Unit commitment accelerated optimization method and system for eliminating redundant constraint based on machine learning, storage medium, electronic device, chip, program product, and program

Abstract

Disclosed in the present invention are a unit commitment accelerated optimization method and system for eliminating a redundant constraint based on machine learning. The method comprises: converting an original unit commitment problem into a line power flow constraint-free problem, and using variables 0-1 to control the establishment of the line power flow constraint, so as to eliminate dense and complex security constraints. Moreover, the change frequency range of each unit in each time period is learned by means of a machine learning method, and a redundant minimum continuous on / off time constraint identification rule is formed, so that a redundant minimum continuous on / off time constraint is eliminated, thereby reducing the number of problem constraints and the problem scale, and improving the unit commitment optimization efficiency.

Description

Machine learning-based methods and systems for accelerating unit configuration optimization to eliminate redundancy constraints, including storage media, electronic devices, chips, software products, and programs.

[0001] Cross-references to related applications

[0002] This application is based on and claims priority to Chinese Patent Application No. 202410314469.6, filed on March 19, 2024, entitled “Method and System for Accelerating Unit Combination Optimization Based on Machine Learning to Eliminate Redundancy Constraints”, the entire contents of which are incorporated herein by reference. Technical Field

[0003] This invention belongs to the field of machine learning and relates to a method and system for accelerating the optimization of unit combination based on machine learning to eliminate redundancy constraints, as well as storage media, electronic devices, chips, program products, and programs. Background Technology

[0004] Complex and dense safety constraints and minimum consecutive start-stop time constraints are among the key factors affecting the solution efficiency of large-scale unit combinations. In fact, some of these constraints are invalid or redundant; removing them from the original model will not affect the overall solution domain, but adding these constraints makes the problem difficult to solve. Reducing safety power flow constraints and minimum consecutive start-stop time constraints can greatly improve the overall solution efficiency of the model. However, accurately identifying the characteristics of these constraints and eliminating redundant constraints is a challenge.

[0005] Summary of the Invention

[0006] The purpose of this invention is to solve the problems in the prior art of being unable to accurately identify the characteristics of safety power flow constraints and minimum continuous start-stop time constraints and to eliminate redundant constraints, and to provide a method and system for accelerating the optimization of unit combination based on machine learning to eliminate redundant constraints, as well as storage media, electronic devices, chips, program products, and programs.

[0007] To achieve the above objectives, the present invention employs the following technical solution:

[0008] Machine learning-based methods for accelerating unit combination optimization to eliminate redundancy constraints include:

[0009] Obtain the semi-continuous constraints and line power flow constraints of the upper and lower limits of unit output in the SCUC model.

[0010] If the semi-continuous constraints on the upper and lower limits of unit output and the line power flow constraints both hold in the SCUC model, then the upper and lower limits of the two constraints are very close, thus introducing a subproblem containing only 0-1 variables and solving the subproblem containing only 0-1 variables.

[0011] The solution of the subproblem containing only 0-1 variables is added to the objective function of the original unit combination problem based on the penalty method, so that the solution of the subproblem is as close as possible to the optimal solution of the original unit combination problem.

[0012] Based on the solutions to the subproblems and the optimal solutions to the original unit combination problem, the offset of the unit state variables of each unit in each time period is obtained; the characteristic demand D of the original unit combination problem is changed to obtain multiple optimal solutions, the relationship between demand D and offset is obtained, and a predictive regression model is learned.

[0013] Given a new unit combination problem, a learning-based prediction model predicts the offset of the unit state variables of each unit in each time period of the new unit combination problem to be solved, identifies and removes redundant constraints in the minimum continuous start-up and shutdown time constraint, so as to accelerate the optimization of unit combination.

[0014] A further improvement of the present invention is that:

[0015] Furthermore, the semi-continuous constraints on the upper and lower limits of the unit output of the SCUC model are as follows:

[0016] Among them, I i,t These are the start-up and shutdown status variables of the generator unit. and These represent the minimum and maximum output of the unit, respectively.

[0017] Furthermore, the line power flow constraints are as follows:

[0018] Among them, GSF i,l The transfer distribution factor represents the proportion of power transferred from unit i to line l.

