A method, system, and apparatus for optimizing simulation accuracy of a regional-scale process model

Abstract

The application discloses a regional scale process model simulation precision optimization method, system and device, relates to the field of model optimization, and comprises the following steps: inputting a crop parameter combination to be optimized into a crop-water productivity model, determining result variables yield and canopy coverage; performing cross-validation on the result variables yield and canopy coverage to determine evaluation indexes; and constructing a multi-objective index evaluation optimization model for optimization of key parameters of a regional scale PCMs; wherein the multi-objective index evaluation optimization model takes the sum of Nash efficiency coefficients NSE of the canopy coverage and the result variable yield in the training set and the test set as an objective function, and takes a root mean square error RMSE as a constraint condition; the evaluation indexes are input into the multi-objective index evaluation optimization model, a parallel algorithm group composed of multiple meta-heuristic optimization methods is adopted to solve the multi-objective index evaluation optimization model, and the key parameters of the regional scale PCMs are optimized to obtain an optimal crop parameter combination.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of model optimization, in particular to a regional scale process model simulation precision optimization method, system and device. BACKGROUND

[0002] Process-based crop models (PCMs) are important tools for modern agricultural science research and resource management. At present, the commonly used PCMs include three types of soil factors (hydrodynamics and water balance models), photosynthesis factors (CO2 and radiation driving) and human factors driving based on the division of main driving factors.

[0003] From the perspective of model calibration, the parameter values cannot be directly measured, are recommended in the form of ranges and have a large number of possible combinations, which makes it difficult for artificial calibration models to obtain the optimal parameter combination, and thus it is difficult to effectively improve the simulation precision of regional scale process models PCMs. SUMMARY

[0004] The present application provides a regional scale process model simulation precision optimization method, system and device to solve the above problems existing in the prior art, i.e. how to improve the simulation precision of regional scale process models PCMs in the prior art. The present application provides a regional scale process model simulation precision optimization method, which comprises the following steps:

[0005] Inputting a crop parameter combination to be optimized into a crop-water productivity model AquaCrop-OSPy to determine simulated crop yield and canopy coverage; wherein the crop parameter combination comprises product data, measured data and calibration data;

[0006] Performing cross-validation on the crop yield and canopy coverage by using a leave-one-out cross-validation method (LOOCV) to determine the evaluation index corresponding to the crop parameter combination; wherein the evaluation index comprises a Nash efficiency coefficient NSE and a root mean square error RMSE;

[0007] Constructing a multi-objective index evaluation optimization model for optimization of key parameters of regional scale PCMs;

[0008] Taking the sum of the Nash efficiency coefficients NSE of the canopy coverage and the crop yield in the training set and the test set as an objective function, and taking the root mean square errors RMSE of the canopy coverage and the crop yield in the training set and the test set as constraint conditions, training the multi-objective index evaluation optimization model;

[0009] The evaluation index is input into the trained multi-objective index evaluation optimization model, a parallel algorithm group composed of multiple meta-heuristic optimization methods is used to solve the multi-objective index evaluation optimization model, the crop parameter combination is optimized, and the optimal crop parameter combination is obtained.

[0010] Optionally, the parallel algorithm group composed of multiple meta-heuristic optimization methods is used to solve the multi-objective index evaluation optimization model, and specifically includes the following steps.

[0011] The parallel algorithm including UNSGA3, MOEA / D, SPEA2, AGEMOEA, CTAEA and SMSEMOA is used to solve the multi-objective index evaluation optimization model.

[0012] Optionally, the product data includes soil hydraulic parameters, soil texture and meteorological data.

[0013] Optionally, the measured data includes yield, leaf area index LAI, irrigation data and field management data.

[0014] Optionally, the calibration data includes a research area, a research object, key crop parameters, a search interval and a search step length.

[0015] The application provides a regional scale process model simulation precision optimization system, which comprises:

[0016] An acquisition module is configured to input a crop parameter combination to be optimized into an AquaCrop-OSPy crop-water productivity model, and determine simulated crop yield and canopy coverage.

[0017] A cross-validation module is configured to perform cross-validation on the crop yield and the canopy coverage by using a leave-one-out cross-validation method (LOOCV), and determine evaluation indexes corresponding to the crop parameter combination.

