screening the number sequences according to the error value corresponding to each number sequence to obtain a screening sequence set; randomly combining the number sequences in the screening sequence set in pairs, respectively and randomly selecting a non-first starting point aiming at the two number sequences of a single pair of combinations, and mutually exchanging the subsequent number values of the starting points in the two number sequences to obtain an exchange sequence set; and reconstructing the sequence set according to the exchange sequence set.
In one embodiment, the processor, when executing the computer program, further performs the steps of:
carrying out duplication removal processing on the exchange sequence set to obtain a duplication-removed exchange sequence set; and randomly generating a supplementary numbering sequence, and supplementing the supplementary numbering sequence to the de-duplicated exchange sequence set to obtain a reconstructed sequence set, wherein the reconstructed sequence set and the sequence set contain the same number of subsets.
In one embodiment, the processor, when executing the computer program, further performs the steps of:
acquiring the weight of the error value corresponding to each serial number sequence; and randomly selecting the number sequence by adopting a roulette algorithm according to the weight to obtain a screening sequence set.
In one embodiment, the processor, when executing the computer program, further performs the steps of:
and reconstructing the sequence set according to the error value corresponding to each numbering sequence, taking the reconstructed sequence set as the sequence set again, and returning to the step of acquiring the super-parameter value of each numbering sequence in the sequence set until the minimum error value obtained in the preset time is returned and is not updated.
In one embodiment, the processor, when executing the computer program, further performs the steps of:
acquiring a preset error function; and calculating an error value between the predicted flow and the real output flow through a preset error function.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring a numbering sequence which is randomly and initially generated according to a hyper-parameter sequence of a hydrological model, and constructing a sequence set;
acquiring a super parameter value, a real input parameter and a real output flow of the hydrological model of each serial number sequence in the sequence set, inputting the super parameter value and the real input parameter into the hydrological model, and acquiring a predicted flow output by the hydrological model;
obtaining an error value corresponding to each number sequence according to an error value between the predicted flow and the real output flow, and recording a minimum error value and a corresponding minimum error number sequence;
reconstructing a sequence set according to the error value corresponding to each serial number sequence, taking the reconstructed sequence set as the sequence set again, and returning to the step of acquiring the super-parameter value of each serial number sequence in the sequence set until the latest minimum error value meets the preset condition;
and reducing the latest minimum error number sequence into a calibration parameter value to obtain the calibration parameter of the hydrological model.
In one embodiment, the computer program when executed by the processor further performs the steps of:
acquiring the number of over-parameters, the over-parameter range and the parameter calibration precision of the hydrological model; generating a hyper-parameter sequence according to the hyper-parameter number, the hyper-parameter range and the parameter calibration precision; based on the hyper-parametric sequence, random initial generation 2nAnd the number sequences form a sequence set, wherein n is parameter calibration precision.
In one embodiment, the computer program when executed by the processor further performs the steps of:
screening the number sequences according to the error value corresponding to each number sequence to obtain a screening sequence set; randomly combining the number sequences in the screening sequence set in pairs, respectively and randomly selecting a non-first starting point aiming at the two number sequences of a single pair of combinations, and mutually exchanging the subsequent number values of the starting points in the two number sequences to obtain an exchange sequence set; and reconstructing the sequence set according to the exchange sequence set.
In one embodiment, the computer program when executed by the processor further performs the steps of:
carrying out duplication removal processing on the exchange sequence set to obtain a duplication-removed exchange sequence set; and randomly generating a supplementary numbering sequence, and supplementing the supplementary numbering sequence to the de-duplicated exchange sequence set to obtain a reconstructed sequence set, wherein the reconstructed sequence set and the sequence set contain the same number of subsets.
In one embodiment, the computer program when executed by the processor further performs the steps of:
acquiring the weight of the error value corresponding to each serial number sequence; and randomly selecting the number sequence by adopting a roulette algorithm according to the weight to obtain a screening sequence set.
In one embodiment, the computer program when executed by the processor further performs the steps of:
and reconstructing the sequence set according to the error value corresponding to each numbering sequence, taking the reconstructed sequence set as the sequence set again, and returning to the step of acquiring the super-parameter value of each numbering sequence in the sequence set until the minimum error value obtained in the preset time is returned and is not updated.