[0019] Furthermore, a subproblem containing only 0-1 variables is introduced and solved. Specifically, if the line power flow constraint holds, the upper and lower limits of the line power flow constraint are very close to the upper and lower limits of the semi-continuous constraint transformation form. A subproblem containing only 0-1 variables is constructed to solve for the corresponding 0-1 variable values. The objective of this subproblem with 0-1 variables is to minimize the distance between the upper and lower limits of the semi-continuous constraint and the upper and lower limits of the line power flow constraint, and the closer the upper and lower limits of the semi-continuous constraint are, the better.

[0020] The semi-continuous constraint is transformed into the same form as the line power flow constraint, as shown in formula (3):

[0021] Construct a subproblem containing only 0-1 variables:

[0022] Among them, the constraint term is that the unit combination problem only contains I.i,t W i,t and Y i,t The constraint H(I) ≥ 0; W i,t and Y i,t These are the decision variables for whether the unit is turned on and off, respectively.

[0023] Transform the above objective function into one containing only I. i,t The linear expression L(I) i,t Solve such subproblems to obtain solutions S with 0-1 variables;

[0024] Furthermore, based on a penalty approach, the solutions to subproblems containing only 0-1 variables are added to the objective function of the original unit combination problem, so that the solutions to the subproblems are as close as possible to the optimal solution of the unit combination problem. Specifically:

[0025] The constraints are G(I,P,W)≥0 and H(I)≥0, expressing the unit combination problem as a set of constraints including unit start-up / shutdown decision variable I, unit output variable P, and other variables w. G(I,P,W) represents the constraint set without line power flow constraints; H(I)≥0 indicates that the unit combination problem only contains I. i,t W i,t Y i,t Constraints on 0-1 variables.

[0026] Furthermore, the minimum continuous start-stop time constraint is as follows:

[0027] Furthermore, based on the solutions to the subproblems and the optimal solutions to the original unit combination problem, the offset of the unit state variables for each unit at each time period is obtained. The characteristic demand D of the original unit combination problem is changed to obtain multiple optimal solutions, and the relationship between demand D and the offsets is obtained and learned into a predictive regression model. Specifically, by changing demand D, the unit combination problem is solved to obtain several optimal solutions I. * Calculate each I * The distance to the solution S of the subproblem containing only 0-1 variables; the degree of offset of each unit in each time period is statistically analyzed to form the relationship between demand D and offset vector v; the SCUC model determined by demand d∈D has a functional relationship. Make

[0028] Among them, v j Each component of the offset degree v for each unit in each time period.

[0029] Furthermore, redundant constraints in the minimum continuous start-up and shutdown time constraint are identified and removed. Specifically, given a new unit combination problem, a learned prediction model predicts the offset of the unit state variables of each unit in each time period for the new unit combination problem to be solved.

[0030] or

[0031] Make a judgment, g(v) i ) represents the function expression of the offset. If formulas (10) and (11) still hold, then the original constraint is redundant and should be removed.

[0032] A machine learning-based system for accelerating and optimizing unit composition by eliminating redundancy constraints includes:

[0033] The first acquisition module acquires the semi-continuous constraints of the upper and lower limits of the unit output of the SCUC model and the line power flow constraints.

[0034] The module introduces a subproblem containing only 0-1 variables. If the semi-continuous constraints of the unit output upper and lower limits and the line power flow constraints in the SCUC model are both valid, then the upper and lower limits of the two constraints are very close. The module then introduces a subproblem containing only 0-1 variables and solves the subproblem containing only 0-1 variables.

[0035] The second acquisition module adds the solution of the subproblem containing only 0-1 variables to the objective function of the original unit combination problem based on a penalty method, so that the solution of the subproblem is as close as possible to the optimal solution of the original unit combination problem.

[0036] The offset data acquisition module obtains the offset of the unit state variable of each unit in each time period based on the solution of the sub-problem and the optimal solution of the original unit combination problem; changes the characteristic demand D of the original unit combination problem, obtains multiple optimal solutions, obtains the relationship between demand D and offset, and learns it into a predictive regression model.

