Business expansion power supply scheme generation method and device based on artificial intelligence, and electronic equipment
By using artificial intelligence-based methods, combining text recognition, intent recognition, and machine learning to generate power supply schemes, and optimizing the schemes using user feedback, the problems of low efficiency and insufficient user satisfaction in traditional power supply scheme generation are solved, achieving efficient and accurate power supply scheme generation.
Patent Information
- Application Number
- CN202511524087.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-23
- Publication Date
- 2026-02-24
AI Technical Summary
Traditional manual design or simple rule-based power supply scheme generation methods are difficult to efficiently handle the diverse power supply needs of business expansion, lack a deep understanding of user needs, resulting in low scheme generation efficiency and difficulty in balancing reliability and user satisfaction.
We employ an AI-based approach, using text recognition, intent recognition, thought chain reasoning, and machine learning to generate power supply solutions. We then optimize these solutions by incorporating user feedback and reward mechanisms to ensure they meet industry standards and user preferences.
The generated power supply schemes are accurate and meet user needs, improving the reliability of the power supply schemes and user satisfaction, and achieving efficient power supply scheme generation.
Smart Images
Figure CN121563409A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power industry expansion technology, specifically a method, apparatus, and electronic device for generating power supply solutions for industry expansion based on artificial intelligence. Background Technology
[0002] Business expansion application is a full-process service provided by the power supply company to electricity customers when they apply for electricity connection, capacity increase, or change of electricity usage, from application acceptance to meter installation and power connection. Its core objective is to ensure that electricity demand is compliantly connected to the power grid, while guaranteeing the security and reliability of power supply.
[0003] As the demand for power supply services for business expansion becomes more diversified and complex, traditional manual design or simple rule-based power supply solution generation methods are insufficient to efficiently handle the diverse needs of business expansion power supply. This results in problems such as inaccurate demand analysis, low solution generation efficiency, insufficient compatibility with industry standards, and poor economic efficiency. While existing technologies attempt to optimize solutions through basic machine learning models, they lack a deep understanding of users' historical dialogues and optimization of long-chain interactions, making it difficult to balance the reliability of the final solution with user satisfaction.
[0004] Against the backdrop of the integration of power grids and artificial intelligence, utilizing technologies such as natural language processing and large language models to automate the analysis of user demand for power supply solutions for business expansion has become a key direction for improving the intelligence level of power supply solutions. Existing power supply solution generation methods based on static rules or segmented processes mainly rely on fixed specifications and operating standards, inputting characteristics such as the geographical distribution and load distribution of the power system to generate business expansion power supply solutions. Generally, these methods possess strong reliability and stability. However, in increasingly complex business expansion environments, this solution generation depends on fixed rules and models, lacking the ability to continuously optimize based on user needs and preferences, and thus failing to meet the multiple requirements of economic efficiency and high efficiency in business expansion solutions. Summary of the Invention
[0005] This invention provides a method, apparatus, and electronic device for generating business expansion power supply schemes based on artificial intelligence, which overcomes the shortcomings of the prior art. It can effectively solve the problem that existing business expansion power supply scheme generation methods often rely on geographical distribution, load distribution, and other features to generate business expansion power supply schemes, and lack consideration for user needs.
[0006] One of the technical solutions of this invention is achieved through the following measures: a method for generating an industrial expansion power supply scheme based on artificial intelligence, comprising: The system obtains the user's input application text for power supply expansion, performs text recognition processing to obtain basic data for the power supply scheme, which includes power supply category, voltage level, equipment selection, line design, and construction cost. Based on industry standard reasoning and verification of basic data for business expansion power supply schemes, a basic power supply scheme is generated and input into the power supply scheme optimization model to obtain a preliminary power supply scheme. The power supply scheme optimization model is obtained by machine learning using multiple samples. Each sample includes historical basic power supply schemes and corresponding historical implemented power supply scheme label data. We obtain user feedback on the initial power supply plan and introduce a reward mechanism and a near-end strategy optimization algorithm to further optimize the initial power supply plan.
