Optimization method and system for mixing and stirring of chemical products

Abstract

The application discloses a kind of optimization method and system of chemical product mixing stirring. By determining original complex variable and original non-complex variable from the related parameter variable of mixing pool stirring, the original model for solving the optimal solution of these variables is determined. And when solving for the first time using the original model, it is converted into a mixed integer linear programming model MILP model with higher solution accuracy to obtain the first model output solution, and the model optimal cut determined thereby is inserted into the LP2 model for solving to obtain the second model output solution. Then according to the first model output solution and the second model output solution, a judgment is made, and whether to output variable optimal solution is decided according to the judgment result. As can be seen, the embodiment of the present application can start from the related parameter variable of mixing pool stirring, and output variable optimal solution is processed by combining MILP model and LP2 model, so as to ensure the maximum utilization of chemical product mixing stirring.

Description

Technical Field

[0001] This application relates to the field of chemical technology, and in particular to an optimized method and system for mixing and stirring chemical products. Background Technology

[0002] In the processing of chemical products such as oil and natural gas, the mixing and stirring of intermediate products is often involved. To produce different series of products (usually containing different substances), chemical companies input raw material streams containing different substances (such as varying ethylene and sulfur contents) into mixing tanks or endpoints with different stirring capacities for mixing. The output streams converge at different endpoints and are mixed again, ultimately resulting in different series of chemical products. The production process is as follows: Figure 1 As shown. However, in actual production environments, the input involves dozens or even hundreds of raw material streams with different components; the number of mixing tanks is usually around ten; and the endpoint may involve hundreds of products with different components. Multiplying these numbers together, we can roughly estimate that the number of variables in the model will reach tens of thousands. Moreover, there are numerous constraints between the various types of variables, reaching thousands or even tens of thousands. Therefore, how to maximize the utilization of chemical product mixing and agitation is a problem that urgently needs to be solved. Summary of the Invention

[0003] This invention provides an optimized method and system for mixing and stirring chemical products, in order to solve or partially solve the problem of not being able to maximize the utilization of chemical product mixing and stirring.

[0004] To solve the above-mentioned technical problems, the present invention provides an optimized method for mixing and stirring chemical products, the method comprising:

[0005] The original complex variables and the original non-complex variables are determined from the relevant parameter variables of the mixing tank; wherein, the relevant parameter variables of the mixing tank include: pipeline flow rate parameter variables, raw material parameter variables, intermediate product parameter variables, and mixing tank mixing capacity parameter variables;

[0006] The original model is determined based on the original complex variables and the original non-complex variables; wherein, the original model includes the original objective function and the original constraints, and the original objective function and / or the original constraints contain the original complex variables and the original non-complex variables;

[0007] If the original model is being solved for the first time, the original model is transformed into a mixed integer linear programming (MILP) model for solution.

[0008] If the MILP model is solved successfully, the corresponding first model output solution is obtained;

[0009] The optimal slicing of the model is obtained based on the output solution of the first model.

[0010] The optimal solution of the second model is obtained by inserting the model into the linear optimization model LP2 and solving it.

[0011] The judgment is made based on the output solutions of the first model and the output solutions of the second model;

[0012] If the judgment is successful, the output solutions of the first model and the second model will be output as the corresponding optimal solutions for the variables.

[0013] Preferably, the output solution of the first model includes: values ​​of uncomplex variables, Lagrange factors, and the optimal value of the objective function of the first model; the output solution of the second model includes: values ​​of complex variables and the optimal value of the objective function of the second model.

[0014] The step of obtaining the optimal slicing of the model based on the output solution of the first model specifically includes:

[0015] The optimal cut of the model is obtained based on the uncomplex variable values ​​of the first model output solution, the Lagrange factor of the first model output solution, the constraints of the MILP model, and the objective function of the MILP model.

[0016] The step of optimally inserting the model into the LP2 model for solution to obtain the output solution of the second model specifically includes:

[0017] The optimal model is then placed into the constraints of the LP2 model for solution, and the output solution of the second model is obtained.

[0018] The judgment based on the output solutions of the first model and the second model specifically includes:

[0019] The optimal values ​​of the first model objective function and the second model objective function are compared. If the optimal value of the first model objective function is less than or equal to the optimal value of the second model objective function, the comparison is successful.

[0020] Preferably, after determining the solution based on the first model output solution and the second model output solution, the method further includes:

[0021] If the judgment fails, the original complex variable in the original model is fixed according to the complex variable value in the solution output by the second model, so as to transform the original model into the first LP1 model for solution;

[0022] Determine whether the first LP1 model has a feasible solution;

[0023] If a feasible solution exists, the optimal cut is obtained by solving the first output solution according to the first LP1 model.

[0024] The optimal solution of the first LP1 model is incorporated into the LP2 model for solution to obtain the second output solution;

[0025] The determination is made based on the first output solution and the second output solution;

[0026] If the judgment is successful, the first output solution and the second output solution will be output as the corresponding optimal solutions for the variables;

[0027] If the judgment fails, the original complex variable in the original model is fixed according to the complex variable value in the second output solution, and the above steps are repeated until the judgment is successful according to the first output solution and the second output solution.

[0028] Preferably, the first output solution includes: uncomplex variable values, Lagrange factors, and the optimal value of the first objective function; the second output solution includes: complex variable values ​​and the optimal value of the second objective function.

[0029] The step of obtaining the optimal cut based on the first output solution specifically includes:

[0030] The optimal cut is obtained based on the uncomplex variable values ​​of the first output solution, the Lagrange factor of the first output solution, the constraints of the first LP1 model, and the objective function of the first LP1 model.

[0031] The step of inserting the first output solution and its optimal value into the LP2 model for solving to obtain the second output solution specifically includes:

[0032] The optimal cut is incorporated into the constraints of the LP2 model and solved to obtain the second output solution;

[0033] The determination based on the first output solution and the second output solution specifically includes:

[0034] Determine the magnitudes of the optimal values ​​of the first and second objective functions. If the optimal value of the first objective function is less than or equal to the optimal value of the second objective function, the determination is successful; and / or

[0035] If the optimal value of the second objective function remains consistent throughout N consecutive iterations, the judgment is successful; where N≥3 and is a positive integer.