[0037] The judgment module, given a new unit combination problem, predicts the offset of the unit state variable of each unit in the new unit combination problem to be solved in each time period based on the learned prediction model, identifies and removes redundant constraints in the minimum continuous start-stop time constraint, so as to realize accelerated optimization of unit combination.

[0038] A terminal device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the steps of the method described above.

[0039] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method described above.

[0040] A chip comprising:

[0041] Memory, used to store computer programs;

[0042] A processor, connected to the memory, is used to retrieve and run a computer program from the memory, causing a device equipped with the chip to perform the steps of the method described above.

[0043] A computer program product includes computer program instructions that cause a computer to perform the steps of the method described above.

[0044] A computer program that, when run on a computer, causes the computer to perform the steps of the method described above.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] This invention transforms the unit combination problem into a problem without line power flow constraints, using 0-1 variables to control the validity of line power flow constraints, thereby eliminating dense and complex safety constraints. Furthermore, it employs machine learning to learn the range of changes for each unit across all time periods, forming redundant minimum consecutive start-stop time constraint identification rules. This eliminates redundant minimum consecutive start-stop time constraints, reducing the number of constraints and the problem size, and improving the efficiency of unit combination optimization. Attached Figure Description

[0047] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 is a flowchart illustrating the unit combination acceleration optimization method based on machine learning to eliminate redundancy constraints according to the present invention.

[0049] Figure 2 is a schematic diagram of the unit combination acceleration optimization system based on machine learning to eliminate redundancy constraints according to the present invention.

[0050] Figure 3 is a schematic diagram of the unit combination acceleration optimization algorithm based on machine learning to eliminate redundancy constraints according to the present invention.

[0051] Figure 4 is a schematic structural diagram of a chip 400 according to an embodiment of the present invention. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0053] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0054] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0055] In the description of the embodiments of the present invention, it should be noted that if terms such as "upper," "lower," "horizontal," or "inner" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of the invention is in use, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. Furthermore, terms such as "first" and "second" are only used to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0056] Furthermore, the use of the term "horizontal" does not imply that the component must be absolutely horizontal, but rather that it can be slightly tilted. For example, "horizontal" simply means that its direction is more horizontal than "vertical," and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0057] In the description of the embodiments of the present invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in the present invention according to the specific circumstances.

[0058] The present invention will now be described in further detail with reference to the accompanying drawings:

[0059] Referring to Figure 1, this invention discloses a method for accelerating the optimization of unit combination based on machine learning to eliminate redundancy constraints, including:

[0060] S101, obtain the semi-continuous constraints and line power flow constraints of the upper and lower limits of the unit output of the SCUC model.

[0061] The semi-continuous constraints on the upper and lower limits of the unit output of the SCUC model are as follows:

[0062] Among them, I i,t These are the start-up and shutdown status variables of the generator unit. and These represent the minimum and maximum output of the unit, respectively.

[0063] The power flow constraints of the line are as follows:

[0064] Among them, GSF i,l The transfer distribution factor represents the proportion of power transferred from unit i to line l.

[0065] S102, If the semi-continuous constraints of the unit output upper and lower limits and the line power flow constraints in the SCUC model are both valid, then the upper and lower limits of the two constraints are very close, and a subproblem containing only 0-1 variables is introduced and the subproblem containing only 0-1 variables is solved.

[0066] If the line flow constraint holds, then the upper and lower limits of the line flow constraint are very close to the upper and lower limits of the semi-continuous constraint transformation form. Construct a subproblem containing only 0-1 variables to solve for the corresponding 0-1 variable values. The goal of this subproblem with 0-1 variables is to minimize the distance between the upper and lower limits of the semi-continuous constraint and the upper and lower limits of the line flow constraint, and the closer the upper and lower limits of the semi-continuous constraint are, the better.

[0067] The semi-continuous constraint is transformed into the same form as the line power flow constraint, as shown in formula (3):

[0068] Construct a subproblem containing only 0-1 variables:

[0069] Among them, the constraint term is that the unit combination problem only contains I. i,t W i,t and Y i,t The constraint H(I) ≥ 0; W i,t and Y i,t These are the decision variables for whether the unit is turned on and off, respectively.