[0018] A construction module is configured to construct a multi-objective index evaluation optimization model for optimization of key parameters of regional scale PCMs.

[0019] A training module is configured to minimize the sum of Nash efficiency coefficients (NSE) of the canopy coverage and the crop yield in the training set and the test set as an objective function, and use the root mean square errors (RMSE) of the canopy coverage and the crop yield in the training set and the test set as constraint conditions to train the multi-objective index evaluation optimization model.

[0020] An optimization module is configured to input the evaluation indexes into the trained multi-objective index evaluation optimization model, solve the multi-objective index evaluation optimization model by using a parallel algorithm group composed of multiple meta-heuristic optimization methods, optimize the crop parameter combination, and obtain the optimal crop parameter combination.

[0021] The application provides a computer device, which comprises a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor implements the regional scale process model simulation precision optimization method when executing the program.

[0022] Compared with the prior art, the application has the following beneficial effects: the application provides a regional scale process model simulation precision optimization method, which combines a meta-heuristic optimization method with LOOCV, so that the key crop parameter combination of PCMs is optimized under limited measured data; the evaluation indexes obtained by the LOOCV process are used to construct a multi-objective index evaluation optimization model, the sum of the Nash efficiency coefficients NSE of the canopy coverage and the result variable yield in the training set and the test set is taken as an objective function, and the root mean square error RMSE is taken as a constraint condition, the multi-objective index evaluation optimization model is solved by using a parallel algorithm group, and the optimal crop parameter combination can be obtained, so that the simulation precision and the calibration efficiency of PCMs on the regional scale are effectively improved. BRIEF DESCRIPTION OF DRAWINGS

[0023] The accompanying drawings, which are incorporated into and form part of the specification, illustrate embodiments consistent with the application and, together with the specification, serve to explain the principles of the application.

[0024] Figure 1 A flowchart of a regional scale process model simulation precision optimization method provided by an embodiment of the application;

[0025] Figure 2 A technical roadmap of a regional scale process model simulation precision optimization method provided by an embodiment of the application;

[0026] Figure 3 A leave-one-out cross-validation method schematic diagram provided by an embodiment of the application;

[0027] Figure 4 A computer device schematic diagram of a regional scale process model simulation precision optimization method provided by an embodiment of the application. DETAILED DESCRIPTION

[0028] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below in combination with the drawings in the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0029] The technical solutions of the present application and how the technical solutions solve the above technical problems will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes can not be described again in some embodiments. The embodiments of the present application will be described below in combination with the drawings.

[0030] As shown in Figure 1 and Figure 2 , the area scale process model simulation precision optimization method shown in the embodiment comprises:

[0031] S1: inputting a crop parameter combination to be optimized into a crop-water productivity model AquaCrop-OSPy to determine simulated crop yield and canopy coverage; wherein the crop parameter combination comprises product data, measured data and calibration data.

[0032] Exemplarily, representative spatiotemporal difference product data, measured data and calibration data can be retrieved from multiple academic databases.

[0033] Optionally, the product data comprises soil hydraulic parameters, soil texture and meteorological data, the measured data comprises yield, LAI and field management data, and the calibration data comprises research area, research object, key crop parameter, search interval and search step.

[0034] S2: cross-validated crop yield and canopy coverage by using a leave-one-out cross-validation method LOOCV to determine the evaluation index corresponding to the crop parameter combination; wherein the evaluation index comprises a Nash efficiency coefficient NSE and a root mean square error RMSE.

[0035] Generally, AquaCrop is a representative daily water balance general crop model based on water-driven, mainly applied in arid regions where water is a key limiting factor for crop production. AquaCrop requires only a small number of explicit parameters, and the input data is intuitive, clear and easy to obtain, which has been widely used in irrigation scheduling, management optimization and crop yield prediction for various crops and different regions. It has the ability to simulate crop yield at point, county and regional scales, and has great potential in guiding water-saving irrigation in arid regions. In the AquaCrop model, CC is an intuitive and direct index associated with crop evapotranspiration and biomass.

[0036] Canopy cover development stage:

[0037]

[0038] Canopy cover senescence process:

[0039]

[0040] Where CC0is the initial CC when the emergence rate is 90%, CC x is the maximum CC, CGC is the canopy growth coefficient, CDC is the canopy senescence coefficient, and t is the number of days after sowing.