[0036] Preferably, after determining whether the first LP1 model has a feasible solution, the method further includes:

[0037] If no feasible solution is found, the first LP1 model is replaced with the second LP1 model for solving, and a third output solution is obtained.

[0038] A feasible cut is obtained based on the third output solution;

[0039] The third output solution and its feasibility are inserted into the LP2 model for solving to obtain the fourth output solution;

[0040] The determination is made based on the third and fourth output solutions;

[0041] If the judgment is successful, the third output solution and the fourth output solution will be output as the corresponding optimal solutions for the variables.

[0042] If the judgment fails, the original complex variable in the original model is fixed according to the complex variable value in the third output solution, and the process is switched to the step of converting the original model into the first LP1 model for solution to obtain the first output solution.

[0043] Preferably, the third output solution includes: uncomplex variable values, Lagrange factors, and the maximum value of the third output solution; the fourth output solution includes: complex variable values ​​and the minimum value of the fourth output solution.

[0044] The step of obtaining a feasible cut based on the third output solution specifically includes:

[0045] The feasible cut is obtained based on the Lagrange factor of the third output solution and the constraints of the second LP1 model;

[0046] The step of inserting the first output solution and its optimal value into the LP2 model for solving to obtain the second output solution specifically includes:

[0047] The feasible solution is then solved by applying the constraints to the LP2 model to obtain the second output solution.

[0048] The determination based on the third output solution and the fourth output solution specifically includes:

[0049] Determine the magnitude of the maximum value of the third output solution and the minimum value of the fourth output solution. If the maximum value of the third output solution is less than or equal to the minimum value of the fourth output solution, the determination is successful; and / or

[0050] If the minimum value of the fourth output solution remains consistent across N consecutive loops, the judgment is successful; where N≥3 and is a positive integer.

[0051] Preferably, the step of converting the original model into a mixed-integer linear programming (MILP) model for solution specifically includes:

[0052] The original objective function in the original model is used as the objective function of the MILP model;

[0053] Obtain the new feasible domains for both the original complex variable and the original non-complex variable;

[0054] The constraints of the MILP model are determined using the new feasible regions of the original complex variables and the original uncomplex variables respectively.

[0055] The solution is obtained based on the objective function and constraints of the MILP model.

[0056] This invention discloses an optimized system for mixing and stirring chemical products, comprising:

[0057] The first determining module is used to determine the original complex variables and the original non-complex variables from the relevant parameter variables of the mixing tank; wherein, the relevant parameter variables of the mixing tank include: pipeline flow rate parameter variables, raw material parameter variables, intermediate product parameter variables, and mixing tank mixing capacity parameter variables;

[0058] The second determining module is used to determine the original model based on the original complex variables and the original non-complex variables; wherein, the original model includes an original objective function and original constraints, and the original objective function and / or the original constraints contain original complex variables and original non-complex variables;

[0059] The transformation module is used to transform the original model into a mixed integer linear programming (MILP) model for solving if the original model is being solved for the first time.

[0060] The first module is used to obtain the corresponding first model output solution if the MILP model is solved successfully.

[0061] The second obtaining module is used to obtain the optimal slicing of the model based on the output solution of the first model.

[0062] The third module is used to optimally insert the model into the LP2 model for solving, and obtain the output solution of the second model.

[0063] The judgment module is used to make a judgment based on the output solution of the first model and the output solution of the second model;

[0064] The output module is used to output the first model's output solution and the second model's output solution as the corresponding optimal solutions if the judgment is successful.

[0065] The present invention discloses a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0066] The present invention discloses a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the above-described method.

[0067] Through one or more technical solutions of the present invention, the present invention has the following beneficial effects or advantages:

[0068] This invention provides an optimization method and system for mixing and stirring chemical products. By identifying the original complex and uncomplex variables from the relevant parameters of the mixing tank, an original model for solving these variables with optimal solutions is determined. When initially using the original model, it is transformed into a Mixed Integer Linear Programming (MILP) model with higher solution accuracy to obtain a first model output solution. This optimal model is then used to solve the second model output solution using an LP2 model. A judgment is then made based on the first and second model output solutions to determine whether to output the optimal solution for the variables. Therefore, this invention can start with the relevant parameters of the mixing tank, and by combining the MILP and LP2 models, output the optimal solution for the variables, thereby ensuring the maximum utilization of the mixing and stirring of chemical products.

[0069] Furthermore, choosing a high-precision mixed-integer linear programming (MILP) model for the initial solution ensures stability and accuracy, and optimizes the accuracy of subsequent models (first LP1 model, second LP1 model, LP2 model, etc.). Specifically, the output solution of the first model obtained using the MILP model affects the optimal tangent of the model, which in turn affects the output quality of the LP2 model. The output of the LP2 model, in turn, affects the output quality and optimal tangent of the first LP1 model in the next iteration, which in turn affects the output quality of the LP2 model in the next iteration. This cycle continues, improving the accuracy of variable solving and thus ensuring the maximum utilization of the chemical product mixing process.

[0070] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0071] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0072] Figure 1 A flowchart of an optimized method for mixing and stirring chemical products according to an embodiment of the present invention is shown;

[0073] Figure 2 A schematic diagram illustrating the principle of an optimized method for mixing and stirring chemical products according to an embodiment of the present invention is shown.

[0074] Figure 3 A schematic diagram of an optimized system for mixing and stirring chemical products according to an embodiment of the present invention is shown. Detailed Implementation

[0075] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0076] To illustrate and explain the embodiments of the present invention, the various names involved will be explained below.

[0077] MILP model: Mixed Integer Linear Programming Model.

[0078] The first LP1 model: a linear optimization model with fixed complex variables.

[0079] The second LP1 model: a relaxed linear optimization model with fixed complex variables, also known as the relaxed LP1 model, which can always find a solution.

[0080] LP2 model: A linear optimization model with uncomplex variables fixed.

[0081] The model solving process involved in this embodiment specifically involves using a solver to run the model for solving.