[0070] Transform the above objective function into one containing only I. i,t The linear expression L(I) i,tSolve such subproblems to obtain solutions S with 0-1 variables;

[0071] S103, based on a penalty approach, adds the solutions to subproblems containing only 0-1 variables to the objective function of the original unit combination problem, making the solutions to the subproblems as close as possible to the optimal solution of the original unit combination problem. Specifically:

[0072] The constraints are G(I,P,W)≥0 and H(I)≥0, expressing the unit combination problem as a set of constraints including unit start-up / shutdown decision variable I, unit output variable P, and other variables w. G(I,P,W) represents the constraint set without line power flow constraints; H(I)≥0 indicates that the unit combination problem only contains I. i,t W i,t Y i,t Constraints on 0-1 variables.

[0073] S104: Based on the solutions to the subproblems and the optimal solutions to the original unit combination problem, obtain the offset of the unit state variables of each unit in each time period; change the characteristic demand D of the original unit combination problem, obtain multiple optimal solutions, obtain the relationship between demand D and offset, and learn it into a predictive regression model.

[0074] By changing demand D, we solve the unit combination problem and obtain several optimal solutions I. * Calculate each I * The distance to the solution S containing only 0-1 variables; the degree of offset of each unit in each time period is statistically analyzed to form the relationship between demand D and offset vector v; the SCUC model determined by demand d∈D has a functional relationship. Make

[0075] Among them, v j Each component of the offset degree v for each unit in each time period.

[0076] S105. Given a new unit combination problem, a learning-based prediction model predicts the offset of the unit state variables of each unit in the new unit combination problem to be solved at each time period, identifies and removes redundant constraints in the minimum continuous start-up and shutdown time constraint, so as to achieve accelerated optimization of unit combination.

[0077] The minimum consecutive start-stop time constraint is as follows:

[0078] To determine whether it is a minimum consecutive start-stop time constraint, specifically: for the minimum consecutive start-stop time constraint...

[0079] or

[0080] Make a judgment, g(v) i ) represents the function expression of the offset. If formulas (10) and (11) still hold, then the original constraint is redundant and should be removed.

[0081] Referring to Figure 2, this invention discloses a unit combination acceleration optimization system based on machine learning to eliminate redundancy constraints, comprising:

[0082] The first acquisition module acquires the semi-continuous constraints of the upper and lower limits of the unit output of the SCUC model and the line power flow constraints.

[0083] The module introduces a subproblem containing only 0-1 variables. If the semi-continuous constraints of the unit output upper and lower limits and the line power flow constraints in the SCUC model are both valid, then the upper and lower limits of the two constraints are very close. The module then introduces a subproblem containing only 0-1 variables and solves the subproblem containing only 0-1 variables.

[0084] The second acquisition module adds the solution of the subproblem containing only 0-1 variables to the objective function of the original unit combination problem based on a penalty method, so that the solution of the subproblem is as close as possible to the optimal solution of the original unit combination problem.

[0085] The offset data acquisition module obtains the offset of the unit state variable of each unit in each time period based on the solution of the sub-problem and the optimal solution of the original unit combination problem; changes the characteristic demand D of the original unit combination problem, obtains multiple optimal solutions, obtains the relationship between demand D and offset, and learns it into a predictive regression model.

[0086] The judgment module, given a new unit combination problem, predicts the offset of the unit state variable of each unit in the new unit combination problem to be solved in each time period based on the learned prediction model, identifies and removes redundant constraints in the minimum continuous start-stop time constraint, so as to realize accelerated optimization of unit combination.

[0087] Example:

[0088] Referring to Figure 3, this invention discloses a unit combination acceleration optimization algorithm based on machine learning to eliminate redundancy constraints, including removing line power flow constraints and removing redundancy minimum continuous start-stop time constraints.

[0089] The first part transforms line power flow constraints into 0-1 variable values ​​by representing them with 0-1 variables. These values ​​are then added as penalties to the objective function of the original problem, ensuring it recovers to the original optimal solution without straying too far from it. The second part uses the solution S from the first part as a baseline. Machine learning methods are used to learn the offset of each unit's variable value in S at time 96 from the corresponding optimal solution. This offset is then used to identify which minimum consecutive start-stop time constraints are redundant and do not significantly affect the solution domain boundary. Once identified, these constraints can be removed, further reducing the problem size and accelerating the process.