[0041] The key process of converting CC into yield and biomass can be expressed as:

[0042] CC * = 1.72CC-CC 2 + 0.30CC 3

[0043]

[0044] B = Ks b WP * ∑Tr / ET0

[0045] Y = f HI HI0B

[0046] Where CC * represents the corrected CC, mainly considering the change of inter-row reflectivity and the masking effect due to partial canopy overlap; T r is the crop transpiration; K s is the soil water stress coefficient, including soil water stress, stomatal closure stress and soil salt stress; is the maximum crop transpiration coefficient corresponding to fully irrigated soil and fully covered canopy; Ks b is the air temperature stress coefficient; WP * is the normalized water productivity; Y represents crop yield; fHI is the adjustment factor for water stress before yield formation, the effects of pollination failure, and the effects of water stress during yield formation; HI0 is the reference harvest index; and B is the final biomass.

[0047] The relationship between canopy cover CC and leaf area index LAI can be expressed as:

[0048] CC=1.005·(1-e -0.6*LAI ) 1.2

[0049] S3: Construct a multi-objective indicator evaluation optimization model for regional-scale PCMs key parameter optimization.

[0050] S4: The multi-objective indicator evaluation optimization model is trained with minimizing the sum of the Nash efficiency coefficients NSE of canopy coverage and crop yield in the training set and the test set as the objective function, and with the root mean square error RMSE of canopy coverage and crop yield in the training set and the test set as the constraint condition.

[0051] The multi-objective evaluation optimization model takes minimizing the sum of the Nash efficiency coefficients NSE of canopy coverage and the result variable yield in the training set and the test set as the objective function, and uses the root mean square error RMSE as the constraint condition.

[0052] like Figure 3 As shown in the figure, by using the leave-one-out cross-validation method LOOCV, by eliminating one sample at a time and using the remaining samples to estimate the regression model parameters, it is suitable for learning small sample data sets. LOOCV can avoid the errors caused by random partitioning of the data set, can better reflect the distribution characteristics of the original samples, and the verification process is completely repeatable; Figure 2 Chinese: RMSE_Train i and NSE_Train i They represent the root mean square error and Nash efficiency coefficient of the training set during the i-th cross-validation (i=1,2,…,n); RMSE_Test and NSE_Test represent the root mean square error and Nash efficiency coefficient of the test set after all cross-validations are completed.

[0053] S5: The evaluation indicators are input into the trained multi-objective indicator evaluation optimization model, and a parallel algorithm group composed of multiple meta-heuristic optimization methods is used to solve the multi-objective indicator evaluation optimization model, optimize the crop parameter combination, and obtain the optimal crop parameter combination.

[0054] Exemplarily, according to the limitations of actually collectable data and research targets, the present application takes the actually measured process variable canopy cover CC and the result variable yield as the calibration indexes of the model, and selects the dimensionless NSE on the training set and the test set as the optimization targets and the dimensioned RMSE as the constraints, respectively.

[0055] Objective 1: Minimize the NSE of the training set:

[0056]

[0057] In the formula, F1 is the sum of the average NSE of the crop yield and CC training set obtained by using the leave-one-out cross validation method, and "-" is to set the fitness value as the negative value of the objective function; is the average NSE of the yield of the training set; is the average NSE of the CC of the training set.

[0058] Objective 2: Minimize the NSE of the test set:

[0059] Min F2 = - {NSE Y,Test + NSE CC,Test}

[0060] In the formula, F2 is the sum of the NSE of the yield and CC test set obtained by using the leave-one-out cross validation method, and "-" is to set the fitness value as the negative value of the objective function; NSE Y,Test is the NSE of the yield of the test set; NSE CC,Test is the NSE of the CC of the test set.

[0061] Constraint condition:

[0062]

[0063] wherein, and are the average RMSE of the yield and CC training set; RMSE Y,Test and RMSE CC,Test are the RMSE of the yield and CC test set; RMSE Y,crop and RMSE CC,crop are the constraint values of the yield and CC, which are respectively taken as different values according to different calibration crops. The penalty function method is used to process the constraints, and the constraints play the role of accelerating the convergence of the algorithm and meeting the needs of the lowest simulation accuracy of the model.

[0064] S5: The evaluation indexes are input into the multi-objective index evaluation optimization model, a parallel algorithm group composed of multiple meta-heuristic optimization methods is used to solve the multi-objective index evaluation optimization model, and the optimal crop parameter combination is obtained through iteration.

[0065] Exemplarily, each parameter in the crop parameter combination has a respective search interval and search step, the crop parameters constitute N combinations, and the present application needs to find the optimal combination. The search interval is determined according to the interval range value of the model default parameter, and can be comprehensively confirmed according to existing knowledge, historical data or through random search and multiple runs; the search step can be set according to the actual demand of each parameter to achieve the best balance. In the iteration process, the crop parameter combination is brought into the PCMs, and according to the evaluation index and the optimization model, the better combination is retained, and then the algorithm generates a new combination, which is brought into the PCMs again, and the above process is repeatedly repeated. The output constitutes the state variable of the optimization model of the next period: the evaluation index is obtained by bringing it into the PCMs, and the optimization is performed according to the index, that is, the state variable of the optimization model of the next period.

[0066] The above is the regional scale process model simulation precision optimization method provided by one or more embodiments of the present specification, based on the same idea, the present specification also provides a corresponding regional scale process model simulation precision optimization system, comprising:

[0067] An acquisition module is configured to input a crop parameter combination to be optimized into a crop-water productivity model AquaCrop-OSPy, and determine simulated crop yield and canopy coverage; wherein the crop parameter combination includes product data, measured data and calibration data.

[0068] A cross-validation module is configured to cross-validate the crop yield and the canopy coverage by using a leave-one-out cross-validation method (LOOCV), and determine an evaluation index corresponding to the crop parameter combination; wherein the evaluation index includes a Nash efficiency coefficient (NSE) and a root mean square error (RMSE).

[0069] A construction module is configured to construct a multi-objective index evaluation optimization model for optimization of key parameters of regional scale PCMs.

[0070] A training module is configured to minimize the sum of the Nash efficiency coefficients (NSE) of the canopy coverage and the crop yield in the training set and the test set as an objective function, and take the root mean square errors (RMSE) of the canopy coverage and the crop yield in the training set and the test set as constraint conditions, and train the multi-objective index evaluation optimization model.

[0071] An optimization module is configured to input the evaluation index into the trained multi-objective index evaluation optimization model, use a parallel algorithm group composed of multiple meta-heuristic optimization methods to solve the multi-objective index evaluation optimization model, optimize the crop parameter combination, and obtain an optimal crop parameter combination.

[0072] The specific definition of the system for optimizing the simulation accuracy of the regional-scale process model can refer to the definition of the method for optimizing the simulation accuracy of the regional-scale process model, which will not be repeated here. Each module in the system for optimizing the simulation accuracy of the regional-scale process model can be realized by software, hardware, or a combination thereof, in whole or in part. Each module can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operation corresponding to each module.

[0073] The application also provides a computer device for implementing the method for optimizing the simulation accuracy of the regional-scale process model. Figure 4 The structural diagram of the computer device is shown in FIG. 1, which includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory. Figure 4 At the hardware level, the computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and can also include other hardware required by the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs to realize the method for optimizing the simulation accuracy of the regional-scale process model provided by the above embodiments.

[0074] Each technical feature of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of each technical feature in the above embodiments are described, however, as long as the combination of these technical features does not exist, it should be considered as the range disclosed by the application.