[0082] Currently, due to the numerous variables and constraints involved in the mixing and stirring of chemical products, most of these variables appear in the form of bilinear terms (such as the multiplication of two variables), which simultaneously exhibit nonlinear and nonconvex characteristics. Therefore, considering these characteristics of the variables involved in mixing and stirring, this invention provides an optimization method for the mixing and stirring of chemical products. By determining the original complex and uncomplex variables from the relevant parameter variables of the mixing tank, an original model for solving these variables with optimal solutions is determined. When the original model is used for the first time, it is transformed into a mixed integer linear programming (MILP) model with higher solution accuracy to obtain the first model output solution. This determined optimal model is then used in the LP2 model for solving, yielding the second model output solution. Finally, a judgment is made based on the first and second model output solutions, and the optimal solution for each variable is output according to the judgment result. As can be seen, the embodiments of the present invention can start from the relevant parameter variables of the mixing tank, and take into account the nonlinear and nonconvex characteristics of the bilinear term. By combining the MILP model and the LP2 model, the optimal solution of the output variable is processed, thereby ensuring the maximum utilization of the mixing of chemical products.

[0083] Please refer to the following. Figures 1-2 The method of this invention includes the following steps:

[0084] Step 101: Determine the original complex variables and the original non-complex variables from the relevant parameter variables of the mixing tank.

[0085] The relevant parameters and variables for mixing in the mixing tank include: pipeline flow rate, raw material parameters, intermediate product parameters, and mixing tank mixing capacity parameters.

[0086] The definitions of the original complex and non-complex variables originate from bilinear terms. For example, a bilinear term might be in the form of x*y, where both x and y are variables. The complexity and non-complexity of these variables can be determined randomly or based on the specific circumstances. For instance, if x is defined as a complex variable, then y is a non-complex variable. Conversely, if x is defined as a non-complex variable, then y is a complex variable. Since the definition of complex and non-complex variables has little impact on the accuracy of the model's processing results, one of the bilinear terms can be randomly chosen as a complex variable and the other as a non-complex variable. Alternatively, it can be determined based on the actual complexity of the variables. For example, in a mixing tank scenario, the mixing capacity of the mixing tank can be defined as a complex variable, and the pipe flow rate as a non-complex variable. If the mixing tank model contains a large number of bilinear terms representing pipe flow rate (x) * mixing capacity (y), then y is a complex variable and x is a non-complex variable.

[0087] Step 102: Determine the original model based on the original complex variables and the original non-complex variables.

[0088] The original model includes an original objective function and original constraints, wherein the original objective function and / or the original constraints contain original complex variables and original non-complex variables.

[0089] In the process of determining the original model, the original objective function and original constraints are determined based on the original complex variables and original non-complex variables, thereby determining the original model so that the optimal solution of the original objective function can be determined under the constraints of the original constraints. This optimal solution is the final optimal solution of the variables.

[0090] Furthermore, based on the original complex variables and the original non-complex variables, the pre-created original objective function and original constraints can be retrieved; alternatively, the original objective function and original constraints can be created directly based on the original complex variables and the original non-complex variables. For example, the original constraints can be determined using each of the original complex variables and the original non-complex variables.

[0091] Step 103: If the original model is being solved for the first time, the original model is transformed into a mixed integer linear programming (MILP) model for solution.

[0092] In this context, "original model" refers to the first solution obtained by using the original complex and non-complex variables. This is because subsequent steps may not yield optimal solutions, requiring further (or multiple) adjustments to the values ​​of the original complex or non-complex variables. Therefore, this embodiment distinguishes between "first time" and "continuously," see [link to documentation]. Figure 2 .

[0093] If the original model is being solved for the first time, it is then converted into a more accurate MILP model for further solving.

[0094] In the specific transformation process, the original objective function in the original model is used as the objective function of the MILP model; the new feasible regions of the original complex variables and the original uncomplex variables are obtained; the constraints of the MILP model are determined using the new feasible regions of the original complex variables and the original uncomplex variables; and the solution is performed based on the objective function and constraints of the MILP model.

[0095] Step 104: If the MILP model is solved successfully, the corresponding first model output solution is obtained.

[0096] Specifically, the output solution of the first model includes: the values ​​of uncomplex variables, the Lagrange factor, the optimal value of the objective function of the first model, that is: the maximum value of the entire original model.

[0097] Step 105: Obtain the optimal cut of the model based on the output solution of the first model.

[0098] In the specific implementation process, the optimal cut of the model is obtained based on the uncomplex variable values ​​of the first model output solution, the Lagrange factor of the first model output solution, the constraints of the MILP model, and the objective function of the MILP model.

[0099] In the specific implementation process of obtaining the optimal cut of the model, the corresponding uncomplex variables in the constraints of the MILP model are fixed according to the uncomplex variable values ​​of the first model output solution. That is, the uncomplex variables in the constraints of the MILP model are replaced with the uncomplex variable values ​​of the first model output solution. Then, all the Langron factors of the MILP model are multiplied by their respective fixed constraints and summed. Finally, the objective function in the MILP model is added to obtain the optimal cut of the model.

[0100] For example, suppose a MILP problem has an objective function a(x) and constraints b(x,y) and c(x,y). Solving the MILP model yields the values ​​of the non-complex variables. And the Lagrange factors d and e (d and e correspond one-to-one with b and c). Then the optimal tangent of the model is the sum of all Lagrange factors multiplied by their corresponding constraints in the MILP model, plus the objective function in the MILP model (i.e., ...). ).

[0101] The role of the optimal tangent in the model is to add it as a new constraint to the original constraints of the linear optimization model LP2, thereby solving the objective function of the LP2 model. It is evident that using a more accurate MILP model to obtain the first model output solution will also improve the solution accuracy of the subsequent linear optimization model LP2.

[0102] Step 106: The model is optimally placed into the linear optimization model LP2 for solution to obtain the output solution of the second model.

[0103] Specifically, the optimal model is placed into the constraints of the LP2 model for solution, and the output solution of the second model is obtained.

[0104] The output solution of the second model includes: the values ​​of complex variables, the optimal value of the objective function of the second model, which is also the minimum value of the entire original model.