[0090] Based on the relationship between safety constraints and upper / lower limits of unit output constraints in hydropower and thermal power problems, the first stage transforms the unit combination problem into a problem without line power flow constraints using traditional optimization methods. 0-1 variables are used to control the validity of line power flow constraints, thereby eliminating dense and complex safety constraints. Building on this, the second stage uses machine learning methods to learn the range of changes for each unit across all time periods, forming a redundant minimum consecutive start-stop time constraint identification rule, thus eliminating redundant minimum consecutive start-stop time constraints.

[0091] (1) Reconstructing line power flow constraints using 0-1 variables

[0092] Because the SCUC model has semi-continuous constraints such as upper and lower limits of unit output, as shown in formula (1):

[0093] Among them, I i,t These are the start-up and shutdown status variables of the generator unit. and These represent the minimum and maximum output of the unit, respectively.

[0094] The power flow constraint of the line is as shown in formula (2):

[0095] Among them, GSF i,l The transfer distribution factor represents the proportion of power transferred from unit i to line l.

[0096] If the semi-continuous constraint holds, it is transformed into the same form as the line power flow constraint, as shown in formula (3):

[0097] If the line power flow constraints hold, then their upper and lower bounds are very close to those of the semi-continuous constraints. Therefore, the overall problem doesn't need to incorporate dense and complex line power flow constraints. Instead, a subproblem containing only 0-1 variables is constructed to solve for their values. The objective of this subproblem is to minimize the distance between the upper and lower bounds of the semi-continuous constraints and the line power flow constraints, with the latter being as close as possible. In other words, the line power flow constraints are made valid through the values ​​of these 0-1 variables. Furthermore, the solutions to these 0-1 variables can be penalized and added to the objective function of the original problem. This allows the solutions to the subproblem to be as close as possible to the true optimal solution, thus fixing variables during the solution process, accelerating the solution efficiency, and ensuring the line power flow constraints hold.

[0098] The steps for constructing the subproblem are as follows:

[0099] Step 1 (Constructing small-scale subproblems):

[0100] Objective function of unit combination problem:

[0101] Constraint: The unit combination problem only contains I i,t W i,t Y i,t The constraint H(I) ≥ 0; where W i,t and Y i,t These are the decision variables for whether the unit is turned on and off, respectively.

[0102] Transform the above objective function into one containing only I. i,t The linear expression L(I) i,t Solve such subproblems to obtain solutions S with 0-1 variables;

[0103] Transform into only I i,t The linear expression L(I) i,t Solve such subproblems to obtain solutions S for 0-1 variables.

[0104] Step 2 (Solve the original problem with the penalty term added)

[0105] The solution to the 0-1 variable subproblem is added to the objective function of the unit combination problem based on a penalty approach, specifically:

[0106] Constraints: G(I,P,W)≥0, H(I)≥0, other constraints

[0107] The unit combination problem is expressed as a set of constraints including unit start-up / shutdown decision variables I, unit output variables P, and other variables w. G(I,P,W) represents the constraint set without line power flow constraints. H(I)≥0 indicates that the original problem only contains I.i,t W i,t Y i,t Constraints with 0-1 variables. This penalizes the distance between the solution with 0-1 variables and the solution to the subproblem, essentially repairing the subproblem's solution to a feasible one while preserving the possibility of the power flow constraints holding. This is because both the subproblem and the original problem contain numerous constraints related to 0-1 variables, and the feasible solutions to the original problem and the subproblem are not significantly different. Solving the modified problem thus achieves accelerated solution.

[0108] (2) Use machine learning methods to predict the offset of minimum consecutive start-stop time.

[0109] By employing the aforementioned method of constructing sub-problems in stages, the problem of the efficiency being affected by dense and complex line power flow constraints is resolved. There are many ways to express the minimum continuous start-stop time constraint; according to this scheme, its form is shown in formulas (7) and (8).

[0110] The algorithm obtains the solution S of 0-1 variables formed by some constraints in the unit combination problem. This solution S is still some distance from the true optimal solution, and is used to measure the change in the true optimal solution. Specifically, it observes the offset between the true optimal solution and solution S. Moreover, it only needs to observe the offset of each unit in 96 time periods, and predict the offset of the unit's state variables in 96 time periods using the information of the true solution and the deviation of this solution S. Solution S is constant and can be solved offline as a benchmark. After predicting the offset of the unit's state variables in 96 time periods, the offset is subtracted or added to the value of the constraint expression of the minimum continuous start-stop time constraint of the 0-1 variable subproblem to see if it exceeds the right-hand side of the expression. If it does not exceed the right-hand side, it means that the expression does not affect the overall optimality and is a redundant constraint, which can be removed. Therefore, the steps of the algorithm are as follows:

[0111] Step 1: Construct a subproblem formed by constraints containing only 0-1 variables in the original problem, as shown in (1). The modified form of the line power flow constraint is used as the objective function, and the solution S is obtained.

[0112] Step 2: Change requirement D, and obtain multiple optimal solutions I by solving the original problem. * Calculate each I * The distance from S. For example, I *= (0,1,1,0), S = (0,0,0,0), and the distance s is calculated as 2. Here, the dimension of the distance vector is the number of aircraft units. The offset (distance) of each aircraft unit at its own 96 time points is statistically analyzed to form the relationship between D and the offset vector v, which is then learned into a prediction model. The prediction model can be used to make predictions using a regression model.

[0113] For the SCUC model determined by the demand parameter d∈D, the following functional relationship exists. Make

[0114] Among them, v j For each unit, each component represents the degree of offset v at its own 96 moments.

[0115] Given a new unit combination problem, a learning-based prediction model predicts the offset of the unit state variables for each unit in each time period in the new unit combination problem to be solved.

[0116] For the minimum continuous start-stop time constraint

[0117] or

[0118] Make a judgment, g(v) i ) represents the function expression of the offset. If formulas (10) and (11) still hold, then the original constraint is redundant. Therefore, the constraint is removed, thereby reducing the problem size and improving the solution efficiency.

[0119] Performance testing:

[0120] An adjustment example based on standard examples (IEEE 30, IEEE 118, WP 2383) is adopted, which contains many complex power flow constraints.

[0121] Table 1 Comparison of Problem Size

[0122] Table 2 Comparison of SCIP solution efficiency

[0123] The analysis of the above results shows that this invention, by reducing the dense, complex, and numerous line flow constraints and minimum continuous start-stop time constraints, can significantly reduce the problem size, thereby accelerating the solution of the unit combination problem and resulting in a significant improvement in solution efficiency, especially in large-scale models. The solution efficiency is generally improved by about 1.7 to 2 times. The number of line flow constraint violations is zero, indicating the effectiveness and reliability of this invention, without significantly altering the solution domain. This demonstrates the innovation and effectiveness of this invention, and it also provides a good acceleration effect for the relatively slow SCIP problem.

[0124] An embodiment of the present invention provides a terminal device. This terminal device includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps in the various method embodiments described above. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the various device embodiments described above.

[0125] The computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention.

[0126] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0127] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0128] The memory can be used to store the computer program and / or module. The processor implements various functions of the terminal device by running or executing the computer program and / or module stored in the memory and calling the data stored in the memory.

[0129] If the modules / units integrated in the terminal device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0130] An embodiment of the present invention provides a chip. FIG4 is a schematic structural diagram of a chip provided in an embodiment of the present invention. The chip 400 shown in FIG4 includes a processor 410, which can call and run computer programs from memory to implement the method in the embodiment of the present invention.

[0131] In some embodiments, as shown in FIG4, chip 400 may further include memory 420. Processor 410 may retrieve and run computer programs from memory 420 to implement the methods in the embodiments of the present invention.

[0132] The memory 420 can be a separate device independent of the processor 410, or it can be integrated into the processor 410.

[0133] In some embodiments, the chip 400 may further include an input interface 430. The processor 410 can control the input interface 430 to communicate with other devices or chips; specifically, it can acquire information or data sent by other devices or chips.