Claims

1. A method for optimizing simulation accuracy of regional scale process models, characterized in that: include: The crop parameter combination to be optimized is input into the crop-water productivity model AquaCrop-OSPy to determine the simulated crop yield and canopy cover; wherein the crop parameter combination includes product data, measured data, and calibration data; the product data includes soil hydraulic parameters, soil texture, and meteorological data; the measured data includes yield, leaf area index (LAI), irrigation data, and field management data; and the calibration data includes the study area, study object, key crop parameters, search interval, and search step size; The crop yield and canopy coverage are cross-validated by using the leave-one-out cross-validation method (LOOCV) to determine the evaluation index corresponding to the crop parameter combination; wherein the evaluation index includes the Nash efficiency coefficient (NSE) and the root mean square error (RMSE); Construct a multi-objective indicator evaluation optimization model for regional-scale PCMs key parameter optimization; The multi-objective indicator evaluation optimization model is trained by minimizing the sum of the Nash efficiency coefficients (NSE) of canopy coverage and crop yield in the training set and the test set as the objective function, and using the root mean square error (RMSE) of canopy coverage and crop yield in the training set and the test set as the constraint condition; The constraints specifically include: RMSE Y，Test ≤RMSE Y，crop RMSE CC,，Test ≤RMSE CC,crop in, and are the RMSE of the average CC training set for crop yield and canopy coverage, respectively; RMSE Y,Test and RMSE CC,Test are the RMSE of the CC test set for crop yield and canopy cover, respectively; Y,crop and RMSE CC,crop are the constraint values ​​of crop yield and canopy coverage CC, respectively; The evaluation indicators are input into the trained multi-objective indicator evaluation optimization model, and a parallel algorithm group composed of multiple meta-heuristic optimization methods is used to solve the multi-objective indicator evaluation optimization model, optimize the crop parameter combination, and obtain the optimal crop parameter combination; Also includes: In the iterative process, the crop parameter combination is brought into PCMs, the optimization model is evaluated according to the evaluation index and multi-objective index, a new crop parameter combination is obtained, the crop parameter combination and the new crop parameter combination are brought into PCMs again, and the above steps are repeated until the optimal crop parameter combination is obtained.

2. The method for optimizing regional scale process model simulation accuracy according to claim 1, characterized in that: The parallel algorithm group composed of multiple meta-heuristic optimization methods is used to solve the multi-objective index evaluation optimization model, specifically including: Parallel algorithms including UNSGA3, MOEA / D, SPEA2, AGEMOEA, CTAEA and SMSEMOA are used to solve the multi-objective index evaluation optimization model.

3. A regional scale process model simulation accuracy optimization system, characterized by: include: An acquisition module is used to input the crop parameter combination to be optimized into the crop-water productivity model AquaCrop-OSPy to determine the simulated crop yield and canopy cover; wherein the crop parameter combination includes product data, measured data, and calibration data; the product data includes soil hydraulic parameters, soil texture, and meteorological data; the measured data includes yield, leaf area index (LAI), irrigation data, and field management data; and the calibration data includes the study area, study object, key crop parameters, search interval, and search step size; A cross-validation module is used to perform cross-validation on crop yield and canopy cover by using a leave-one-out cross-validation method (LOOCV) to determine evaluation indicators corresponding to crop parameter combinations; wherein the evaluation indicators include Nash efficiency coefficient (NSE) and root mean square error (RMSE); A construction module for building a multi-objective indicator evaluation optimization model for regional-scale PCMs key parameter optimization; The training module is used to train the multi-objective indicator evaluation optimization model by minimizing the sum of the Nash efficiency coefficients (NSE) of canopy coverage and crop yield in the training set and the test set as the objective function, and using the root mean square error (RMSE) of canopy coverage and crop yield in the training set and the test set as the constraint condition; the constraint conditions specifically include: RMSE Y，Test ≤RMSE Y，crop RMSE CC,Test ≤RMSE CC,crop in, and are the RMSE of the average CC training set for crop yield and canopy coverage, respectively; RMSE Y,Test and RMSE CC,Test are the RMSE of the CC test set for crop yield and canopy cover, respectively; Y,crop and RMSE CC,crop are the constraint values ​​of crop yield and canopy coverage CC, respectively; The optimization module is used to input the evaluation indicators into the trained multi-objective indicator evaluation optimization model, solve the multi-objective indicator evaluation optimization model using a parallel algorithm group composed of multiple meta-heuristic optimization methods, optimize the crop parameter combination, and obtain the optimal crop parameter combination; Also includes: The iterative optimization module is used to obtain a new crop parameter combination by bringing the crop parameter combination into PCMs during the iteration process, evaluating the optimization model according to the evaluation index and multi-objective index, bringing the crop parameter combination and the new crop parameter combination into PCMs again, and repeating the above steps until the optimal crop parameter combination is obtained.

4. A computer device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for optimizing the simulation accuracy of a regional-scale process model according to any one of claims 1 to 2 is implemented.

Images