[0007] The following are further optimizations and / or improvements to the above-mentioned technical solution: The above-mentioned acquisition of user-inputted power supply application text, followed by text recognition processing to obtain basic power supply scheme data, includes: Obtain the user-inputted application text for power supply expansion; Input the user's business expansion power supply application text into the intent recognition model to obtain the matching power supply scheme category. The intent recognition model is obtained by learning the initial network model using multiple samples. Each sample includes the user's historical business expansion power supply application text and the corresponding power supply scheme category label data. The initial network model is a network model formed by combining a BERT pre-trained model and a dynamic attention mechanism. Input the user's business expansion power supply application text into the pre-built power supply thinking chain reasoning framework, and call the function to obtain power supply scheme reasoning data.
[0008] The construction process of the above power supply thinking chain reasoning framework includes: A power supply reasoning process is constructed, which establishes a reasoning chain in the order of voltage level, equipment selection, line design, and cost estimation. The reasoning process yields power supply scheme reasoning data, and the formal expression of the reasoning chain is as follows: Voltage level → Equipment selection → Circuit design → Cost estimation.
[0009] The above-mentioned business expansion power supply scheme rule reasoning verifies the basic data of the business expansion power supply scheme, generates a basic power supply scheme, and inputs the basic power supply scheme into the power supply scheme optimization model to obtain a preliminary power supply scheme, including: Based on State Grid standards and power industry specifications, a knowledge graph-based reasoning engine is constructed. The reasoning engine is used to verify the basic data of the business expansion power supply scheme and generate a basic power supply scheme. The basic power supply scheme is input into the power supply scheme optimization model to obtain the preliminary power supply scheme. The power supply scheme optimization model is obtained by learning the initial network model using multiple samples. Each sample includes the label data of the historical basic power supply scheme and the corresponding historical implemented power supply scheme. The initial network model includes the random forest network model, the XGBoost network model, and the MLP network model.
[0010] The above-mentioned user feedback on the initial power supply plan is used to further optimize the initial power supply plan by introducing a reward mechanism and a near-end strategy optimization algorithm, including: Obtain user feedback on the initial power supply plan and obtain corresponding reward values based on the set reward mechanism; By using a near-end policy optimization algorithm and combining it with reward values to optimize the initial power supply strategy, an optimized power supply scheme is obtained.
[0011] The second technical solution of the present invention is achieved through the following measures: an artificial intelligence-based power supply solution generation device, comprising: The interactive analysis unit obtains the power supply application text input by the user, performs text recognition processing to obtain basic data of the power supply scheme, which includes power supply category, voltage level, equipment selection, line design, and construction cost; The scheme generation unit, based on industry standard reasoning and verification of the basic data of the business expansion power supply scheme, generates a basic power supply scheme and inputs the basic power supply scheme into the power supply scheme optimization model to obtain a preliminary power supply scheme. The power supply scheme optimization model is obtained by machine learning using multiple samples. Each sample includes the label data of the historical basic power supply scheme and the corresponding historical implemented power supply scheme. The scheme optimization unit obtains user feedback on the initial power supply scheme and introduces a reward mechanism and a near-end strategy optimization algorithm to further optimize the initial power supply scheme.
[0012] The following are further optimizations and / or improvements to the above-mentioned technical solution: The aforementioned interactive analysis unit includes: The interactive information acquisition module acquires the user's input of the application text for business expansion power supply; The intent recognition module takes the user's business expansion power supply application text as input into the intent recognition model and obtains the matching power supply scheme category. The intent recognition model is learned by using multiple samples to train the initial network model. Each sample includes the user's historical business expansion power supply application text and the corresponding power supply scheme category label data. The initial network model is a network model formed by combining a BERT pre-trained model and a dynamic attention mechanism. The thinking chain reasoning module takes the user's business expansion power supply application text into the pre-built power supply thinking chain reasoning framework and calls the function to obtain power supply scheme reasoning data. The function call module calls functions, including JSON structured API calls and multi-step function calls.
[0013] The above-mentioned scheme generation unit includes: The rule-based reasoning module, based on State Grid standards and power industry specifications, constructs a knowledge graph-based reasoning engine. The reasoning engine is used to verify the basic data of the business expansion power supply scheme and generate a basic power supply scheme. The model optimization module takes the basic power supply scheme as input to the power supply scheme optimization model to obtain the preliminary power supply scheme. The power supply scheme optimization model is obtained by learning the initial network model using multiple samples. Each sample includes the label data of the historical basic power supply scheme and the corresponding historical implemented power supply scheme. The initial network model includes the random forest network model, the XGBoost network model, and the MLP network model.