[0134] In some embodiments, the chip 400 may further include an output interface 440. The processor 410 can control the output interface 440 to communicate with other devices or chips; specifically, it can output information or data to other devices or chips.

[0135] In some embodiments, the chip can be applied to the electronic device in the embodiments of the present invention, and the chip can implement the corresponding processes implemented by the electronic device in the various methods of the embodiments of the present invention. For the sake of brevity, it will not be described in detail here.

[0136] It should be understood that the chip mentioned in the embodiments of the present invention may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0137] This invention also provides a computer program product, including computer program instructions.

[0138] Optionally, the computer program product can be applied to the electronic device in the embodiments of the present invention, and the computer program instructions cause the computer to execute the corresponding processes implemented by the electronic device in the various methods of the embodiments of the present invention. For the sake of brevity, they will not be described in detail here.

[0139] This invention also provides a computer program.

[0140] Optionally, the computer program can be applied to the electronic device in the embodiments of the present invention. When the computer program is run on a computer, it causes the computer to execute the corresponding processes implemented by the electronic device in the various methods of the embodiments of the present invention. For the sake of brevity, it will not be described in detail here.

[0141] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A machine learning-based method for accelerating unit combination optimization to eliminate redundancy constraints, including: Obtain the semi-continuous constraints and line power flow constraints of the upper and lower limits of unit output in the SCUC model. If the semi-continuous constraints on the upper and lower limits of unit output and the line power flow constraints both hold in the SCUC model, then the upper and lower limits of the two constraints are very close, thus introducing a subproblem containing only 0-1 variables and solving the subproblem containing only 0-1 variables. The solution of the subproblem containing only 0-1 variables is added to the objective function of the original unit combination problem based on the penalty method, so that the solution of the subproblem is as close as possible to the optimal solution of the original unit combination problem. Based on the solutions to the subproblems and the optimal solutions to the original unit combination problem, the offset of the unit state variables of each unit in each time period is obtained; the characteristic demand D of the original unit combination problem is changed to obtain multiple optimal solutions, the relationship between demand D and offset is obtained, and a predictive regression model is learned. Given a new unit combination problem, a learning-based prediction model predicts the offset of the unit state variables of each unit in each time period of the new unit combination problem to be solved, identifies and removes redundant constraints in the minimum continuous start-up and shutdown time constraint, so as to accelerate the optimization of unit combination.

2. The method for accelerating unit combination optimization based on machine learning to eliminate redundancy constraints according to claim 1, wherein, The semi-continuous constraints on the upper and lower limits of the unit output of the SCUC model are as follows: Among them, I i,t P is the start-stop state variable of the unit. i L and P i U These represent the minimum and maximum output of the unit, respectively.

3. The method for accelerating unit combination optimization based on machine learning to eliminate redundancy constraints according to claim 2, wherein, The power flow constraints of the line are specifically as follows: Among them, GSF i,l The transfer distribution factor represents the proportion of power transferred from unit i to line l.

4. The method for accelerating unit combination optimization based on machine learning to eliminate redundancy constraints according to claim 3, wherein, The introduction of a subproblem containing only 0-1 variables and the solution of the subproblem containing only 0-1 variables are as follows: If the line power flow constraint holds, the upper and lower limits of the line power flow constraint are very close to the upper and lower limits of the semi-continuous constraint transformation form; construct a subproblem containing only 0-1 variables to solve for the corresponding 0-1 variable values. The objective of this subproblem containing only 0-1 variables is to minimize the distance between the upper and lower limits of the semi-continuous constraint and the upper and lower limits of the line power flow constraint, and the closer the upper and lower limits of the semi-continuous constraint are, the better. The semi-continuous constraint is transformed into the same form as the line power flow constraint, as shown in formula (3): Construct a subproblem containing only 0-1 variables: Among them, the constraint term is that the unit combination problem only contains I. i,t W i,t and Y i,t The constraint H(I) ≥ 0; W i,t and Y i,t These are the decision variables for whether the unit is turned on and off, respectively. Transform the above objective function into one containing only I. i,t The linear expression L(I) i,t Solve such subproblems to obtain solutions S with 0-1 variables; 5. The method for accelerating unit combination optimization based on machine learning to eliminate redundancy constraints according to claim 4, wherein, The penalty-based approach adds the solutions to subproblems containing only 0-1 variables to the objective function of the original unit combination problem, making the solutions to the subproblems as close as possible to the optimal solution of the unit combination problem. Specifically: The constraints are G(I,P,W)≥0 and H(I)≥0, expressing the unit combination problem as a set of constraints including unit start-up / shutdown decision variable I, unit output variable P, and other variables w. G(I,P,W) is the constraint set without line power flow constraints; H(I)≥0 indicates that the unit combination problem only contains I. i,t W i,t Y i,t Constraints on 0-1 variables.