[0105] Step 107: Make a judgment based on the output solution of the first model and the output solution of the second model.

[0106] Specifically, during the judgment process, the optimal values ​​of the first model's objective function and the second model's objective function are compared. If the optimal value of the first model's objective function is less than or equal to the optimal value of the second model's objective function, the judgment is successful, indicating that the currently obtained uncomplex variable values ​​and complex variable values ​​are the optimal solutions for the variables. If the optimal value of the first model's objective function is greater than the optimal value of the second model's objective function, the judgment fails, indicating that the currently obtained uncomplex variable values ​​and complex variable values ​​are not the optimal solutions for the variables. Furthermore, the process also checks whether the iteration has ended. If the iteration has ended, the judgment is also successful, and the output solution corresponding to the last iteration is output.

[0107] Step 108: If the judgment is successful, output the first model output solution and the second model output solution as the corresponding optimal solutions for the variables.

[0108] As can be seen, the embodiments of the present invention can start from the relevant parameter variables of the mixing tank, and process the output variables by combining the MILP model and the LP2 model, thereby ensuring the maximum utilization of the mixing of chemical products.

[0109] After judging based on the output solutions of the first model and the second model, if the judgment fails, the complex variable values ​​in the output solution of the second model are put into the original model for solving.

[0110] In the specific implementation process, the original complex variables in the original model are fixed according to the complex variable values ​​in the output solution of the second model, so as to transform the original model into the first LP1 model for solving and obtain the first output solution. Due to the solution accuracy of the LP2 model, using the complex variable values ​​in the output solution of the second model to input into the original model for solving will also improve the solution accuracy of the original model.

[0111] Specifically, during the fixation process, the complex variable values ​​in the output solution of the second model are used to replace the original complex variables in the original model, thus transforming the original model into the first LP1 model. Therefore, the objective function and / or constraints in the first LP1 model are only related to non-complex variables, thereby reducing the difficulty of solving the model.

[0112] The solution determines whether the first LP1 model has a feasible solution. Specifically, when the solver runs the first LP1 model, if the solver can solve it, it returns a real value with the information "feasible"; if the first LP1 model cannot be solved, the solver does not output any result and returns the information "infeasible". Therefore, the feasible or infeasible information is used to determine whether the first LP1 model has a feasible solution.

[0113] If a feasible solution exists, the optimal tangent is obtained by solving the first output solution according to the first LP1 model. The first output solution includes: values ​​of uncomplex variables, Lagrange factors, and the optimal value of the first objective function.

[0114] Specifically, the implementation process of the optimal cut of the first LP1 model is similar to that of the optimal cut of the aforementioned model. Based on the non-complex variable values ​​of the first output solution, the corresponding non-complex variables in the constraints of the first LP1 model are fixed. In other words, the non-complex variables in the constraints of the first LP1 model are replaced by the non-complex variable values ​​of the first output solution. Then, all the Langran factors of the first output solution are multiplied by their respective fixed constraints and summed. Finally, the objective function of the first LP1 model is added, thus obtaining the optimal cut of the first LP1 model.

[0115] The optimal solution of the first LP1 model is incorporated into the LP2 model for solution, yielding a second output solution. Specifically, the optimal solution is incorporated into the constraints of the LP2 model for solution, resulting in the second output solution. The second output solution includes: values ​​of complex variables and the optimal value of the second objective function. It is evident that the solution accuracy of the first LP1 model affects the solution accuracy of the LP2 model.

[0116] The judgment is made based on the first output solution and the second output solution.

[0117] In the specific judgment process, the optimal values ​​of the first objective function and the second objective function are compared. If the optimal value of the first objective function is less than or equal to the optimal value of the second objective function, the judgment is successful. And / or whether the optimal value of the second objective function remains consistent across N consecutive iterations is also checked; if so, the judgment is successful. Here, N ≥ 3 and is a positive integer. Furthermore, the process also checks whether the iteration has ended. If the iteration has ended, the judgment is also successful, and the output solution corresponding to the last iteration is output.

[0118] If the judgment is successful, it means that the currently obtained non-complex variable values ​​and complex variable values ​​are the optimal solutions for the variables. Then, the first output solution and the second output solution are output as the corresponding optimal solutions for the variables.

[0119] If the judgment fails, it means that the currently obtained non-complex variable values ​​and complex variable values ​​are not the optimal solution for the variable. Based on the complex variable values ​​in the second output solution, the original complex variable in the original model is fixed, and the above steps are repeated sequentially until the judgment based on the first output solution and the second output solution is successful.

[0120] As an optional embodiment, after determining whether the first LP1 model has a feasible solution, if there is no feasible solution, it indicates that the first LP1 model is unsolvable. Then, the first LP1 model is replaced with a second LP1 model for solving, obtaining a third output solution. The second LP1 model is also called a relaxed LP1 model, meaning that running the solver on this relaxed LP1 model will definitely yield a solution. In this embodiment, the solution result of the relaxed LP1 model is referred to as the third output solution. During the process of replacing the first LP1 model with the second LP1 model, relaxation variables of the second LP1 model are obtained. The objective function and constraints of the first LP1 model are updated based on these relaxation variables to obtain the objective function and constraints of the second LP1 model. Specifically, relaxation variables are used to replace variables in the objective function of the first LP1 model, making the objective function of the first LP1 model become an objective function related to the relaxation variables, i.e., the objective function of the second LP1 model. Relaxation variables are inserted into the constraints of the first LP1 model to obtain the constraints of the second LP1 model. In this model, the obtained third output solution is independent of the objective function and only related to the constraints. (Relaxation variables) Its function is to determine which constraint makes the first LP1 model unsolvable, and to make the second LP1 model solvable based on that constraint.

[0121] Therefore, the role of the relaxed LP1 model is to ensure that a solution is obtained (i.e., the third output solution) when the complex variable values ​​obtained by the LP2 model are not good (e.g., the accuracy is not high) and to guarantee the smooth output of subsequent models.

[0122] Furthermore, a feasible cut is obtained based on the third output solution. The third output solution includes: uncomplex variable values, Lagrange factors, and the maximum value of the third output solution. In the process of obtaining the feasible cut, the feasible cut is obtained based on the uncomplex variable values ​​of the third output solution, the Lagrange factors of the third output solution, and the constraints of the second LP1 model. Specifically, the uncomplex variables corresponding to the constraints of the second LP1 model are fixed based on the uncomplex variable values ​​of the third output solution; that is, the uncomplex variables corresponding to the constraints of the second LP1 model are replaced by the uncomplex variable values ​​of the third output solution. Then, all the Lagrange factors of the second LP1 model are multiplied by their respective fixed constraints and summed to obtain the feasible cut.

[0123] For example, suppose a relaxed LP1 model has constraints b(x,y) and c(x,y). The values ​​of the uncomplex variables obtained after solving the relaxed LP1 model are... And the Lagrange factors d and e (d and e correspond one-to-one with b and c). The feasible solution is obtained by multiplying all Lagrange factors by their corresponding constraint expressions in the relaxed LP1 model and then summing the results.

[0124] The third output solution and its feasible solutions are then incorporated into the LP2 model for solving to obtain the fourth output solution. Specifically, the feasible solutions are incorporated into the constraints of the LP2 model for solving to obtain the fourth output solution. The fourth output solution includes: the values ​​of the complex variables, and the minimum value of the fourth output solution.

[0125] Furthermore, a judgment is made based on the third and fourth output solutions. During the judgment process, the maximum value of the third output solution and the minimum value of the fourth output solution are compared. If the maximum value of the third output solution is less than or equal to the minimum value of the fourth output solution, the judgment is successful. And / or whether the minimum value of the fourth output solution remains consistent across N consecutive iterations is also checked. If so, the judgment is successful. Here, N ≥ 3 and is a positive integer. Additionally, it is determined whether the iteration has ended. If the iteration has ended, the judgment is also successful, and the output solution corresponding to the last iteration is output.

[0126] If the judgment is successful, it means that the currently obtained non-complex variable values ​​and complex variable values ​​are the optimal solutions for the variables. Then, the third output solution and the fourth output solution will be output as the corresponding optimal solutions for the variables.

[0127] If the judgment fails, it means that the currently obtained non-complex variable values ​​and complex variable values ​​are not the optimal solution for the variables. Based on the complex variable values ​​in the third output solution, the original complex variables in the original model are fixed, and the process proceeds to the step of converting the original model into a first LP1 model for solution, to obtain the first output solution.

[0128] To more clearly illustrate and explain the embodiments of the present invention, specific examples are used below for illustration.

[0129] In the first example, the model of the mixing tank contains a large number of bilinear terms of pipeline flow rate (x) * mixing tank agitation capacity (y), where y is a complex variable and x is a non-complex variable.

[0130] Determine the original complex variable (mixing capacity y of the mixing tank) and the original uncomplex variable (pipe flow rate x) from the relevant parameters of the mixing tank.

[0131] The following sections will introduce each model.

[0132] Original model.

[0133] parameter:

[0134] I: Mixing pool input and output flow rates;

[0135] X: The feasible region of variable x;

[0136] Y: The feasible region of variable y;

[0137] i: number;

[0138] variable:

[0139] x: a continuous variable;

[0140] y: a continuous variable;

[0141] Objective function:

[0142] Among them, c i This represents the cost of the corresponding raw material flow, where i represents the raw material flow.

[0143] Constraints:

[0144] g(x,y)≤0(1), where x is the pipeline flow rate (variable) and y is the mixing capacity of the mixing tank (variable). The general form of g(x,y) is a*x*y+b*x+c (a, b, and c are all coefficients).

[0145] h(x, y) = 0 (2), where x is the pipeline flow rate (variable) and y is the mixing capacity of the mixing tank (variable). The general form of h(x, y) is a*x*y + b*x + c (a, b, and c are all coefficients);

[0146] g(x)≤0(3), where x is the pipe flow rate (variable). The general form of g(x) is b*x+c (where b and c are coefficients);

[0147] h(x) = 0 (4), where x is the pipe flow rate (variable). The general form of h(x) is b*x + c (where b and c are coefficients).

[0148] h(y) = 0 (5), where y is the mixing capacity (variable) of the mixing tank. The general form of h(y) is b*y + c (where b and c are coefficients);

[0149] g(y)≤0(6), where y is the mixing capacity (variable) of the mixing tank. The general form of g(y) is b*y+c (where b and c are coefficients);

[0150] x∈X, y∈Y

[0151] Among them, the objective function is a linear function; (1) is an inequality with bilinear terms; (2) is an equation with bilinear terms; (3) and (6) are linear inequalities; (4) and (5) are linear equations.

[0152] MILP model.

[0153] parameter:

[0154] I: Mixing pool input and output flow rates;

[0155] x l x u : The minimum and maximum values ​​of the feasible region of x;

[0156] y l y u : The minimum and maximum values ​​of the feasible region of y;

[0157] d: Select the number of points (at least 2);

[0158] variable:

[0159] x: a continuous variable;

[0160] y: a continuous variable;

[0161] z i : 0-1 variable;

[0162] w: a continuous variable, representing x*y;

[0163] Objective function:

[0164]

[0165] Constraints:

[0166] g′(x)+w≤0(22), where g′(x) is b*x+c in the original problem;

[0167] h′(x)+w=0(23), where h′(x) is b*x+c in the original problem;

[0168]

[0169]

[0170]

[0171] g(x)≤0(27)

[0172] h(x)=0(28)

[0173] x∈X, y∈Y

[0174] Wherein, the objective function is a linear function; g′(x) in (22) is a linear inequality; h′(x) in (23) is a linear equality; and (24) indicates that in [y l y u Only one value can be selected from ]; (25), (26) represent z i When = 1,

[0175] First LP1 model.

[0176] parameter:

[0177] I: Mixing pool input and output flow rates;

[0178] k: indicates the iteration number;

[0179] The constant value is derived from LP2 of the (k-1)th iteration;

[0180] variable:

[0181] x: a continuous variable;

[0182] Objective function:

[0183]

[0184] Constraints:

[0185] In this context, x represents the pipe flow rate (variable). This refers to the mixing capacity of the mixing tank (a fixed value). Among them... The general form is The form (a, b, c are all coefficients);

[0186] In this context, x represents the pipe flow rate (variable). This refers to the mixing capacity of the mixing tank (a fixed value). Among them... The general form is The form (a, b, c are all coefficients);

[0187] g(x)≤0(9)

[0188] h(x)=0(10)

[0189] x∈X

[0190] Among them, the objective function is a linear function; (7) is a linear inequality; (8) and (10) are linear equations; and (9) is a linear inequality.

[0191] Second LP1 model.

[0192] parameter:

[0193] X: The feasible region of variable x;

[0194] k: indicates the iteration number;

[0195] The constant value is derived from LP2 of the (k-1)th iteration;

[0196] variable:

[0197] x: a continuous variable;

[0198] α: slack variable;

[0199] Objective function:

[0200] Minα

[0201] Constraints:

[0202]

[0203]

[0204]

[0205] g(x)-α≤0(14)

[0206] h(x)-α≤0(15)

[0207] -h(x)-α≤0(16)

[0208] α≥0(17)

[0209] Among them, (11) is a linear inequality; (12), (13), (14), (15), and (16) are linear inequalities.

[0210] LP2 model:

[0211] parameter:

[0212] Y: The feasible region of variable y;

[0213] K feasible The set of iterations for the first LP1 model when a feasible solution exists;

[0214] K infeasible The set of iterations when the first LP1 model has no feasible solution;

[0215] u k , λ k ,λ1 k ,λ2 k: The Lagrange factor at the k-th iteration;

[0216] variable:

[0217] y: a continuous variable;

[0218] u: continuous variable;

[0219] Objective function:

[0220] Min u

[0221] L * (y;u k , λ k )≤u(18),k∈K feasible

[0222] L * (y;u k ,λ1 k ,λ2 k )≤0(19),k∈K infeasible

[0223] h(y)=0(20)

[0224] g(y)≤0(21)

[0225] Among them, the left part of (18) comes from the optimal cut; the left part of (19) comes from the feasible cut; (20) is a linear equation; and (21) is a linear inequality.

[0226] Optimal cut.

[0227]

[0228] u k , λ k ≥0

[0229] In the kth iteration, the values ​​of the non-complex variables output when solving the MILP or the first LP1 model are calculated.

[0230] u k : The Lagrange factor obtained from inequality (7) when solving the MILP or first LP1 model in the kth iteration;

[0231] λ k : The Lagrange factor obtained from equation (8) when solving the MILP or first LP1 model in the kth iteration.

[0232] It is feasible to cut.

[0233]

[0234] u k,λ1 k ,λ2 k ≥0

[0235] In the kth iteration, the values ​​of the uncomplex variables output by the second LP1 model are calculated.

[0236] u k : The Lagrange factor obtained from inequality (11) when solving the second LP1 model in the kth iteration;

[0237] λ1 k : The Lagrange factor obtained from inequality (12) when solving the second LP1 model in the kth iteration;

[0238] λ2 k : The Lagrange factor obtained from inequality (13) when solving the second LP1 model in the kth iteration.

[0239] Before implementing the technical solution of this invention, it is necessary to initialize various parameters, such as UB. 0 : Initial upper bound value, initially set to Infinity; N Conv : Number of convergence iterations, initial value is 1; Max Conv : Maximum number of convergence iterations. D: Number of segments in the y variable of the MILP model.

[0240] Based on the models described above, examples of technical solutions for each major stage are given.

[0241] 1.1: Model Transformation

[0242] If k = 1 and d = D-1 (the value of d affects the solution time and the optimality of the solution), then the original model is transformed into a MILP model; if this is not the first solution of the original model, then the solution in section 1.5 is transformed into a MILP model. Input into the original model to generate the first LP1 model;

[0243] 1.2: Solving for non-complex variables

[0244] If the original model is converted to a MILP model, the MILP model, d, preprocessed data, and the new feasible region of the variables are input into the solver (e.g., Gurobi, Cplex, COPT, etc.), and the solver's gap, time limit, and other parameters are set for solving. If the Original Problem is converted to a Primal Problem, the model, preprocessed data, and the new feasible region of the variables are input into the solver, and the solver's gap, time limit, and other parameters are set for solving. If the solver's solution is infeasible, slack variables are introduced to transform the Primal Problem into a Relaxed-Primal Problem, and the solver is used to solve it again. The above solution outputs... The Lagrange factor corresponding to the model constraints and the objective optimum Obj1 k (The optimal value of the target in the Relaxed-Primal Problem is Infeasible.)

[0245] 1.3: Upper Bound (UB, maximum value) updated

[0246] If the optimal target value output in 1.1 is Infeasible, then UB k =min(UB) k-1 (infinity); conversely, update UB. k =min(UB) k-1 Obj1 k );

[0247] 1.4: Generation of Optimal and Feasible Cuts

[0248] If the optimal value of the output objective in 1.1 is Infeasible, then the value of variable x and the Lagrange factor of the constraint are processed using the above-disclosed expression for feasible cut to obtain a feasible cut; otherwise, the value of variable x and the Lagrange factor of the constraint are processed using the above-disclosed expression for optimal cut to obtain an optimal cut.

[0249] 1.5: Solving Complex Variables

[0250] Add the feasible or optimal cut from section 1.4 to the LP2 model above, and input the model, preprocessed data, and the new feasible domain of the variables into the solver, setting the solution parameters to perform the solution. Output Target optimal value Obj2 k and Obj2 k Store in LB. LB = {Obj2} 1 Obj2 2 , ...,Obj2 k If Obj2k =Obj2 k-1 Then N Conv =N Conv +1.

[0251] 1.6: Loop Stop Detection

[0252] There are three criteria for determining whether an algorithm has stopped, and one or more of them can be used for the determination.