[0014] The above-mentioned optimization unit includes: The reward analysis module obtains user feedback on the initial power supply plan and generates corresponding reward values based on the set reward mechanism. The strategy optimization module uses a near-end strategy optimization algorithm and combines reward values to optimize the initial power supply strategy and obtain an optimized power supply scheme.
[0015] The third technical solution of the present invention is achieved through the following measures: an electronic device, including a processor and a memory, wherein the memory stores a computer program, which is loaded and executed by the processor to realize the steps in generating an artificial intelligence-based power supply scheme.
[0016] This invention, based on text recognition and machine learning, yields a power supply scheme that conforms to industry rules and meets user needs and preferences. Specifically, it utilizes intent recognition, thought chain reasoning, and function calls to effectively identify explicit and implicit user needs, thereby extracting key information from the power supply application text and reasoning to obtain basic data for the power supply scheme. This provides effective and user-compliant data support for accurately generating the power supply scheme. Based on State Grid standards and power industry specifications, and combined with the basic data for the power supply expansion scheme, a basic power supply scheme is generated. Machine learning is then introduced to derive an initial power supply scheme based on this basic scheme. The entire process is accurate and efficient. Taking into full account user feedback on the initial power supply scheme, a reward mechanism and near-end strategy optimization algorithm are introduced to further optimize the initial power supply scheme, outputting a standardized, economical, and user-preferred final power supply scheme. This improves both the reliability of the power supply scheme and user satisfaction. Attached Figure Description
[0017] Appendix Figure 1 This is a schematic diagram of the process for generating the power supply solution for business expansion provided by the present invention.
[0018] Appendix Figure 2 This is a schematic diagram of the method for obtaining basic data for the power supply scheme provided by the present invention.
[0019] Appendix Figure 3A schematic diagram of the process for obtaining the preliminary power supply scheme provided by the present invention.
[0020] Appendix Figure 4 A schematic diagram of the further optimization method provided by the present invention.
[0021] Appendix Figure 5 This is a schematic diagram of the device for generating the power supply solution for business expansion provided by the present invention. Detailed Implementation
[0022] The present invention is not limited to the following embodiments, and the specific implementation can be determined according to the technical solution of the present invention and the actual situation.
[0023] Those skilled in the art will understand that, unless otherwise stated, in the embodiments of this application, "module" or "unit" refers to a computer program or part of a computer program with a predetermined function, which works together with other related parts to achieve a predetermined goal, and can be implemented, wholly or partially, using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.
[0024] In addition, in the embodiments of this application, "multiple" refers to two or more, and "first" and "second" are used to distinguish descriptions and should not be construed as implying relative importance.
[0025] This application provides a method, apparatus, and electronic device for generating power supply solutions for business expansion based on artificial intelligence. This AI-based power supply solution generation apparatus can be integrated into a computer device, which can be a server, a terminal, or other similar device; it can also be executed jointly by a terminal and a server. The above examples should not be construed as limiting this application.
[0026] The aforementioned terminals may include mobile phones, wearable smart devices, tablets, laptops, personal computers (PCs), and in-vehicle computers, etc., and this application does not limit them. This application also does not limit the number of terminal devices.
[0027] The aforementioned server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. This application does not impose any restrictions on this.
[0028] For example, computer equipment acquires the user's input of a power supply application text, performs text recognition processing to obtain basic power supply plan data, which includes power supply category, voltage level, equipment selection, line design, and construction cost. Based on industry standards, the basic power supply plan data is verified to generate a basic power supply plan, which is then input into a power supply plan optimization model to obtain a preliminary power supply plan. The power supply plan optimization model is obtained through machine learning using multiple samples, each of which includes historical basic power supply plans and corresponding historical implemented power supply plan label data. User feedback on the preliminary power supply plan is obtained, and a reward mechanism and a near-end strategy optimization algorithm are introduced to further optimize the preliminary power supply plan.
[0029] Based on this, the technical solution of this application will be described and explained below with reference to several examples.
[0030] Example 1: As shown in the attached document Figure 1 As shown in the figure, this invention discloses a method for generating a business expansion power supply scheme based on artificial intelligence, including: Step S110: Obtain the power supply application text input by the user, perform text recognition processing on it to obtain basic power supply scheme data, which includes power supply category, voltage level, equipment selection, line design, and construction cost; Step S120: Based on industry standard reasoning, verify the basic data of the business expansion power supply scheme, generate a basic power supply scheme, and input the basic power supply scheme into the power supply scheme optimization model to obtain a preliminary power supply scheme. The power supply scheme optimization model is obtained by machine learning using multiple samples. Each sample includes the label data of the historical basic power supply scheme and the corresponding historical implemented power supply scheme. Step S130: Obtain user feedback on the initial power supply scheme, and introduce a reward mechanism and a near-end strategy optimization algorithm to further optimize the initial power supply scheme.
[0031] This invention discloses an artificial intelligence-based method for generating power supply schemes for business expansion. Based on text recognition and machine learning, it generates a preliminary power supply scheme by combining industry rules. Then, it collects user feedback on the preliminary power supply scheme, introduces a reward mechanism and a near-end strategy optimization algorithm to further optimize the preliminary power supply scheme, and outputs a standardized, economical final power supply scheme that meets user preferences. This not only improves the reliability of the power supply scheme but also enhances user satisfaction.
[0032] Example 2: As shown in the attached document Figure 2 As shown, this embodiment of the invention is a further optimization of the above embodiment, wherein the power supply application text input by the user is obtained, and text recognition processing is performed on it to obtain basic data of the power supply scheme, including: Step S210: Obtain the user-inputted application text for power supply expansion; Step S220: Input the user's business expansion power supply application text into the intent recognition model to obtain the matching power supply scheme category. The intent recognition model is obtained by learning the initial network model using multiple samples. Each sample includes the user's historical business expansion power supply application text and the corresponding power supply scheme category label data. The initial network model is a network model formed by combining the BERT pre-trained model and the dynamic attention mechanism.
[0033] In this step, the initial network model is a combination of a BERT pre-trained model and a dynamic attention mechanism. Specifically, the BERT pre-trained model uses a multi-layer Transformer architecture to build global context awareness capabilities, capturing the semantic relationships between explicit statements (e.g., "requires high-reliability power supply") and implicit requirements (e.g., "cannot be interrupted" implies a requirement for backup power) in the user's power supply application text. Through dynamic weight allocation by the self-attention layer, it automatically focuses on key information (e.g., industry terminology, numerical parameters, constraints) and suppresses irrelevant noise (e.g., redundant descriptions or colloquial expressions), thereby capturing key information in the user's context and improving contextual understanding. This enables the tracking of long user dialogue chains, accurate identification of the needs of the power supply solution, and determination of the power supply solution category, which includes new power supply, capacity expansion, power supply path change, and cost consultation.
[0034] The learning process of the intent recognition model is a general machine learning process, which involves acquiring multiple samples and dividing them into training and test sets according to a certain ratio. The initial network model is trained using the training set. Training ends when the stopping condition is met (which may be the maximum number of iterations or the loss function stabilization), resulting in the intent recognition model. The trained intent recognition model is then tested using the test set to optimize the model parameters and output an intent recognition model that meets the test evaluation requirements.
[0035] For example: (1) The user inputs the power supply application text for business expansion as “I want to build a new 500kW load factory and need a power supply plan”. When it is input into the intent recognition model, the output result is “Add power supply”.
[0036] (2) The user inputs the power supply application text "My factory originally had 200kW, and now it needs to be increased to 600kW". When it is input into the intent recognition model, the output result is "increase capacity".
[0037] (3) The user inputs the application text for power supply expansion, which is “The existing power supply scheme is too expensive to construct. Is there a cheaper scheme?”. When this text is input into the intent recognition model, the output result is “Scheme optimization”.