In one embodiment, the computer program when executed by the processor further performs the steps of:
acquiring a preset error function; and calculating an error value between the predicted flow and the real output flow through a preset error function.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A hydrological model parameter calibration method, the method comprising:
acquiring a numbering sequence which is randomly and initially generated according to a hyper-parameter sequence of a hydrological model, and constructing a sequence set;
acquiring a super parameter value, a real input parameter and a real output flow of a hydrological model of each serial number sequence in the sequence set, inputting the super parameter value and the real input parameter into the hydrological model, and acquiring a predicted flow output by the hydrological model;
obtaining an error value corresponding to each number sequence according to an error value between the predicted flow and the real output flow, and recording a minimum error value and a corresponding minimum error number sequence;
reconstructing a sequence set according to the error value corresponding to each numbering sequence, taking the reconstructed sequence set as the sequence set again, and returning to the step of obtaining the over-parameter value of each numbering sequence in the sequence set until the latest minimum error value meets the preset condition;
and reducing the latest minimum error number sequence into a calibration parameter value to obtain the calibration parameter of the hydrological model.
2. The method of claim 1, wherein the obtaining of the number sequences randomly and initially generated according to the hyper-parameter sequences of the hydrological model comprises:
acquiring the number of over-parameters, the over-parameter range and the parameter calibration precision of the hydrological model;
generating a hyper-parameter sequence according to the hyper-parameter number, the hyper-parameter range and the parameter calibration precision;
based on the hyper-parametric sequence, randomly generating initially 2nAnd the number sequences form a sequence set, wherein n is parameter calibration precision.
3. The method of claim 1, wherein reconstructing the sequence set according to the error value corresponding to each numbering sequence comprises:
screening the number sequences according to the error value corresponding to each number sequence to obtain a screening sequence set;
randomly combining the number sequences in the screening sequence set in pairs, respectively and randomly selecting a non-first starting point aiming at the two number sequences combined in a single pair, and mutually exchanging the subsequent number values of the starting points in the two number sequences to obtain an exchange sequence set;
and reconstructing a sequence set according to the exchange sequence set.
4. The method of claim 3, wherein reconstructing the set of sequences from the set of exchange sequences comprises:
carrying out duplication removal processing on the exchange sequence set to obtain a duplication-removed exchange sequence set;
and randomly generating a supplementary numbering sequence, and supplementing the supplementary numbering sequence to the de-duplicated exchange sequence set to obtain a reconstructed sequence set, wherein the reconstructed sequence set and the sequence set contain the same number of subsets.
5. The method of claim 3, wherein the filtering the number sequences according to the error value corresponding to each number sequence to obtain a filtered sequence set comprises:
acquiring the weight of the error value corresponding to each serial number sequence;
and randomly selecting the number sequence by adopting a roulette algorithm according to the weight to obtain a screening sequence set.
6. The method according to claim 1, wherein the reconstructing the sequence set according to the error value corresponding to each numbering sequence, and using the reconstructed sequence set as the sequence set again, and returning to the step of obtaining the hyper-parameter value of each numbering sequence in the sequence set until the latest minimum error value satisfies a preset condition comprises:
and reconstructing a sequence set according to the error value corresponding to each numbering sequence, taking the reconstructed sequence set as the sequence set again, and returning to the step of obtaining the super-parameter value of each numbering sequence in the sequence set until the minimum error value obtained in a preset time is returned and is not updated.
7. The method of claim 1, wherein the obtaining an error value corresponding to each number sequence according to an error value between the predicted flow and the real output flow, and before recording a minimum error value and a corresponding minimum error number sequence, further comprises:
acquiring a preset error function;
and calculating an error value between the predicted flow and the real output flow through the preset error function.
8. A hydrological model parameter calibration apparatus, the apparatus comprising:
the sequence set construction module is used for acquiring a numbering sequence which is randomly and initially generated according to the hyper-parameter sequence of the hydrological model and constructing a sequence set;
the calculation module is used for acquiring a super-parameter value, a real input parameter and a real output flow of a hydrological model of each serial number sequence in the sequence set, inputting the super-parameter value and the real input parameter into the hydrological model and acquiring a predicted flow output by the hydrological model;
the recording module is used for obtaining an error value corresponding to each number sequence according to the error value between the predicted flow and the real output flow, and recording a minimum error value and a corresponding minimum error number sequence;
the cyclic updating module is used for reconstructing a sequence set according to the error value corresponding to each serial number sequence, taking the reconstructed sequence set as the sequence set again, and returning to the step of acquiring the super-parameter value of each serial number sequence in the sequence set until the latest minimum error value meets the preset condition;
and the calibration module is used for reducing the latest minimum error serial number into a calibration parameter value to obtain the calibration parameter of the hydrological model.
9. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911336959.1A CN110990761B (en) | 2019-12-23 | 2019-12-23 | Hydrological model parameter calibration method, hydrological model parameter calibration device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911336959.1A CN110990761B (en) | 2019-12-23 | 2019-12-23 | Hydrological model parameter calibration method, hydrological model parameter calibration device, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110990761A true CN110990761A (en) | 2020-04-10 |
| CN110990761B CN110990761B (en) | 2023-09-08 |
Family
ID=70074660
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911336959.1A Active CN110990761B (en) | 2019-12-23 | 2019-12-23 | Hydrological model parameter calibration method, hydrological model parameter calibration device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110990761B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105389469A (en) * | 2015-11-09 | 2016-03-09 | 中山大学 | Automatic calibration method of storm water management model parameters |
| CN106682355A (en) * | 2017-01-12 | 2017-05-17 | 中国水利水电科学研究院 | Hydrological model parameter calibration method based on PSO (particle swarm optimization)-GA (genetic algorithm) mixed algorithm |
| CN109034507A (en) * | 2018-10-11 | 2018-12-18 | 华自科技股份有限公司 | Hydropower station amount prediction technique and device, computer equipment and storage medium |
| CN109740113A (en) * | 2018-12-03 | 2019-05-10 | 东软集团股份有限公司 | Hyper parameter threshold range determines method, apparatus, storage medium and electronic equipment |
| CN109993372A (en) * | 2019-04-12 | 2019-07-09 | 淮河水利委员会水文局(信息中心) | One kind being based on the probabilistic flood probability forecasting procedure of multi-source |
-
2019
- 2019-12-23 CN CN201911336959.1A patent/CN110990761B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105389469A (en) * | 2015-11-09 | 2016-03-09 | 中山大学 | Automatic calibration method of storm water management model parameters |
| CN106682355A (en) * | 2017-01-12 | 2017-05-17 | 中国水利水电科学研究院 | Hydrological model parameter calibration method based on PSO (particle swarm optimization)-GA (genetic algorithm) mixed algorithm |
| CN109034507A (en) * | 2018-10-11 | 2018-12-18 | 华自科技股份有限公司 | Hydropower station amount prediction technique and device, computer equipment and storage medium |
| CN109740113A (en) * | 2018-12-03 | 2019-05-10 | 东软集团股份有限公司 | Hyper parameter threshold range determines method, apparatus, storage medium and electronic equipment |
| CN109993372A (en) * | 2019-04-12 | 2019-07-09 | 淮河水利委员会水文局(信息中心) | One kind being based on the probabilistic flood probability forecasting procedure of multi-source |
Non-Patent Citations (2)
| Title |
|---|
| DMITRI KAVETSKI 等: "Parameter Estimation and Predictive Uncertainty Quantification in Hydrological Modelling" * |
| 李金城;常学秀;高伟;: "目标函数权重对PEST-HSPF水文率定的影响研究" * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110990761B (en) | 2023-09-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| AU2017202259B2 (en) | Methods and systems for reservoir history matching for improved estimation of reservoir performance | |
| CN108563660B (en) | Service recommendation method, system and server | |
| Zhang et al. | Advances in surrogate modeling for storm surge prediction: storm selection and addressing characteristics related to climate change | |
| JP6791780B2 (en) | Sentence writing device | |
| CN111080397A (en) | Credit evaluation method and device and electronic equipment | |
| CN108171010B (en) | Protein complex detection method and device based on semi-supervised network embedded model | |
| CN113901679B (en) | Reliability analysis method and device for power system and computer equipment | |
| CN111832693B (en) | Neural network layer operation and model training method, device and equipment | |
| CN109767034B (en) | Relay protection constant value optimization method and device, computer equipment and storage medium | |
| CN113110961B (en) | Equipment abnormality detection method and device, computer equipment and readable storage medium | |
| CN110990761A (en) | Hydrological model parameter calibration method and device, computer equipment and storage medium | |
| CN110232437B (en) | CNN-based time series prediction method and model determination method | |
| CN115186810A (en) | Training-free neural network architecture searching method and related equipment | |
| CN111694814B (en) | Batch expansion method and device for date partition table, computer equipment and storage medium | |
| Ayvaz et al. | Identification of the aquifer parameters from pumping test data by using a hybrid optimization approach | |
| CN111722594A (en) | Industrial process monitoring method, device, equipment and readable storage medium | |
| CN113420876A (en) | Real-time operation data processing method, device and equipment based on unsupervised learning | |
| JP2004334480A (en) | Stock price variation value analyzing system | |
| Kwak et al. | Binary NCP: a new approach for solving nonlinear complementarity problems | |
| CN104570757B (en) | Method and device for operating an integrated control assembly | |
| CN116629348B (en) | Intelligent workshop data acquisition and analysis method and device and computer equipment | |
| CN117454668B (en) | Method, device, equipment and medium for predicting failure probability of parts | |
| CN111126650B (en) | Model construction method and device for oil field simulation system | |
| CN114923544B (en) | Method and device for predicting water level of outlet section and electronic equipment | |
| Smith et al. | Simulated annealing of constrained statistical functions |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |