Processing method and device for unit combination in power system and electronic equipment

Abstract

The invention discloses a unit commitment processing method and device in a power system and electronic equipment. The method comprises the following steps: acquiring a target mathematical model corresponding to a target unit combination problem in the power system; according to the similarity between the target mathematical model and sample mathematical models corresponding to the multiple sample unit commitment problems, a target sample mathematical model is determined from the multiple sample mathematical models, and hyper-parameters corresponding to the sample mathematical models are obtained through screening according to solving durations corresponding to the sample mathematical models; and taking the hyper-parameter corresponding to the target sample mathematical model as the hyper-parameter of a solver, and solving the target mathematical model through the solver to obtain a solving result for the target unit commitment problem. According to the method and the device, the technical problem of low solving speed due to the fact that a solver adopts a fixed hyper-parameter to solve the unit commitment problem in a power system in the prior art is solved.

Description

Technical Field

[0001] The present application relates to the field of power systems, and in particular to a method, device and electronic equipment for processing unit combinations in a power system. Background Art

[0002] The power system unit combination is the basis and key issue for efficient dispatch of the power system. In essence, it makes decisions on the on / off status and power generation of each unit in the power system to meet the actual load demand. This decision problem is usually solved using solvers, which have embedded frameworks such as branch pricing and a large number of heuristic strategies. The supporting operating behaviors of the solver's internal framework and heuristic strategies are controlled by a set of configurable hyperparameters. The default hyperparameter settings are usually set directly by the solver developer to meet the needs of solving general problems, but for specific categories of problems, the parameter settings may not be the optimal settings.

[0003] Specifically, for the unit combination problem, since the grid load, grid topology results, maintenance plan, etc. make the unit combination problem have the characteristics of real-time variability, the fixed configuration hyperparameters cannot be suitable for the unit combination problem in all scenarios, so there is a problem of slow solution speed when solving the unit combination problem.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present application provide a method, device and electronic device for processing unit combination in an electric power system, so as to at least solve the technical problem that the solver in the related art uses fixed hyperparameters to solve the unit combination problem in the electric power system, resulting in a slow solution speed.

[0006] According to one aspect of an embodiment of the present application, a method for processing unit combination in an electric power system is provided, including: obtaining a target mathematical model corresponding to a target unit combination problem in the electric power system; determining a target sample mathematical model from multiple sample mathematical models based on a similarity between the target mathematical model and sample mathematical models corresponding to multiple sample unit combination problems, wherein the optimal hyperparameters corresponding to the sample mathematical model are obtained by screening according to the solution time corresponding to the sample mathematical model; using the optimal hyperparameters corresponding to the target sample mathematical model as hyperparameters of a solver, and solving the target mathematical model through the solver to obtain a solution result for the target unit combination problem.

[0007] Furthermore, before determining the target sample mathematical model from multiple sample mathematical models based on the similarity between the target mathematical model and the sample mathematical models corresponding to the multiple sample unit combination problems, the method also includes: obtaining first feature information corresponding to the sample mathematical model in the multiple sample mathematical models; performing feature extraction on the target mathematical model through a feature extraction model to obtain second feature information; and determining the similarity between the target mathematical model and the multiple sample mathematical models based on the second feature information and the first feature information corresponding to the multiple sample mathematical models.

[0008] Furthermore, the feature extraction model is obtained in the following manner: for a sample mathematical model among multiple sample mathematical models, a target solution time for the solver to solve the sample mathematical model is determined; training samples are determined according to the sample mathematical models among the multiple sample mathematical models, and the target solution time is used as a true label to construct a training sample set; an initial feature extraction model is trained according to the training sample set to obtain the feature extraction model.

[0009] Furthermore, the feature extraction model is a graph neural network model, and the feature extraction model is used to extract features of the target mathematical model to obtain the second feature information, including: determining a bipartite graph according to the target mathematical model; and extracting features of the bipartite graph through the feature extraction model to obtain the second feature information.

[0010] Furthermore, the target mathematical model includes at least one constraint and at least one decision variable, and determining the bipartite graph according to the target mathematical model includes: constructing nodes in a first node subset according to at least one decision variable, and constructing nodes in a second node subset according to at least one constraint; constructing edges between nodes in the first node subset and nodes in the second node subset according to an association relationship between at least one decision variable and at least one constraint; and determining the bipartite graph according to the first node subset, the second node subset, and the edges.

[0011] Furthermore, constructing nodes in the first node subset according to at least one decision variable includes: using a decision variable in the at least one decision variable as a node in the first node subset; and determining variable information of the decision variable as an attribute of the node corresponding to the decision variable.

[0012] Furthermore, the target mathematical model also includes an objective function, and determining the variable information of the decision variable as the attribute of the node corresponding to the decision variable includes: determining the variable information of the decision variable and the coefficient of the decision variable in the objective function as the attribute of the node corresponding to the decision variable.

[0013] Furthermore, constructing nodes in the second node subset according to at least one constraint condition includes: taking a target parameter of the constraint condition in at least one constraint condition as a node in the second node subset; and determining a parameter value of the target parameter of the constraint condition as an attribute of the node corresponding to the constraint condition.

[0014] Furthermore, the hyperparameters corresponding to the sample mathematical model are determined in the following manner: obtaining multiple sets of hyperparameters; for the hyperparameters in the multiple sets of hyperparameters, determining the time it takes for the solver to solve the sample mathematical model under the hyperparameters; and determining the hyperparameters corresponding to the sample mathematical model from the multiple sets of hyperparameters based on the solution time corresponding to the sample mathematical model under the multiple sets of hyperparameters.

[0015] According to another aspect of an embodiment of the present application, a method for processing unit combination in an electric power system is also provided, including: obtaining a target mathematical model corresponding to a target unit combination problem in the electric power system uploaded by a client; determining a target sample mathematical model from multiple sample mathematical models based on the similarity between the target mathematical model and sample mathematical models corresponding to multiple sample unit combination problems in a cloud server, using hyperparameters corresponding to the target sample mathematical model as hyperparameters of a solver, and solving the target mathematical model through the solver to obtain a solution result for the target unit combination problem, wherein the hyperparameters corresponding to the sample mathematical model are obtained by screening according to the solution time corresponding to the sample mathematical model; and feeding back the solution result to the client.

[0016] According to another aspect of an embodiment of the present application, a processing device for unit combination in an electric power system is also provided, including: a first acquisition unit, used to acquire a target mathematical model corresponding to a target unit combination problem in the electric power system; a first determination unit, used to determine a target sample mathematical model from multiple sample mathematical models based on a similarity between the target mathematical model and sample mathematical models corresponding to multiple sample unit combination problems, wherein the hyperparameters corresponding to the sample mathematical model are obtained by screening according to the solution time corresponding to the sample mathematical model; a processing unit, used to use the hyperparameters corresponding to the target sample mathematical model as the hyperparameters of a solver, and solve the target mathematical model through the solver to obtain a solution result for the target unit combination problem.

[0017] Furthermore, the processing device for the unit combination in the power system also includes: a second acquisition unit, used to obtain first feature information corresponding to a sample mathematical model in multiple sample mathematical models; a feature extraction unit, used to extract features of the target mathematical model through the feature extraction model to obtain second feature information; a second determination unit, used to determine the similarity between the target mathematical model and multiple sample mathematical models based on the second feature information and the first feature information corresponding to the multiple sample mathematical models.

[0018] Furthermore, the processing device for the unit combination in the power system also includes: a third determination unit, used to determine, for a sample mathematical model among multiple sample mathematical models, a target solution time of the sample mathematical model by the solver under preset hyperparameters; a construction unit, used to determine training samples according to the sample mathematical models among the multiple sample mathematical models, and use the target solution time as a true label to construct a training sample set; a training unit, used to train an initial feature extraction model according to the training sample set to obtain a feature extraction model.

[0019] Furthermore, the feature extraction model is a graph neural network model, and the feature extraction unit includes: a determination subunit, used to determine the bipartite graph according to the target mathematical model; and a feature extraction subunit, used to extract features of the bipartite graph through the feature extraction model to obtain second feature information.

[0020] Furthermore, the target mathematical model includes at least one constraint and at least one decision variable, and the determination subunit includes: a first construction module, used to construct nodes in the first node subset according to at least one decision variable, and to construct nodes in the second node subset according to at least one constraint; a second construction module, used to construct edges between nodes in the first node subset and nodes in the second node subset according to an association relationship between at least one decision variable and at least one constraint; and a determination module, used to determine a bipartite graph based on the first node subset, the second node subset and the edges.

[0021] Furthermore, the first construction module includes: a first processing submodule, used to use a decision variable in at least one decision variable as a node in the first node subset; and a first determining submodule, used to determine variable information of the decision variable as an attribute of the node corresponding to the decision variable.

[0022] Furthermore, the target mathematical model also includes an objective function, and the first determining submodule includes: a second determining submodule, which is used to determine the variable information of the decision variable and the coefficient of the decision variable in the objective function as the attribute of the node corresponding to the decision variable.

[0023] Furthermore, the first construction module includes: a second processing submodule, used to use the target parameter of the constraint condition in at least one constraint condition as a node in the second node subset; and a third determination submodule, used to determine the parameter value of the target parameter of the constraint condition as the attribute of the node corresponding to the constraint condition.

[0024] Furthermore, the processing device for the unit combination in the power system also includes: a third acquisition unit, used to acquire multiple sets of hyperparameters; a fourth determination unit, used to determine, for the hyperparameters in the multiple sets of hyperparameters, the solution time of the sample mathematical model by the solver under the hyperparameters; and a fifth determination unit, used to determine the hyperparameters corresponding to the sample mathematical model from the multiple sets of hyperparameters based on the corresponding solution time of the sample mathematical model under the multiple sets of hyperparameters.

[0025] Furthermore, the first acquisition unit acquires the following data: Acquire first feature information corresponding to a sample mathematical model among the multiple sample mathematical models; Extracting features from the target mathematical model using a feature extraction model to obtain second feature information; Determine the similarity between the target mathematical model and the multiple sample mathematical models according to the second feature information and the first feature information corresponding to the multiple sample mathematical models.

[0026] Furthermore, the feature extraction model method acquired by the first acquisition unit is specifically as follows: For a sample mathematical model among the plurality of sample mathematical models, determining a target solving time for the solver to solve the sample mathematical model; Determine a training sample according to a sample mathematical model among the multiple sample mathematical models, and use the target solution time as a true label to construct a training sample set; An initial feature extraction model is trained according to the training sample set to obtain the feature extraction model.

[0027] Furthermore, the feature extraction model acquired by the first acquisition unit is a graph neural network model, and the feature extraction of the target mathematical model is performed by the feature extraction model to obtain the second feature information including: Determining a bipartite graph according to the target mathematical model; The feature extraction model is used to extract features from the bipartite graph to obtain the second feature information.

[0028] Furthermore, the target mathematical model obtained by the first determination unit determines the bipartite graph including: constructing nodes in the first node subset according to the at least one decision variable, and constructing nodes in the second node subset according to the at least one constraint condition; Building edges between nodes in the first node subset and nodes in the second node subset according to an association relationship between the at least one decision variable and the at least one constraint condition; The bipartite graph is determined based on the first subset of nodes, the second subset of nodes, and the edges.

[0029] Further, the nodes in the first node subset acquired by the first acquisition unit include: using a decision variable in the at least one decision variable as a node in the first node subset; The variable information of the decision variable is determined as the attribute of the node corresponding to the decision variable.

[0030] Furthermore, the attributes of the nodes corresponding to the decision variables determined by the first determination unit include: The variable information of the decision variable and the coefficient of the decision variable in the objective function are determined as the attributes of the node corresponding to the decision variable.

[0031] Furthermore, the nodes in the second node subset acquired by the first acquisition unit include: using a target parameter of a constraint condition in the at least one constraint condition as a node in the second node subset; The parameter value of the target parameter of the constraint condition is determined as the attribute of the node corresponding to the constraint condition.

[0032] Furthermore, the first determining unit determines the hyperparameters corresponding to the mathematical model in the following manner: Get multiple sets of hyperparameters; For a hyperparameter in the plurality of sets of hyperparameters, determining a time duration for the solver to solve the sample mathematical model under the hyperparameter; According to the solution time corresponding to the sample mathematical model under the multiple sets of hyperparameters, the optimal hyperparameter corresponding to the sample mathematical model is determined from the multiple sets of hyperparameters.

[0033] According to another aspect of an embodiment of the present invention, there is further provided an electronic device, comprising: a memory storing an executable program; and a processor for running the program, wherein when the program is run, any one of the above-mentioned methods for processing a unit combination in a power system is executed.

[0034] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the computer program implements any one of the above methods for processing unit combinations in a power system.

[0035] In an embodiment of the present application, a target mathematical model corresponding to a target unit combination problem in a power system is obtained; a target sample mathematical model is determined from multiple sample mathematical models based on a similarity between the target mathematical model and sample mathematical models corresponding to multiple sample unit combination problems, wherein a hyperparameter corresponding to the sample mathematical model is obtained by screening according to a solution time corresponding to the sample mathematical model; the hyperparameter corresponding to the target sample mathematical model is used as a hyperparameter of a solver, and the target mathematical model is solved by the solver to obtain a solution result for the target unit combination problem, wherein the hyperparameter corresponding to the sample mathematical model obtained by screening according to the solution time corresponding to the sample mathematical model makes the sample mathematical model The solution time under the corresponding hyperparameters can be limited within a certain range. Since similar mathematical models have similar solution characteristics, a target sample mathematical model similar to the target mathematical model is screened out from multiple sample mathematical models, and two similar mathematical models are set to use the same hyperparameters, so that the solution time of the target mathematical model under the hyperparameters can also be limited within a certain range, thereby achieving the purpose of targeted setting of hyperparameters for the target mathematical model corresponding to the target unit combination problem, and realizing the technical effect of effectively improving the solution speed, thereby solving the technical problem of slow solution speed in the related technology that the solver uses fixed hyperparameters to solve the unit combination problem in the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings: Figure 1 is a schematic diagram of a computer terminal provided according to Embodiment 1 of the present application; Figure 2 is a flow chart of a method for processing a unit combination in a power system provided in Embodiment 1 of the present application; Figure 3 It is a schematic diagram of a process of obtaining a feature extraction model through optional training provided in Embodiment 1 of the present application; Figure 4 It is a schematic diagram of an optional process for processing a target unit combination problem provided in accordance with the first embodiment of the present application; Figure 5 is a schematic diagram of determining a bipartite graph according to Embodiment 1 of the present application; Figure 6 is a flow chart of a method for processing unit combination in a power system provided in Embodiment 2 of the present application; Figure 7 is a schematic diagram of a processing device for a unit combination in a power system provided according to Embodiment 3 of the present application; Figure 8 It is a structural block diagram of an electronic device provided according to Embodiment 4 of the present application. DETAILED DESCRIPTION

[0037] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0038] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0039] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards in the relevant regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0040] First, some nouns or terms that appear in the description of the embodiments of the present application are subject to the following explanations: Unit Commitment: The optimization problem of deciding which generators should be running and their output at a specific time period (e.g., every hour of a day) based on load demand, generation cost, startup and shutdown costs, emission limits, and other operating constraints. Example 1

[0041] According to an embodiment of the present application, a method for processing a unit combination in an electric power system is also provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0042] The method embodiment provided in the first embodiment of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 The hardware structure block diagram of a computer terminal (or mobile device) for implementing a method for processing unit combination in a power system is shown. Figure 1 As shown, the computer terminal (or mobile device) 10 may include a processor set 102 (the processor set 102 may include but is not limited to a processing device such as a microprocessor (Microcontroller Unit, referred to as MCU) or a programmable logic device (Field Programmable Gate Array, referred to as FPGA), and the processor set 102 may include a processor set, Figure 1 102a, 102b, ..., 102n are used to illustrate), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a Universal Serial Bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. Those skilled in the art can understand that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0043] It should be noted that the one or more processors 102 and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuits may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10 (or mobile device).

[0044] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the method for processing the combination of units in the power system in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, the processing method for the combination of units in the power system described above is realized. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0045] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0046] The display may be a touch screen type liquid crystal display, which may enable a user to interact with a user interface of the computer terminal 10 (or mobile device).

[0047] The power system unit combination is the basis and key issue of efficient power system dispatch. In essence, it makes decisions on the on / off state and power generation of each unit in the power system to meet the actual load demand. This decision problem is usually solved by using solvers, which have embedded frameworks such as branch pricing and a large number of heuristic strategies. The supporting operation behaviors of the solver's internal framework and heuristic strategies are controlled by a set of configurable hyperparameters. The default hyperparameter settings are usually set directly by the solver developer to meet the needs of solving general problems, but for specific categories of problems, the parameter settings may not be the optimal settings.

[0048] Specifically, for the unit combination problem, since the grid load, grid topology results, maintenance plan, etc. make the unit combination problem have the characteristics of real-time variability, the fixed configuration hyperparameters cannot be suitable for the unit combination problem in all scenarios, so there is a problem of slow solution speed when solving the unit combination problem.

[0049] In the above technical background, the present application provides Figure 2 The method for handling unit combination in the power system shown. Figure 2 1 is a flow chart of a method for processing a unit combination in a power system according to the first embodiment of the present application. Figure 2 As shown, the method includes: Step S201, obtaining a target mathematical model corresponding to a target unit commitment problem in a power system.

[0050] Optionally, electronic devices, application systems, servers and other devices can be used as the execution subject of the present application. In this embodiment, the automatic tuning parameter adjuster is used as the execution subject to obtain the target mathematical model corresponding to the target unit combination problem in the power system.

[0051] The unit combination problem in the power system is characterized by real-time variability, depending on changes in events such as grid load, grid topology results, and maintenance plans in the power system. The above-mentioned target unit combination problem can be a new unit combination problem generated during the operation of the power system.

[0052] Optionally, the target mathematical model is obtained by mathematically modeling the target unit combination problem, and the target mathematical model may be a mathematical programming model, for example, the target mathematical model may be a mixed integer linear programming model, a mixed integer nonlinear programming model, an integer programming model or the like.

[0053] When the target mathematical model is a mathematical programming model, the target mathematical model includes an objective function, decision variables and constraints, wherein the objective function is the target to be optimized in the mathematical programming model, the decision variables are unknowns in the mathematical programming model, and they represent the control quantities or parameters to be optimized, and the constraints are mathematical expressions that limit the value range of the decision variables in the mathematical programming model.

[0054] For example, in the unit combination scenario of the power system, the objective function may be a function that makes the total operating cost of the generator set relatively small. The total operating cost includes but is not limited to the linear coal consumption cost, no-load cost, startup cost, shutdown cost, etc. of the unit. The decision variable may be one or more, and the decision variable may be the startup state of unit g at time t, the shutdown state of unit g at time t, the output power of unit g at time t, etc. The constraint condition may be one or more, and the constraint condition may be output constraint, start-stop time constraint, power balance constraint, etc.

[0055] Step S202, determining a target sample mathematical model from multiple sample mathematical models based on the similarity between the target mathematical model and the sample mathematical models corresponding to the multiple sample unit combination problems, wherein the hyperparameters corresponding to the sample mathematical model are obtained by screening according to the solution time corresponding to the sample mathematical model.

[0056] Since the hyperparameters of the solver affect the performance and solution quality of the solver, the solver takes different time to solve the mathematical model of the same problem under different hyperparameters. Therefore, in order to improve the speed of solving the new unit combination problem (i.e., the aforementioned target unit combination problem) each time a new unit combination problem appears in the power system, the automatic tuning parameter adjuster can first determine the relatively preferred hyperparameters when solving the mathematical models of multiple known unit combination problems (i.e., the aforementioned sample unit combination problems), and then determine the hyperparameters required to solve the new unit combination problem based on the similarity between the mathematical model of the new unit combination problem and the mathematical models of multiple known unit combination problems. Among them, the hyperparameters of the solver include but are not limited to the iteration limit parameter, the accuracy requirement parameter, the number of threads, the algorithm selection parameter, etc.

[0057] Optionally, the model types of the multiple sample mathematical models and the target mathematical model are the same, for example, the multiple sample mathematical models and the target mathematical model are both mixed integer linear programming models.

[0058] The hyperparameters corresponding to the multiple sample mathematical models are not exactly the same. The hyperparameters corresponding to the sample mathematical models are obtained by screening according to the solution time corresponding to the sample mathematical models. In other words, the solution time of the sample mathematical model under the hyperparameters corresponding to the sample mathematical model is less than the preset time. For example, the automatic tuning tuner can test the solution time of the sample mathematical model under multiple sets of hyperparameters, and thus determine the hyperparameters corresponding to the sample mathematical model from the multiple sets of hyperparameters according to the solution time obtained by the test.

[0059] Optionally, after obtaining the target mathematical model, the automatic tuning parameter adjuster may first determine the similarity between the target mathematical model and sample mathematical models corresponding to the plurality of sample unit combination problems. For example, the automatic tuning parameter tuner can directly determine the similarity between two mathematical models (i.e., any sample mathematical model in multiple sample mathematical models and the target mathematical model) based on the number of constraints, data of decision variables and other information of the two mathematical models. The automatic tuning parameter tuner can also first extract the feature information of the two mathematical models mentioned above, and then determine the similarity between the two mathematical models based on the feature information. The automatic tuning parameter tuner also uses a solver (such as MindOpt, etc.) to respectively configure multiple groups of different hyperparameter values ​​(including the default hyperparameter values ​​of the solver) and then solve the solution time of multiple sample mathematical models as a label to train a preset neural network model (such as the neural network model in the figure, etc.), so that the neural network model can output the representation of the mathematical model, thereby having the mathematical model classification function, and at the same time, the representation of all sample mathematical models used for training and the corresponding optimal parameter configuration are stored in the representation and parameter library, so that the target mathematical model is input into the trained neural network model, and the trained neural network model represents and classifies the target mathematical model to obtain a classification result, wherein the classification result includes the probability value that the target mathematical model belongs to the category of the sample mathematical model in the multiple sample mathematical models.

[0060] After determining the similarity between the target mathematical model and multiple sample mathematical models, the automatic tuning parameter adjuster can determine the target sample mathematical model from the multiple sample mathematical models according to the similarity. For example, the sample mathematical model with higher similarity is determined as the target sample mathematical model, and for another example, any sample mathematical model among the sample mathematical models with similarity greater than a threshold is determined as the target sample mathematical model.

[0061] It should be noted that by determining the target sample mathematical model from multiple sample mathematical models based on the similarity between the models, it is convenient to determine the sample mathematical model that is similar to the target mathematical model, thereby facilitating the accurate determination of the hyperparameters required to solve the target mathematical model.

[0062] Step S203, using the hyperparameters corresponding to the target sample mathematical model in the parameter library as the hyperparameters of the solver, configuring them into the solver, and then solving the target mathematical model through the solver to obtain the solution result for the target unit combination problem.

[0063] Optionally, the solver corresponds to the model type of the target mathematical model. For example, if the target mathematical model is a mixed integer linear programming model, the solver is a mixed integer linear programming solver.

[0064] Optionally, the target mathematical model can be input into the solver in a data format that the solver can recognize, so as to be solved by the solver. For example, when the target mathematical model is a linear programming model, the MPS (Mathematical Programming System) file corresponding to the target mathematical model is input into the solver for solving, wherein the MPS file is used to describe the mathematical model of the linear programming problem. The solution result output by the solver is the solution result for the target unit combination problem.

[0065] It should be noted that since the solution time of the sample mathematical model is relatively short when it is solved under its corresponding hyperparameters, the solution speed of the target mathematical model can be specifically improved by setting the target mathematical model similar to the sample mathematical model to adopt the same hyperparameters.

[0066] In this scheme, the hyperparameters corresponding to the sample mathematical models are obtained by screening according to the solution time corresponding to the sample mathematical models, so that the solution time of the sample mathematical models under the corresponding hyperparameters can be limited within a certain range. Since similar mathematical models have similar solution characteristics, a target sample mathematical model similar to the target mathematical model is screened out from multiple sample mathematical models, and two similar mathematical models are set to use the same hyperparameters, so that the solution time of the target mathematical model under the hyperparameters can also be limited within a certain range, thereby achieving the purpose of targeted setting of hyperparameters for the target mathematical model corresponding to the target unit combination problem, and realizing the technical effect of effectively improving the solution speed, thereby solving the technical problem of slow solution speed in the related technology that the solver uses fixed hyperparameters to solve the unit combination problem in the power system.

[0067] How to determine the similarity between the target mathematical model and multiple sample mathematical models is crucial. Therefore, in the method for processing unit combination in the power system provided in Example 1 of the present application, before determining the target sample mathematical model from multiple sample mathematical models based on the similarity between the target mathematical model and the sample mathematical models corresponding to the multiple sample unit combination problems, the method for processing unit combination in the power system includes: obtaining first feature information corresponding to the sample mathematical model in the multiple sample mathematical models; performing feature extraction on the target mathematical model through a feature extraction model to obtain second feature information; and determining the similarity between the target mathematical model and the multiple sample mathematical models based on the second feature information and the first feature information corresponding to the multiple sample mathematical models.

[0068] In order to improve the accuracy of the determined similarity, feature extraction is performed on both the sample mathematical model and the target mathematical model, so that the similarity is determined by the extracted feature information. Optionally, feature extraction can be performed on multiple sample mathematical models and target mathematical models respectively through a feature extraction model to obtain multiple first feature information and second feature information. The feature extraction model can be a graph neural network model, a convolutional neural network model, etc.

[0069] Optionally, the feature extraction model can be trained based on the above-mentioned multiple sample mathematical models and their corresponding true labels, and the feature extraction model can also be trained based on other mathematical models (i.e., mathematical models other than the above-mentioned multiple sample mathematical models) and the true labels corresponding to other mathematical models. The true label can be the model category of the mathematical model, or the solution time of the mathematical model.

[0070] The first feature information and the second feature information are in vector form. Therefore, after determining the first feature information and the second feature information, the automatic tuning parameter tuner can determine the similarity between the first feature information and the second feature information based on Euclidean distance, Manhattan distance, cosine similarity, Pearson correlation coefficient, etc., thereby determining the similarity between the first feature information and the second feature information as the similarity between the sample mathematical model and the target mathematical model.

[0071] It should be noted that by obtaining the first characteristic information and the second characteristic information, the redundant information in the sample mathematical model and the target mathematical model is effectively eliminated. By determining the similarity between the two mathematical models based on the characteristic information, the accuracy of the determined similarity is effectively improved, thereby facilitating the improvement of the accuracy of the determined hyperparameters and the efficiency of determining the hyperparameters. This avoids the problem that when predicting the hyperparameters corresponding to the target mathematical model based on machine learning, the machine learning method can only learn the shallow features of the problem, and cannot achieve effective learning for ultra-large-scale and complex unit combination problems. In addition, in the actual power system dispatching and operation, solving the problem is too time-consuming and cannot meet the real-time response requirements.

[0072] In order to improve the accuracy of feature extraction, in the processing method of unit combination in the power system provided in Example 1 of the present application, the feature extraction model is obtained in the following manner: for a sample mathematical model among multiple sample mathematical models, determining the target solution time of the sample mathematical model by the solver under preset hyperparameters; determining training samples according to the sample mathematical models among the multiple sample mathematical models, and taking the target solution time as the true label to construct a training sample set; training an initial feature extraction model according to the training sample set to obtain a feature extraction model.

[0073] Optionally, the sample mathematical model and the above-mentioned target mathematical model are solved using the same solver.

[0074] In order to determine the true labels of multiple sample mathematical models, the same solver is used to solve the sample mathematical models in the multiple sample mathematical models under the same preset hyperparameters to obtain the target solution time corresponding to the sample mathematical models.

[0075] Optionally, in the process of determining the training samples, a sample mathematical model among multiple sample mathematical models can be directly used as a training sample, or data processing can be performed on the sample mathematical model among multiple sample mathematical models, so that the data processing result is determined as the training sample. For example, a bipartite graph conversion process (i.e., the aforementioned data processing) is performed on the sample mathematical model among multiple sample mathematical models to obtain a sample bipartite graph (i.e., the aforementioned data processing result), so that the sample bipartite graph is determined as the training sample.

[0076] After the training samples are determined, the target solution time corresponding to the sample mathematical model is used as the true label of the training sample corresponding to the sample mathematical model, thereby constructing a training sample set. Afterwards, the initial feature extraction model is trained according to the training sample set until the preset iteration condition is reached to obtain a feature extraction model. During the training process, the initial feature extraction model will not only output the predicted label of the sample mathematical model, but also output the feature information obtained by extracting the features of the training samples corresponding to the sample mathematical model. Optionally, the feature extraction model can be a graph neural network model, a convolutional neural network model or other models with feature extraction functions.

[0077] It should be noted that since different sample mathematical models have different solution times under the same solver and the same hyperparameters, the target solution time of the sample mathematical model by the solver under preset hyperparameters is used as the true label, the initial feature extraction model can learn the feature patterns and differences of different mathematical models based on the dimension of solution time during the training process, thereby improving the accuracy of feature extraction of the feature extraction model.

[0078] In order to obtain the second characteristic information, in the processing method of the unit combination in the power system provided in Example 1 of the present application, the feature extraction model is a graph neural network model, and the feature extraction model is used to extract features of the target mathematical model to obtain the second characteristic information, including: determining a bipartite graph according to the target mathematical model; and extracting features of the bipartite graph through the feature extraction model to obtain the second characteristic information.

[0079] Alternatively, the complex relationships and patterns inherent in unit commitment problems can be visualized and analyzed through bipartite graphs. A bipartite graph is a special type of graph whose vertices can be clearly divided into two non-overlapping subsets. Each edge connects a vertex from its own subset, and there are no edges connecting vertices in the same subset. This structural property makes the bipartite graph particularly suitable for representing binary relationships between two types of entities. Applying bipartite graphs to unit commitment problems in power optimization can clearly and concisely describe the relationship between decision variables and constraints. Therefore, after obtaining the target mathematical model, the target mathematical model can be modeled as a bipartite graph.

[0080] For example, nodes in a first node subset may be constructed according to at least one decision variable in a target mathematical model, and nodes in a second node subset may be constructed according to at least one constraint condition in the target mathematical model. Edges may be constructed between nodes in the first node subset and nodes in the second node subset according to an association relationship between at least one decision variable and at least one constraint condition. A bipartite graph may be determined according to the first node subset, the second node subset, and the edges.

[0081] For another example, the nodes in the third node subset can be constructed according to at least one decision variable in the target mathematical model, and the nodes in the fourth node subset can be constructed according to the objective function in the target mathematical model. According to the association between at least one decision variable and the objective function, edges can be constructed between the nodes in the third node subset and the nodes in the fourth node subset, and a bipartite graph can be determined according to the third node subset, the fourth node subset and the edges.

[0082] In some embodiments, a decision variable in at least one decision variable can be used as a node in a third node subset, the variable information of the decision variable can be determined as an attribute of the node corresponding to the decision variable, the objective function can be used as a node in a fourth node subset, and the coefficient in the objective function can be determined as an attribute of the node corresponding to the objective function.

[0083] For another example, the nodes in the fifth node subset can be constructed according to at least one constraint in the target mathematical model, and the nodes in the sixth node subset can be constructed according to the objective function in the target mathematical model. According to the association between the decision variables in at least one constraint and the decision variables in the objective function, edges are constructed between the nodes in the fifth node subset and the nodes in the sixth node subset, and a bipartite graph is determined based on the fifth node subset, the sixth node subset and the edges.

[0084] In some embodiments, at least one constraint in the constraint conditions can be used as a node in the fifth node subset, the coefficients and target parameters in the constraint conditions can be determined as attributes of the node corresponding to the constraint conditions, the objective function can be used as a node in the fourth node subset, and the coefficients in the objective function can be determined as attributes of the node corresponding to the objective function.

[0085] After determining the bipartite graph according to the target mathematical model, the trained feature extraction model can be used to extract features from the bipartite graph to obtain the second feature information. Figure 3 The schematic diagram shown in FIG. 1 realizes the training of the above-mentioned feature extraction model. Figure 3 is a flow chart of an optional training process for obtaining a feature extraction model according to the first embodiment of the present application, such as Figure 3 As shown, after obtaining multiple sample mathematical models, first determine the target solution time of the solver for the multiple sample mathematical models under preset hyperparameters, obtain the true label corresponding to the sample mathematical model, and then convert the sample mathematical models in the multiple sample mathematical models into sample bipartite graphs to obtain the training samples corresponding to the sample mathematical models. After that, construct a training sample set based on the training samples and true labels corresponding to the multiple sample mathematical models to train the initial graph neural network model (i.e., the initial feature extraction model), thereby obtaining the graph neural network model (i.e., the feature extraction model). In addition, the first feature information corresponding to the sample mathematical model in the multiple sample mathematical models can be determined according to the feature extraction model, and the first feature information can be stored in the example representation library for use.

[0086] In an optional embodiment, it is possible to use Figure 4 The schematic diagram shown in FIG. 1 is used to solve the target unit combination problem. Figure 4 is a schematic diagram of an optional process for processing a target unit combination problem according to the first embodiment of the present application, such as Figure 4 As shown, after obtaining the target mathematical model corresponding to the new unit combination problem in the power system (that is, the target unit combination problem), the bipartite graph is determined according to the target mathematical model, and then the graph neural network model (that is, the feature extraction model) is used to extract features of the bipartite graph to obtain the second feature information, and then the first feature information with a relatively high similarity to the second feature information is searched in the above-mentioned example characterization library, so that the sample mathematical model corresponding to the first feature information is determined as the target sample mathematical model, and the hyperparameters corresponding to the target sample mathematical model are used as the hyperparameters of the solver, so that the target mathematical model is solved by the solver to obtain the solution result for the target unit combination problem.

[0087] It should be noted that since the use of a bipartite graph to represent the unit combination problem can visually identify which variables are involved in the constraints and clearly explain the constraint matrix structure of the unit combination problem, it can help the feature extraction model understand the overall sparsity of the problem, better learn the deep representations of different unit combination problems (that is, different mathematical models), and learn the characteristic paradigms and differences of different unit combination problems, thereby improving the accuracy of feature extraction of the feature extraction model, and then improving the accuracy of hyperparameter determination and the speed of problem solving.

[0088] In order to effectively generate a bipartite graph, in the method for processing a unit combination in a power system provided in Example 1 of the present application, the target mathematical model includes at least one constraint and at least one decision variable, and determining the bipartite graph according to the target mathematical model includes: constructing nodes in a first node subset according to at least one decision variable, and constructing nodes in a second node subset according to at least one constraint; constructing edges between nodes in the first node subset and nodes in the second node subset according to an association relationship between at least one decision variable and at least one constraint; and determining the bipartite graph according to the first node subset, the second node subset, and the edges.

[0089] Optionally, the automatic tuning parameter adjuster can construct nodes in the first node subset based on at least one decision variable. For example, a decision variable in at least one decision variable is used as a node in the first node subset, and the variable information of the decision variable is determined as the attribute of the node corresponding to the decision variable. For another example, a decision variable in at least one decision variable is used as a node in the first node subset, and the variable information of the decision variable and the coefficient of the decision variable in the objective function are determined as the attributes of the node corresponding to the decision variable.

[0090] Optionally, the automatic tuning parameter tuner can construct nodes in the second node subset according to at least one constraint condition. For example, the target parameter of the constraint condition in at least one constraint condition is used as a node in the second node subset, and the parameter value of the target parameter of the constraint condition is determined as the attribute of the node corresponding to the constraint condition. For another example, the constraint condition in at least one constraint condition is used as a node in the second node subset, and the parameter value of the target parameter of the constraint condition and the coefficient in the constraint condition are determined as the attribute of the node corresponding to the constraint condition.

[0091] Optionally, the automatic tuning parameter tuner may construct edges between nodes in the first node subset and nodes in the second node subset based on the association between at least one decision variable and at least one constraint. For example, when a decision variable appears in a constraint, an edge is constructed between the node corresponding to the constraint and the node corresponding to the decision variable. Optionally, the constructed edge may or may not have attributes. In the case where the constructed edge has attributes, the coefficient of the decision variable in the constraint corresponding to the edge may be determined as the attribute of the edge.

[0092] After the first node subset, the second node subset and the edges are determined, the first node subset, the second node subset and the edges together constitute a bipartite graph corresponding to the target mathematical model.

[0093] In an optional embodiment, it is possible to use Figure 5The schematic diagram shown realizes the determination of the bipartite graph corresponding to the target mathematical model. Figure 5 is a schematic diagram of determining a bipartite graph according to the first embodiment of the present application. Figure 5 In the above example, the target mathematical model is a mixed integer linear programming model. is the decision variable, and the objective function is as follows:

[0094] The constraints are as follows:

[0095] like Figure 5 As shown, according to at least one decision variable in the target mathematical model Constructing nodes in the first node subset according to the constraint right value in at least one constraint condition in the target mathematical model (i.e., target parameters) construct nodes in the second node subset, construct edges between nodes in the first node subset and nodes in the second node subset according to the association relationship between at least one decision variable and at least one constraint condition, and convert the coefficient of the decision variable in the constraint condition Determine the attributes of the edge.

[0096] It should be noted that by constructing node subsets according to decision variables and constraints respectively, and constructing edges between nodes in node subsets according to the association between decision variables and constraints, the constructed bipartite graph can effectively, clearly and concisely describe the relationship between decision variables and constraints, thereby facilitating feature extraction by the feature extraction model.

[0097] In an optional embodiment, in the process of constructing nodes in the first node subset according to at least one decision variable, a decision variable in at least one decision variable can be used as a node in the first node subset; and variable information of the decision variable can be determined as an attribute of the node corresponding to the decision variable.

[0098] Optionally, the variable information of the decision variable includes but is not limited to the variable type, the upper limit of the variable value, the lower limit of the variable value, etc. The variable type includes but is not limited to binary, integer variable, continuous variable, etc.

[0099] It should be noted that by determining the variable information of the decision variable as the attribute of the node corresponding to the decision variable, the information of the decision variable in the bipartite graph is enriched, thereby further improving the accuracy of feature extraction.

[0100] In order to further improve the richness of information of decision variables in the bipartite graph, in the method for processing unit combination in the power system provided in Example 1 of the present application, the target mathematical model also includes an objective function, and determining the variable information of the decision variable as the attribute of the node corresponding to the decision variable includes: determining the variable information of the decision variable and the coefficient of the decision variable in the objective function as the attribute of the node corresponding to the decision variable.

[0101] For example, Figure 5 In It is the coefficient of the decision variable in the objective function. is the decision variable The coefficients in the objective function are, is the decision variable The coefficients in the objective function.

[0102] It should be noted that by taking the coefficients of the decision variables in the objective function as the attributes of the nodes corresponding to the decision variables, the richness of the information of the decision variables in the bipartite graph is effectively improved, thereby further improving the accuracy of feature extraction.

[0103] In an optional embodiment, in the process of constructing nodes in the second node subset according to at least one constraint condition, the target parameter of the constraint condition in at least one constraint condition can be used as the node in the second node subset, and the parameter value of the target parameter of the constraint condition can be determined as the attribute of the node corresponding to the constraint condition.

[0104] Optionally, the target parameter is a constant value other than the coefficient in the constraint, for example, Figure 5 In This is the target parameter mentioned above.

[0105] It should be noted that by determining the parameter value of the target parameter of the constraint condition as the attribute of the node corresponding to the constraint condition, the information of the constraint condition in the bipartite graph is enriched, thereby further improving the accuracy of feature extraction.

[0106] In order to improve the learning effect of the feature extraction model on the feature paradigm and differences of different unit combination problems during the training process, in the method for processing unit combination in the power system provided in the first embodiment of the present application, the automatic tuning and adjusting device can construct nodes in the seventh node subset according to at least one decision variable in the sample mathematical model, and construct nodes in the eighth node subset according to at least one constraint condition in the sample mathematical model, and construct edges between nodes in the seventh node subset and nodes in the eighth node subset according to the association relationship between at least one decision variable in the sample mathematical model and at least one constraint condition in the sample mathematical model, so as to determine the sample bipartite graph corresponding to the sample mathematical model according to the seventh node subset, the eighth node subset and the edges between the seventh node subset and the eighth node subset. Afterwards, the obtained sample bipartite graph is used as a training sample to train the initial feature extraction model to obtain the feature extraction model.

[0107] The process of determining the sample bipartite graph according to the sample mathematical model is the same as the process of determining the bipartite graph according to the target mathematical model, so it will not be repeated here.

[0108] It should be noted that, in the process of constructing training samples, node subsets are constructed according to the decision variables and constraints of the sample mathematical model, and the edges between the nodes in the node subsets are constructed according to the association between the decision variables and the constraints. The constructed sample bipartite graph can effectively, clearly and concisely describe the relationship between the decision variables and the constraints in the sample mathematical model, thereby facilitating the feature extraction model to effectively learn the characteristic paradigms and differences of different unit combination problems during the training process.

[0109] In order to effectively determine the hyperparameters corresponding to the sample mathematical model, in the method for processing the unit combination in the power system provided in Example 1 of the present application, the automatic tuning parameter adjuster can obtain multiple sets of hyperparameters; for the hyperparameters in the multiple sets of hyperparameters, determine the solution time of the sample mathematical model by the solver under the hyperparameters; and determine the hyperparameters corresponding to the sample mathematical model from the multiple sets of hyperparameters based on the solution time corresponding to the sample mathematical model under the multiple sets of hyperparameters.

[0110] Optionally, multiple sets of hyperparameters may be preset. For a hyperparameter in a plurality of sets of hyperparameters, the automatic tuning device determines the length of time it takes the solver to solve the sample mathematical model under the hyperparameter. For example, the solver performs a single solution on the sample mathematical model under the hyperparameter, and the length of time for the single solution is determined as the length of time for the solver to solve the hyperparameter. For another example, the solver performs multiple solutions on the sample mathematical model under the hyperparameter, and the average length of time for the multiple solutions is determined as the length of time for the solver to solve the hyperparameter.

[0111] After determining the solution time corresponding to the sample mathematical model under multiple sets of hyperparameters, the hyperparameters corresponding to the sample mathematical model can be determined from the multiple sets of hyperparameters according to the solution time. For example, for the sample mathematical model, the hyperparameters with relatively short solution time are determined as the hyperparameters corresponding to the sample mathematical model. For another example, any set of hyperparameters among the hyperparameters with a solution time shorter than a preset time is determined as the hyperparameters corresponding to the sample mathematical model.

[0112] It should be noted that by determining the hyperparameters with relatively short solution time as the hyperparameters corresponding to the sample mathematical model, the hyperparameters that can optimize the solution speed of the sample mathematical model are effectively determined. Therefore, when the hyperparameters of the target mathematical model are determined according to the similarity, the determined hyperparameters can also effectively optimize the solution speed of the target mathematical model.

[0113] In an embodiment of the present application, a target mathematical model corresponding to a target unit combination problem in a power system is obtained; a target sample mathematical model is determined from multiple sample mathematical models based on a similarity between the target mathematical model and sample mathematical models corresponding to multiple sample unit combination problems, wherein a hyperparameter corresponding to the sample mathematical model is obtained by screening according to a solution time corresponding to the sample mathematical model; the hyperparameter corresponding to the target sample mathematical model is used as a hyperparameter of a solver, and the target mathematical model is solved by the solver to obtain a solution result for the target unit combination problem, wherein the hyperparameter corresponding to the sample mathematical model obtained by screening according to the solution time corresponding to the sample mathematical model makes the sample mathematical model The solution time under the corresponding hyperparameters can be limited within a certain range. Since similar mathematical models have similar solution characteristics, a target sample mathematical model similar to the target mathematical model is screened out from multiple sample mathematical models, and two similar mathematical models are set to use the same hyperparameters, so that the solution time of the target mathematical model under the hyperparameters can also be limited within a certain range, thereby achieving the purpose of targeted setting of hyperparameters for the target mathematical model corresponding to the target unit combination problem, and realizing the technical effect of effectively improving the solution speed, thereby solving the technical problem of slow solution speed in the related technology that the solver uses fixed hyperparameters to solve the unit combination problem in the power system.

[0114] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0115] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, disk, CD), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of each embodiment of the present application. Example 2

[0116] According to an embodiment of the present application, a method for processing unit combination in a power system is also provided. Figure 6 is a flow chart of a method for processing a unit combination in a power system provided in Embodiment 2 of the present application, such as Figure 6 As shown, the method includes: Step S601, obtaining a target mathematical model corresponding to a target unit combination problem in a power system uploaded by a client.

[0117] Step S602: In the cloud server, a target sample mathematical model is determined from multiple sample mathematical models based on the similarity between the target mathematical model and the sample mathematical models corresponding to the multiple sample unit combination problems, and the hyperparameters corresponding to the target sample mathematical model are used as the hyperparameters of the solver. The target mathematical model is solved by the solver to obtain a solution result for the target unit combination problem, wherein the hyperparameters corresponding to the sample mathematical model are obtained by screening according to the solution time corresponding to the sample mathematical model.

[0118] Step S603: Feedback the solution result to the client.

[0119] Through the above scheme, the hyperparameters corresponding to the sample mathematical models are obtained by screening according to the solution time corresponding to the sample mathematical models, so that the solution time of the sample mathematical models under the corresponding hyperparameters can be limited within a certain range. Since similar mathematical models have similar solution characteristics, a target sample mathematical model similar to the target mathematical model is screened out from multiple sample mathematical models, and two similar mathematical models are set to use the same hyperparameters, so that the solution time of the target mathematical model under the hyperparameters can also be limited within a certain range, thereby achieving the purpose of targeted setting of hyperparameters for the target mathematical model corresponding to the target unit combination problem, and realizing the technical effect of effectively improving the solution speed, thereby solving the technical problem of slow solution speed in the related technology that the solver uses fixed hyperparameters to solve the unit combination problem in the power system.

[0120] In the cloud server, the specific processing method for the unit combination in the power system is the same as the method in the first embodiment, and will not be repeated here.

[0121] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0122] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, disk, CD), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of each embodiment of the present application. Example 3

[0123] According to an embodiment of the present application, a device for processing a unit combination in a power system for implementing the method for processing a unit combination in a power system is also provided. Figure 7 is a schematic diagram of a processing device for a unit combination in a power system provided in Embodiment 3 of the present application, such as Figure 7 As shown, the device includes: a first acquisition unit 701, a first determination unit 702, and a processing unit 703.

[0124] The first acquisition unit 701 is used to acquire a target mathematical model corresponding to a target unit commitment problem in a power system; A first determining unit 702 is used to determine a target sample mathematical model from multiple sample mathematical models according to the similarity between the target mathematical model and the sample mathematical models corresponding to the multiple sample unit combination problems, wherein the hyperparameters corresponding to the sample mathematical model are obtained by screening according to the solution time corresponding to the sample mathematical model; The processing unit 703 is used to use the hyperparameters corresponding to the target sample mathematical model as the hyperparameters of the solver, and solve the target mathematical model through the solver to obtain a solution result for the target unit commitment problem.

[0125] In the processing device for unit combination in the power system provided in Example 3 of the present application, the target mathematical model corresponding to the target unit combination problem in the power system is obtained by the first acquisition unit 701; the first determination unit 702 determines the target sample mathematical model from multiple sample mathematical models according to the similarity between the target mathematical model and the sample mathematical models corresponding to the multiple sample unit combination problems, wherein the hyperparameters corresponding to the sample mathematical model are obtained by screening according to the solution time corresponding to the sample mathematical model; the processing unit 703 uses the hyperparameters corresponding to the target sample mathematical model as the hyperparameters of the solver, and solves the target mathematical model through the solver to obtain the solution result for the target unit combination problem. In this scheme, the hyperparameters corresponding to the sample mathematical models are obtained by screening according to the solution time corresponding to the sample mathematical models, so that the solution time of the sample mathematical models under the corresponding hyperparameters can be limited within a certain range. Since similar mathematical models have similar solution characteristics, a target sample mathematical model similar to the target mathematical model is screened out from multiple sample mathematical models, and two similar mathematical models are set to use the same hyperparameters, so that the solution time of the target mathematical model under the hyperparameters can also be limited within a certain range, thereby achieving the purpose of targeted setting of hyperparameters for the target mathematical model corresponding to the target unit combination problem, and realizing the technical effect of effectively improving the solution speed, thereby solving the technical problem of slow solution speed in the related technology that the solver uses fixed hyperparameters to solve the unit combination problem in the power system.

[0126] Optionally, in the processing device for the unit combination in the power system provided in Example 3 of the present application, the processing device for the unit combination in the power system also includes: a second acquisition unit, used to acquire first feature information corresponding to a sample mathematical model in multiple sample mathematical models; a feature extraction unit, used to extract features of the target mathematical model through the feature extraction model to obtain second feature information; and a second determination unit, used to determine the similarity between the target mathematical model and multiple sample mathematical models based on the second feature information and the first feature information corresponding to the multiple sample mathematical models.

[0127] Optionally, in the processing device for the unit combination in the power system provided in Example 3 of the present application, the processing device for the unit combination in the power system also includes: a third determination unit, used to determine, for a sample mathematical model among multiple sample mathematical models, a target solution time of the solver for the sample mathematical model under preset hyperparameters; a construction unit, used to determine training samples according to the sample mathematical models among the multiple sample mathematical models, and use the target solution time as a true label to construct a training sample set; a training unit, used to train an initial feature extraction model according to the training sample set to obtain a feature extraction model.

[0128] Optionally, in the processing device for the unit combination in the power system provided in Example 3 of the present application, the feature extraction model is a graph neural network model, and the feature extraction unit includes: a determination subunit, used to determine the bipartite graph according to the target mathematical model; and a feature extraction subunit, used to extract features from the bipartite graph through the feature extraction model to obtain second feature information.

[0129] Optionally, in the processing device for the combination of units in the power system provided in Example 3 of the present application, the target mathematical model includes at least one constraint and at least one decision variable, and the determination subunit includes: a first construction module, used to construct nodes in the first node subset according to at least one decision variable, and to construct nodes in the second node subset according to at least one constraint; a second construction module, used to construct edges between nodes in the first node subset and nodes in the second node subset according to the association relationship between at least one decision variable and at least one constraint; a determination module, used to determine a bipartite graph based on the first node subset, the second node subset and the edges.

[0130] Optionally, in the processing device for the unit combination in the power system provided in Example 3 of the present application, the first construction module includes: a first processing sub-module, used to take a decision variable in at least one decision variable as a node in a first node subset; and a first determination sub-module, used to determine the variable information of the decision variable as an attribute of the node corresponding to the decision variable.

[0131] Optionally, in the processing device for the unit combination in the power system provided in Example 3 of the present application, the target mathematical model also includes an objective function, and the first determination submodule includes: a second determination submodule, which is used to determine the variable information of the decision variable and the coefficient of the decision variable in the objective function as the attributes of the node corresponding to the decision variable.

[0132] Optionally, in the processing device for the unit combination in the power system provided in Example 3 of the present application, the first construction module includes: a second processing sub-module, used to take the target parameter of the constraint condition in at least one constraint condition as a node in the second node subset; and a third determination sub-module, used to determine the parameter value of the target parameter of the constraint condition as the attribute of the node corresponding to the constraint condition.

[0133] Optionally, in the processing device for the combination of units in the power system provided in Example 3 of the present application, the processing device for the combination of units in the power system also includes: a third acquisition unit, used to acquire multiple sets of hyperparameters; a fourth determination unit, used to determine, for the hyperparameters in the multiple sets of hyperparameters, the solution time of the sample mathematical model by the solver under the hyperparameters; and a fifth determination unit, used to determine the hyperparameters corresponding to the sample mathematical model from the multiple sets of hyperparameters based on the corresponding solution times of the sample mathematical model under the multiple sets of hyperparameters.

[0134] It should be noted that the first acquisition unit 701, the first determination unit 702, and the processing unit 703 correspond to steps S201 to S203 in Example 1, and the examples and application scenarios implemented by the above units and the corresponding steps are the same, but are not limited to the contents disclosed in the above Example 1. It should be noted that the above modules as part of the device can be run in the computer terminal 10 provided in Example 1.

[0135] Optionally, in the processing device for the unit combination in the power system provided in the third embodiment of the present application, the first acquisition unit 701 acquires the following data: Acquire first feature information corresponding to a sample mathematical model among the multiple sample mathematical models; Extracting features from the target mathematical model using a feature extraction model to obtain second feature information; Determine the similarity between the target mathematical model and the multiple sample mathematical models according to the second feature information and the first feature information corresponding to the multiple sample mathematical models.

[0136] Optionally, in the processing device for the unit combination in the power system provided in the third embodiment of the present application, the feature extraction model method acquired by the first acquisition unit 701 is specifically as follows: For a sample mathematical model among the plurality of sample mathematical models, determining a target solving time for the solver to solve the sample mathematical model; Determine a training sample according to a sample mathematical model among the multiple sample mathematical models, and use the target solution time as a true label to construct a training sample set; An initial feature extraction model is trained according to the training sample set to obtain the feature extraction model.

[0137] Optionally, in the processing device for a unit combination in a power system provided in Embodiment 3 of the present application, the feature extraction model acquired by the first acquisition unit 701 is a graph neural network model, and the feature extraction of the target mathematical model is performed by the feature extraction model to obtain the second feature information including: Determining a bipartite graph according to the target mathematical model; The feature extraction model is used to extract features from the bipartite graph to obtain the second feature information.

[0138] Optionally, in the processing device for unit combination in the power system provided in the third embodiment of the present application, the target mathematical model determined by the first determination unit 702 includes: constructing nodes in the first node subset according to the at least one decision variable, and constructing nodes in the second node subset according to the at least one constraint condition; Building edges between nodes in the first node subset and nodes in the second node subset according to an association relationship between the at least one decision variable and the at least one constraint condition; The bipartite graph is determined based on the first subset of nodes, the second subset of nodes, and the edges.

[0139] Optionally, in the processing device for a unit combination in a power system provided in Embodiment 3 of the present application, the nodes in the first node subset acquired by the first acquisition unit 701 include: using a decision variable in the at least one decision variable as a node in the first node subset; The variable information of the decision variable is determined as the attribute of the node corresponding to the decision variable.

[0140] Optionally, in the processing device for unit combination in the power system provided in the third embodiment of the present application, the attributes of the nodes corresponding to the decision variables determined by the first determination unit 702 include: The variable information of the decision variable and the coefficient of the decision variable in the objective function are determined as the attributes of the node corresponding to the decision variable.

[0141] Optionally, in the processing device for a unit combination in a power system provided in Embodiment 3 of the present application, the nodes in the second node subset acquired by the first acquisition unit 701 include: using a target parameter of a constraint condition in the at least one constraint condition as a node in the second node subset; The parameter value of the target parameter of the constraint condition is determined as the attribute of the node corresponding to the constraint condition.

[0142] Optionally, in the processing device for unit combination in the power system provided in the third embodiment of the present application, the first determining unit 702 determines the hyperparameter corresponding to the mathematical model in the following manner: Get multiple sets of hyperparameters; For a hyperparameter in the plurality of sets of hyperparameters, determining a time duration for the solver to solve the sample mathematical model under the hyperparameter; According to the solution time corresponding to the sample mathematical model under the multiple sets of hyperparameters, the optimal hyperparameter corresponding to the sample mathematical model is determined from the multiple sets of hyperparameters.

[0143] It should be noted that the preferred implementation scheme involved in the above embodiments of the present application is the same as the scheme provided in Example 1, as well as the application scenario and implementation process, but is not limited to the scheme provided in Example 1. Example 4

[0144] The embodiment of the present application may provide an electronic device, which may be any electronic device in a group of electronic devices. Optionally, in this embodiment, the electronic device may also be replaced by a terminal device such as a mobile terminal.

[0145] Optionally, in this embodiment, the electronic device may be located in at least one network device among a plurality of network devices of a computer network.

[0146] In this embodiment, the above-mentioned electronic device can execute the program code of the following steps in the method for processing unit combination in the power system: obtaining a target mathematical model corresponding to the target unit combination problem in the power system; determining the target sample mathematical model from multiple sample mathematical models based on the similarity between the target mathematical model and the sample mathematical models corresponding to multiple sample unit combination problems, wherein the hyperparameters corresponding to the sample mathematical model are obtained by screening according to the solution time corresponding to the sample mathematical model; using the hyperparameters corresponding to the target sample mathematical model as the hyperparameters of the solver, and solving the target mathematical model through the solver to obtain a solution result for the target unit combination problem.

[0147] The above-mentioned computer terminal can also execute the program code of the following steps in the method for determining the target object: before determining the target sample mathematical model from multiple sample mathematical models based on the similarity between the target mathematical model and the sample mathematical models corresponding to the combination problems of multiple sample units, obtaining the first feature information corresponding to the sample mathematical model in the multiple sample mathematical models; extracting features from the target mathematical model through the feature extraction model to obtain the second feature information; and determining the similarity between the target mathematical model and the multiple sample mathematical models based on the second feature information and the first feature information corresponding to the multiple sample mathematical models.

[0148] The above-mentioned computer terminal can also execute the program code of the following steps in the method for determining the target object: for a sample mathematical model among multiple sample mathematical models, determine the target solution time of the sample mathematical model by the solver under preset hyperparameters; determine the training samples according to the sample mathematical models among the multiple sample mathematical models, and use the target solution time as the true label to construct a training sample set; train the initial feature extraction model according to the training sample set to obtain a feature extraction model.

[0149] The above-mentioned computer terminal can also execute the program code of the following steps in the method for determining the target object: the feature extraction model is a graph neural network model, and the bipartite graph is determined according to the target mathematical model; features of the bipartite graph are extracted through the feature extraction model to obtain second feature information.

[0150] The above-mentioned computer terminal can also execute the program code of the following steps in the method for determining the target object: the target mathematical model includes at least one constraint and at least one decision variable, constructs nodes in the first node subset according to the at least one decision variable, and constructs nodes in the second node subset according to the at least one constraint; constructs edges between nodes in the first node subset and nodes in the second node subset according to the association relationship between at least one decision variable and at least one constraint; determines a bipartite graph according to the first node subset, the second node subset and the edges.

[0151] The computer terminal can also execute the program code of the following steps in the method for determining the target object: using a decision variable in at least one decision variable as a node in the first node subset; and determining variable information of the decision variable as an attribute of the node corresponding to the decision variable.

[0152] The above-mentioned computer terminal can also execute the program code of the following steps in the method for determining the target object: the target mathematical model also includes an objective function, and the variable information of the decision variable and the coefficient of the decision variable in the objective function are determined as the attributes of the node corresponding to the decision variable.

[0153] The above-mentioned computer terminal can also execute the program code of the following steps in the method for determining the target object: using the target parameter of at least one constraint as a node in the second node subset; determining the parameter value of the target parameter of the constraint as the attribute of the node corresponding to the constraint.

[0154] The above-mentioned computer terminal can also execute the program code of the following steps in the method for determining the target object: obtaining multiple sets of hyperparameters; for the hyperparameters in the multiple sets of hyperparameters, determining the solution time of the sample mathematical model by the solver under the hyperparameters; and determining the hyperparameters corresponding to the sample mathematical model from the multiple sets of hyperparameters based on the corresponding solution time of the sample mathematical model under the multiple sets of hyperparameters.

[0155] Optionally, Figure 8 is a structural block diagram of an electronic device provided according to Embodiment 4 of the present application. Figure 8 As shown, the electronic device 20 may include: one or more ( Figure 8 Only one is shown in the figure) processor 202, memory 204. The electronic device 20 may also include a storage controller, through which the memory 204 is controlled and managed; the electronic device 20 may also include a peripheral interface, through which the radio frequency module, audio module and display screen are connected.

[0156] Among them, the memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the method and device for processing the combination of units in the power system in the embodiment of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned method for processing the combination of units in the power system. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0157] The processor can call the information and application programs stored in the memory through the transmission device to execute the following steps: obtain the target mathematical model corresponding to the target unit combination problem in the power system; determine the target sample mathematical model from multiple sample mathematical models based on the similarity between the target mathematical model and the sample mathematical models corresponding to the multiple sample unit combination problems, wherein the hyperparameters corresponding to the sample mathematical model are obtained by screening according to the solution time corresponding to the sample mathematical model; use the hyperparameters corresponding to the target sample mathematical model as the hyperparameters of the solver, and solve the target mathematical model through the solver to obtain the solution result for the target unit combination problem.

[0158] Optionally, the processor may also execute the program code of the following steps: before determining the target sample mathematical model from multiple sample mathematical models based on the similarity between the target mathematical model and the sample mathematical models corresponding to the combination problems of multiple sample units, obtaining first feature information corresponding to a sample mathematical model in multiple sample mathematical models; performing feature extraction on the target mathematical model through a feature extraction model to obtain second feature information; and determining the similarity between the target mathematical model and multiple sample mathematical models based on the second feature information and the first feature information corresponding to the multiple sample mathematical models.

[0159] Optionally, the processor may also execute the program code of the following steps: for a sample mathematical model among multiple sample mathematical models, determining a target solution time of the sample mathematical model by the solver under preset hyperparameters; determining training samples according to the sample mathematical models among multiple sample mathematical models, and taking the target solution time as a true label to construct a training sample set; training an initial feature extraction model according to the training sample set to obtain a feature extraction model.

[0160] Optionally, the processor may also execute the program code of the following steps: the feature extraction model is a graph neural network model, and a bipartite graph is determined according to a target mathematical model; features of the bipartite graph are extracted through the feature extraction model to obtain second feature information.

[0161] Optionally, the processor may also execute program code of the following steps: the target mathematical model includes at least one constraint and at least one decision variable, nodes in a first node subset are constructed according to at least one decision variable, and nodes in a second node subset are constructed according to at least one constraint; edges are constructed between nodes in the first node subset and nodes in the second node subset according to an association relationship between at least one decision variable and at least one constraint; a bipartite graph is determined according to the first node subset, the second node subset, and the edges.

[0162] Optionally, the processor may further execute program code of the following steps: using a decision variable in at least one decision variable as a node in the first node subset; and determining variable information of the decision variable as an attribute of the node corresponding to the decision variable.

[0163] Optionally, the processor may also execute program code of the following steps: the target mathematical model also includes an objective function, and variable information of the decision variable and the coefficient of the decision variable in the objective function are determined as attributes of the node corresponding to the decision variable.

[0164] Optionally, the processor may also execute program code of the following steps: using a target parameter of at least one constraint as a node in the second node subset; and determining a parameter value of the target parameter of the constraint as an attribute of the node corresponding to the constraint.

[0165] Optionally, the processor may also execute the program code of the following steps: obtaining multiple sets of hyperparameters; for the hyperparameters in the multiple sets of hyperparameters, determining the solution time of the sample mathematical model by the solver under the hyperparameters; and determining the hyperparameters corresponding to the sample mathematical model from the multiple sets of hyperparameters based on the corresponding solution time of the sample mathematical model under the multiple sets of hyperparameters.

[0166] It can be understood by those skilled in the art that Figure 8 The structure shown is for illustration only, and the electronic device may also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (Mobile Internet Devices, MID), a PAD, and other terminal devices. Figure 8 The structure of the electronic device is not limited. Figure 8 More or fewer components (such as network interfaces, display devices, etc.) shown in, or with Figure 8Different configurations are shown.

[0167] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc. Example 5

[0168] The embodiment of the present application further provides a computer program product. Optionally, in this embodiment, the computer program product may include a computer program, and when the computer program is executed by a processor, the method for processing the unit combination in the power system provided in the first embodiment above is implemented.

[0169] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0170] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0171] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0172] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0173] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0174] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0175] The above are only preferred implementations of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for processing unit combination in a power system, characterized in that: include: Obtaining a target mathematical model corresponding to a target unit commitment problem in a power system; Determining a target sample mathematical model from multiple sample mathematical models according to the similarity between the target mathematical model and the sample mathematical models corresponding to the multiple sample unit combination problems, wherein the optimal hyperparameter corresponding to the sample mathematical model is obtained by screening according to the solution time corresponding to the sample mathematical model; The optimal hyperparameters corresponding to the target sample mathematical model are used as hyperparameters of the solver, and the target mathematical model is solved by the solver to obtain a solution result for the target unit combination problem.

2. The method according to claim 1, characterized in that Before determining the target sample mathematical model from the plurality of sample mathematical models according to the similarity between the target mathematical model and the sample mathematical models corresponding to the plurality of sample unit combination problems, the method further comprises: Acquire first feature information corresponding to a sample mathematical model among the multiple sample mathematical models; Extracting features from the target mathematical model using a feature extraction model to obtain second feature information; Determine the similarity between the target mathematical model and the multiple sample mathematical models according to the second feature information and the first feature information corresponding to the multiple sample mathematical models.

3. The method according to claim 2, characterized in that The feature extraction model is obtained in the following way: For a sample mathematical model among the plurality of sample mathematical models, determining a target solving time for the solver to solve the sample mathematical model; Determine a training sample according to a sample mathematical model among the multiple sample mathematical models, and use the target solution time as a true label to construct a training sample set; An initial feature extraction model is trained according to the training sample set to obtain the feature extraction model.

4. The method according to claim 2, characterized in that: The feature extraction model is a graph neural network model, and the feature extraction model is used to extract features from the target mathematical model to obtain the second feature information including: Determining a bipartite graph according to the target mathematical model; The feature extraction model is used to extract features from the bipartite graph to obtain the second feature information.

5. The method according to claim 4, characterized in that The target mathematical model includes at least one constraint condition and at least one decision variable, and determining the bipartite graph according to the target mathematical model includes: constructing nodes in the first node subset according to the at least one decision variable, and constructing nodes in the second node subset according to the at least one constraint condition; Building edges between nodes in the first node subset and nodes in the second node subset according to an association relationship between the at least one decision variable and the at least one constraint condition; The bipartite graph is determined based on the first subset of nodes, the second subset of nodes, and the edges.

6. The method according to claim 5, characterized in that Constructing nodes in the first node subset according to the at least one decision variable comprises: using a decision variable in the at least one decision variable as a node in the first node subset; The variable information of the decision variable is determined as the attribute of the node corresponding to the decision variable.

7. The method according to claim 6, characterized in that The target mathematical model also includes a target function, and determining the variable information of the decision variable as the attribute of the node corresponding to the decision variable includes: The variable information of the decision variable and the coefficient of the decision variable in the objective function are determined as the attributes of the node corresponding to the decision variable.

8. The method according to claim 5, characterized in that Constructing nodes in the second node subset according to the at least one constraint condition comprises: using a target parameter of a constraint condition in the at least one constraint condition as a node in the second node subset; The parameter value of the target parameter of the constraint condition is determined as the attribute of the node corresponding to the constraint condition.

9. The method according to claim 1, characterized in that: The hyperparameters corresponding to the sample mathematical model are determined in the following way: Get multiple sets of hyperparameters; For a hyperparameter in the plurality of sets of hyperparameters, determining a time duration for the solver to solve the sample mathematical model under the hyperparameter; According to the solution time corresponding to the sample mathematical model under the multiple sets of hyperparameters, the optimal hyperparameter corresponding to the sample mathematical model is determined from the multiple sets of hyperparameters.

10. The method according to claim 1, characterized in that The mathematical model is uploaded by the client to the cloud server for solution calculation, and then the cloud server feeds back the solution result to the client.

11. A processing device for unit combination in a power system, characterized in that: include: A first acquisition unit is used to acquire a target mathematical model corresponding to a target unit combination problem in a power system; A first determination unit is used to determine a target sample mathematical model from multiple sample mathematical models according to a similarity between the target mathematical model and sample mathematical models corresponding to multiple sample unit combination problems, wherein the hyperparameter corresponding to the sample mathematical model is obtained by screening from multiple groups of hyperparameters according to the solution time corresponding to the sample mathematical model; The processing unit is used to use the hyperparameters corresponding to the target sample mathematical model as the hyperparameters of the solver, and solve the target mathematical model through the solver to obtain a solution result for the target unit combination problem.

12. The processing device for unit combination in the power system according to claim 11, characterized in that: The first acquisition unit acquires the following data: Acquire first feature information corresponding to a sample mathematical model among the multiple sample mathematical models; Extracting features from the target mathematical model using a feature extraction model to obtain second feature information; Determine the similarity between the target mathematical model and the multiple sample mathematical models according to the second feature information and the first feature information corresponding to the multiple sample mathematical models.

13. The processing device for unit combination in the power system according to claim 12, characterized in that: The feature extraction model method obtained by the first acquisition unit is specifically as follows: For a sample mathematical model among the plurality of sample mathematical models, determining a target solving time for the solver to solve the sample mathematical model; Determine a training sample according to a sample mathematical model among the multiple sample mathematical models, and use the target solution time as a true label to construct a training sample set; An initial feature extraction model is trained according to the training sample set to obtain the feature extraction model.

14. The processing device for unit combination in the power system according to claim 12, characterized in that: The feature extraction model acquired by the first acquisition unit is a graph neural network model. The feature extraction model is used to extract features from the target mathematical model, and the second feature information obtained includes: Determining a bipartite graph according to the target mathematical model; The feature extraction model is used to extract features from the bipartite graph to obtain the second feature information.

15. The processing device for unit combination in the power system according to claim 14, characterized in that: The target mathematical model obtained by the first determination unit determines the bipartite graph including: constructing nodes in the first node subset according to the at least one decision variable, and constructing nodes in the second node subset according to the at least one constraint condition; Building edges between nodes in the first node subset and nodes in the second node subset according to an association relationship between the at least one decision variable and the at least one constraint condition; The bipartite graph is determined based on the first subset of nodes, the second subset of nodes, and the edges.

16. The processing device for unit combination in the power system according to claim 14, characterized in that: The nodes in the first node subset acquired by the first acquisition unit include: using a decision variable in the at least one decision variable as a node in the first node subset; The variable information of the decision variable is determined as the attribute of the node corresponding to the decision variable.

17. The processing device for unit combination in the power system according to claim 12, characterized in that: The attributes of the node corresponding to the decision variable determined by the first determination unit include: The variable information of the decision variable and the coefficient of the decision variable in the objective function are determined as the attributes of the node corresponding to the decision variable.

18. The processing device for unit combination in the power system according to claim 15, characterized in that: The nodes in the second node subset acquired by the first acquisition unit include: using a target parameter of a constraint condition in the at least one constraint condition as a node in the second node subset; The parameter value of the target parameter of the constraint condition is determined as the attribute of the node corresponding to the constraint condition.

19. The processing device for unit combination in the power system according to claim 12, characterized in that: The first determination unit determines the hyperparameters corresponding to the mathematical model in the following manner: Get multiple sets of hyperparameters; For a hyperparameter in the plurality of sets of hyperparameters, determining a time duration for the solver to solve the sample mathematical model under the hyperparameter; According to the solution time corresponding to the sample mathematical model under the multiple sets of hyperparameters, the optimal hyperparameter corresponding to the sample mathematical model is determined from the multiple sets of hyperparameters.

20. An electronic device, characterized in that: include: A memory storing an executable program; A processor is used to run the program, wherein the program, when running, executes the method for processing unit combination in the power system of any one of claims 1 to 10.

21. A computer program product, characterized in that The invention comprises a computer program, which, when executed by a processor, implements a method for processing a unit combination in a power system according to any one of claims 1 to 10.

Images