6. The method for accelerating unit combination optimization based on machine learning to eliminate redundancy constraints according to claim 5, wherein, The minimum continuous start-stop time constraint is specifically as follows:

7. The method for accelerating unit combination optimization based on machine learning to eliminate redundancy constraints according to claim 6, wherein, The method involves obtaining the offset of each unit's state variable at various time periods based on the solutions to the sub-problems and the optimal solutions to the original unit combination problem. By changing the characteristic demand D of the original unit combination problem, multiple optimal solutions are obtained, and the relationship between demand D and the offsets is derived. This relationship is then learned into a predictive regression model. Specifically, by changing demand D, the unit combination problem is solved to obtain several optimal solutions I. * Calculate each I * Distance to the solution S of the subproblem containing only 0-1 variables; statistical analysis of the offset degree of each unit in each time period to form the relationship between demand D and offset vector v; The SCUC model, determined by demand d∈D, has a functional relationship. Make Among them, v j Each component of the offset degree v for each unit in each time period.

8. The method for accelerating unit combination optimization based on machine learning to eliminate redundancy constraints according to claim 7, wherein, The process of identifying and removing redundant constraints in the minimum continuous start-up and shutdown time constraint involves: given a new unit combination problem, a learned prediction model predicts the offset of the unit state variables for each unit in each time period of the new unit combination problem to be solved, for the minimum continuous start-up and shutdown time constraint... or Make a judgment, g(v) i ) represents the function expression of the offset. If formulas (10) and (11) still hold, then the original If a constraint is redundant, then that constraint should be removed.

9. A unit combination acceleration optimization system based on machine learning to eliminate redundancy constraints, including: The first acquisition module acquires the semi-continuous constraints of the upper and lower limits of the unit output of the SCUC model and the line power flow constraints. The module introduces a subproblem containing only 0-1 variables. If the semi-continuous constraints of the unit output upper and lower limits and the line power flow constraints in the SCUC model are both valid, then the upper and lower limits of the two constraints are very close. The module then introduces a subproblem containing only 0-1 variables and solves the subproblem containing only 0-1 variables. The second acquisition module adds the solution of the subproblem containing only 0-1 variables to the objective function of the original unit combination problem based on a penalty method, so that the solution of the subproblem is as close as possible to the optimal solution of the original unit combination problem. The offset data acquisition module obtains the offset of the unit state variable of each unit in each time period based on the solution of the sub-problem and the optimal solution of the original unit combination problem; changes the characteristic demand D of the original unit combination problem, obtains multiple optimal solutions, obtains the relationship between demand D and offset, and learns it into a predictive regression model. The judgment module, given a new unit combination problem, predicts the offset of the unit state variable of each unit in the new unit combination problem to be solved in each time period based on the learned prediction model, identifies and removes redundant constraints in the minimum continuous start-stop time constraint, so as to realize accelerated optimization of unit combination.

10. A terminal device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method as claimed in any one of claims 1-8.

11. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method as claimed in any one of claims 1-8.

12. A chip, the chip comprising: Memory, used to store computer programs; A processor, connected to the memory, is configured to retrieve and run a computer program from the memory, causing a device on which the chip is mounted to perform the method as described in any one of claims 1 to 8.

13. A computer program product comprising a computer program or instructions which, when executed by a processor, implement the method as described in any one of claims 1 to 7.

14. A computer program that causes a computer to perform the method as described in any one of claims 1 to 8.

Images