[0253] (1) If min(LB)≥UB k Then the iteration ends and UB is output. k For the optimal value, output This is the optimal solution;

[0254] (2) If min(LB) < UB k N Conv ≥Max Conv Then the iteration ends and UB is output. k For the optimal value, output This is the optimal solution;

[0255] (3) If min(LB) < UB k If the iteration continues, jump to 1.1 and input... And update k = k + 1. Based on the same inventive concept, see [link to relevant documentation]. Figure 3 The following embodiments illustrate an optimized system for mixing and stirring chemical products, including:

[0256] The first determining module 31 is used to determine the original complex variables and the original non-complex variables from the relevant parameter variables of the mixing tank; wherein, the relevant parameter variables of the mixing tank include: pipeline flow rate parameter variables, raw material parameter variables, intermediate product parameter variables, and mixing tank mixing capacity parameter variables;

[0257] The second determining module 32 is used to determine the original model based on the original complex variables and the original non-complex variables; wherein, the original model includes the original objective function and the original constraints, and the original objective function and / or the original constraints contain the original complex variables and the original non-complex variables;

[0258] The conversion module 33 is used to convert the original model into a mixed integer linear programming (MILP) model for solving if the original model is being solved for the first time.

[0259] The first module 34 is used to obtain the corresponding first model output solution if the MILP model is successfully solved.

[0260] The second obtaining module 35 is used to obtain the optimal slicing of the model based on the output solution of the first model;

[0261] The third module 36 is used to optimally insert the model into the LP2 model for solving, and obtain the output solution of the second model.

[0262] The judgment module 37 is used to make a judgment based on the output solution of the first model and the output solution of the second model;

[0263] The output module 38 is used to output the first model output solution and the second model output solution as the corresponding optimal solutions if the judgment is successful.

[0264] Based on the same inventive concept as in the foregoing embodiments, this embodiment of the invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the methods described above.

[0265] Based on the same inventive concept as in the foregoing embodiments, this embodiment of the invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of any of the methods described above.

[0266] The algorithms and displays provided herein are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used in conjunction with the teachings herein. The required structure for constructing such systems is apparent from the above description. Furthermore, this invention is not directed to any particular programming language. It should be understood that the contents of the invention described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing the best mode of implementation of the invention.

[0267] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0268] Similarly, it should be understood that, in order to simplify this disclosure and aid in understanding one or more of the various aspects of the invention, in the above description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof. However, this method of disclosure should not be construed as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim. Rather, as reflected in the following claims, inventive aspects lie in fewer than all features of a single foregoing disclosed embodiment. Therefore, the claims following the detailed description are hereby expressly incorporated into this detailed description, wherein each claim itself is a separate embodiment of the invention.

[0269] Those skilled in the art will understand that modules in the device of the embodiments can be adaptively changed and placed in one or more devices different from that embodiment. Modules, units, or components in the embodiments can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components. Except where at least some of such features and / or processes or units are mutually exclusive, any combination can be used to combine all features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all processes or units of any method or device so disclosed. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose.

[0270] Furthermore, those skilled in the art will understand that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are intended to be within the scope of the invention and form different embodiments. For example, in the following claims, any of the claimed embodiments can be used in any combination.

[0271] The various component embodiments of the present invention can be implemented in hardware, or as software modules running on one or more processors, or a combination thereof. Those skilled in the art will understand that microprocessors or digital signal processors (DSPs) can be used in practice to implement some or all of the functions of some or all of the components of the gateway, proxy server, or system according to embodiments of the present invention. The present invention can also be implemented as a device or apparatus program (e.g., a computer program and computer program product) for performing some or all of the methods described herein. Such programs implementing the present invention can be stored on a computer-readable medium or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0272] It should be noted that the above embodiments are illustrative of the invention and not restrictive, and that those skilled in the art can devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names.

Claims

1. An optimized method for mixing and stirring chemical products, characterized in that, The method includes: The original complex variable and the original uncomplex variable are determined from the relevant parameter variables of the mixing tank; wherein, the relevant parameter variables of the mixing tank include: pipeline flow rate parameter variable, raw material parameter variable, intermediate product parameter variable, and mixing tank mixing capacity parameter variable; the original complex variable and the original uncomplex variable are derived from the bilinear term formed by multiplying two variables, and one variable in the bilinear term is randomly determined as the original complex variable, and the other variable is determined as the original uncomplex variable; The original model is determined based on the original complex variables and the original non-complex variables; wherein, the original model includes an original objective function and original constraints, and the original objective function and / or the original constraints contain original complex variables and original non-complex variables; the original model is used to determine the optimal solution of the original objective function under the constraints of the original constraints; If the original model is being solved for the first time, it is transformed into a Mixed Integer Linear Programming (MILP) model for solution. Specifically, the original objective function of the original model is used as the objective function of the MILP model. New feasible regions for the original complex variables and the original uncomplex variables are obtained. The constraints of the MILP model are determined using these new feasible regions, and the solution is then performed based on the objective function and constraints of the MILP model. If the MILP model is solved successfully, the corresponding first model output solution is obtained; The optimal slicing of the model is obtained based on the output solution of the first model. The optimal cut of the model is placed into the linear optimization model LP2 for solution, and the output solution of the second model is obtained. The LP2 model is a linear optimization model after fixing the original uncomplex variables, and the optimal cut of the model is placed into the constraint conditions of the LP2 model for solution. The judgment is made based on the output solutions of the first model and the output solutions of the second model; If the judgment is successful, the output solutions of the first model and the second model will be output as the corresponding optimal solutions for the variables.

2. The method as described in claim 1, characterized in that, The first model output solution includes: values ​​of uncomplex variables, Lagrange factors, and the optimal value of the objective function of the first model; the second model output solution includes: values ​​of complex variables and the optimal value of the objective function of the second model. The step of obtaining the optimal slicing of the model based on the output solution of the first model specifically includes: The optimal cut of the model is obtained based on the uncomplex variable values ​​of the first model output solution, the Lagrange factor of the first model output solution, the constraints of the MILP model, and the objective function of the MILP model. The step of optimally inserting the model into the LP2 model for solution to obtain the output solution of the second model specifically includes: The optimal model is then placed into the constraints of the LP2 model for solution, and the output solution of the second model is obtained. The judgment based on the output solutions of the first model and the second model specifically includes: The optimal values ​​of the first model objective function and the second model objective function are compared. If the optimal value of the first model objective function is less than or equal to the optimal value of the second model objective function, the comparison is successful.

3. The method as described in claim 2, characterized in that, After determining the solution based on the first model output solution and the second model output solution, the method further includes: If the judgment fails, the original complex variable in the original model is fixed according to the complex variable value in the solution output by the second model, so as to transform the original model into the first LP1 model for solution; wherein, the first LP1 model is a linear optimization model after fixing the original complex variable, and the objective function and / or constraint conditions of the first LP1 model are only related to the original non-complex variable; Determine whether the first LP1 model has a feasible solution; If a feasible solution exists, the optimal cut is obtained by solving the first output solution based on the first LP1 model. The optimal solution of the first LP1 model is incorporated into the LP2 model for solution to obtain the second output solution; The determination is made based on the first output solution and the second output solution; If the judgment is successful, the first output solution and the second output solution will be output as the corresponding optimal solutions for the variables; If the judgment fails, the original complex variable in the original model is fixed according to the complex variable value in the second output solution, and the above steps are repeated until the judgment is successful according to the first output solution and the second output solution.

4. The method as described in claim 3, characterized in that, The first output solution includes: values ​​of uncomplex variables, Lagrange factors, and the optimal value of the first objective function; the second output solution includes: values ​​of complex variables and the optimal value of the second objective function. The step of obtaining the optimal cut based on the first output solution specifically includes: The optimal cut is obtained based on the uncomplex variable values ​​of the first output solution, the Lagrange factor of the first output solution, the constraints of the first LP1 model, and the objective function of the first LP1 model. The step of inserting the first output solution and its optimal value into the LP2 model for solving to obtain the second output solution specifically includes: The optimal cut is incorporated into the constraints of the LP2 model and solved to obtain the second output solution; The determination based on the first output solution and the second output solution specifically includes: Determine the magnitudes of the optimal values ​​of the first and second objective functions. If the optimal value of the first objective function is less than or equal to the optimal value of the second objective function, the determination is successful; and / or If the optimal value of the second objective function remains consistent throughout N consecutive iterations, the judgment is successful; where N≥3 and is a positive integer.

5. The method as described in claim 3, characterized in that, After determining whether the first LP1 model has a feasible solution, the method further includes: If no feasible solution is found, a relaxation variable is obtained, and the objective function and constraints of the first LP1 model are updated according to the relaxation variable to obtain a second LP1 model. The second LP1 model is then solved to obtain a third output solution. The second LP1 model is a relaxed linear optimization model with the original complex variables fixed. A feasible cut is obtained based on the third output solution; The third output solution and its feasibility are inserted into the LP2 model for solving to obtain the fourth output solution; The determination is made based on the third and fourth output solutions; If the judgment is successful, the third output solution and the fourth output solution will be output as the corresponding optimal solutions for the variables. If the judgment fails, the original complex variable in the original model is fixed according to the value of the complex variable in the fourth output solution, and the process is switched to the step of converting the original model into the first LP1 model for solving to obtain the first output solution.

6. The method as described in claim 5, characterized in that, The third output solution includes: non-complex variable values, Lagrange factors, and the maximum value of the third output solution; the fourth output solution includes: complex variable values ​​and the minimum value of the fourth output solution. The step of obtaining a feasible cut based on the third output solution specifically includes: The feasible cut is obtained based on the Lagrange factor of the third output solution and the constraints of the second LP1 model; The step of inserting the feasible solution corresponding to the third output solution into the LP2 model for solving to obtain the fourth output solution specifically includes: The feasible solution is then applied to the constraints in the LP2 model to obtain the fourth output solution. The determination based on the third output solution and the fourth output solution specifically includes: Determine the magnitude of the maximum value of the third output solution and the minimum value of the fourth output solution. If the maximum value of the third output solution is less than or equal to the minimum value of the fourth output solution, the determination is successful; and / or If the minimum value of the fourth output solution remains consistent across N consecutive loops, the judgment is successful; where N≥3 and is a positive integer.

7. An optimized system for mixing and stirring chemical products, characterized in that, include: The first determining module is used to determine the original complex variable and the original uncomplex variable from the relevant parameter variables of the mixing tank; wherein, the relevant parameter variables of the mixing tank include: pipeline flow rate parameter variable, raw material parameter variable, intermediate product parameter variable, and mixing tank mixing capacity parameter variable; the original complex variable and the original uncomplex variable are derived from a bilinear term formed by multiplying two variables, and one variable in the bilinear term is randomly determined as the original complex variable, and the other variable is determined as the original uncomplex variable; The second determining module is used to determine the original model based on the original complex variables and the original non-complex variables; wherein, the original model includes an original objective function and original constraints, and the original objective function and / or the original constraints contain original complex variables and original non-complex variables; the original model is used to determine the optimal solution of the original objective function under the constraints of the original constraints; The transformation module is used to transform the original model into a Mixed Integer Linear Programming (MILP) model for solving if the original model is being solved for the first time. Specifically, the original objective function in the original model is used as the objective function of the MILP model; new feasible regions for the original complex variables and the original non-complex variables are obtained; the constraints of the MILP model are determined using the new feasible regions of the original complex variables and the original non-complex variables; and the solution is performed based on the objective function and constraints of the MILP model. The first module is used to obtain the corresponding first model output solution if the MILP model is solved successfully. The second obtaining module is used to obtain the optimal slicing of the model based on the output solution of the first model. The third module is used to input the optimal cut of the model into the LP2 model for solving, and obtain the output solution of the second model; wherein, the LP2 model is a linear optimization model after fixing the original uncomplex variables, and the optimal cut of the model is used as a new constraint condition to input the constraint condition of the LP2 model for solving. The judgment module is used to make a judgment based on the output solution of the first model and the output solution of the second model; The output module is used to output the first model's output solution and the second model's output solution as the corresponding optimal solutions if the judgment is successful.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1-6.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method according to any one of claims 1-6.

Images