[0038] Step S230: Input the user-inputted power supply application text into the pre-built power supply thinking chain reasoning framework, and call the function to obtain power supply scheme reasoning data. The construction process of the power supply thinking chain reasoning framework includes: A power supply reasoning process is constructed, which establishes a reasoning chain in the order of voltage level, equipment selection, line design, and cost estimation. The reasoning process yields power supply scheme reasoning data, and the formal expression of the reasoning chain is as follows: Voltage level → Equipment selection → Circuit design → Cost estimation.
[0039] It should also be noted that this embodiment is based on the State Grid Guangming Big Model or Big Language Model (LLM), and uses Natural Language Processing (NLP) to parse the user's input of the business expansion power supply application text (such as explicit requirements such as power capacity, load characteristics, and industry attributes). At the same time, it combines historical case databases and real-time data (such as implicit constraints such as regional power grid load and electricity price policies), and extracts deep intent through multi-turn dialogue or document analysis to decompose complex business expansion tasks. It also uses knowledge graphs to construct logical relationships between power supply scheme elements (such as voltage level → equipment selection → line design → cost estimation) to form a traceable reasoning chain. Furthermore, it can optimize the reasoning path through dynamic feedback mechanisms (such as expert verification or user correction) to realize the mapping from fuzzy requirements to precise solutions, and gradually reason out the optimal solution, thereby improving the interpretability and accuracy of decision-making.
[0040] In the above power supply reasoning framework, the reasoning process for voltage level involves identifying the power supply category (which can be directly obtained from step 220), extracting key parameters, calculating the voltage level, and determining the voltage level; for example: The user entered the power supply application text for business expansion as "I want to build a new factory with a load of 500kW and need a power supply solution"; Power supply type identification: This requirement is for new power supply; Key parameters extracted: Load 500kW, User type: Industrial user (factory); Calculated voltage level: 500kW load is suitable for 10kV power supply.
[0041] In the aforementioned power supply thought chain reasoning framework, equipment selection, line design, and cost estimation are all completed through function calls. In this embodiment, function calling enables AI to convert natural language input into structured data and call backend business interfaces to execute queries. Specifically: (1) By using JSON structured API calls, the AI model can directly return an executable parameter structure; (2) Multi-step function calls: AI can make cascading API calls, such as calling multiple business interfaces such as the backend GIS interface, power supply equipment database, and cost assessment API, to perform tasks such as geographic information query, equipment information query, and cost estimation.
[0042] For example: GIS query: Call the GIS system to obtain geographic information such as substations and transmission lines, and determine the line design.
[0043] Equipment selection: The power supply equipment database is accessed to match equipment information such as transformer and cable specifications to determine the appropriate equipment.
[0044] Construction cost estimation: Call the cost assessment API to calculate the construction cost of different power supply schemes and determine the cost estimation results.
[0045] Step S240: Combine the power supply scheme category and power supply scheme reasoning data to obtain the corresponding power supply scheme basic data.
[0046] In this embodiment, intent recognition, thought chain reasoning, and function calls are used to effectively identify the explicit and implicit requirements of user needs, thereby extracting key information from the power supply application text and reasoning to obtain basic data for the power supply scheme, providing effective and user-compliant data support for accurately generating the power supply scheme.
[0047] Example 3: As shown in the attached document Figure 3 As shown, this embodiment of the invention is a further optimization of the above embodiment. It verifies the basic data of the business expansion power supply scheme based on rule-based reasoning, generates a basic power supply scheme, and inputs the basic power supply scheme into the power supply scheme optimization model to obtain a preliminary power supply scheme, including: Step S310: Based on the State Grid standards and power industry specifications, construct a knowledge graph-based reasoning engine, use the reasoning engine to verify the basic data of the business expansion power supply scheme, and generate a basic power supply scheme. It should be noted that the knowledge graph-based inference engine enables complex rule-based combined reasoning. When using it, the basic data of the power supply expansion scheme is input into the inference engine to verify voltage levels, equipment selection, and line design. If the verification fails, the error message is replaced using the inference engine's reasoning result. The knowledge graph-based inference engine has inference rules set up, for example: Voltage level matching (i.e., voltage level matching): For example, a 10kV power supply is recommended for a 500kW load, and a 35kV power supply is recommended for loads above 2000kW. Line design optimization: Calculate the shortest path (overhead line or cable) based on GIS data. Equipment selection: Match transformers and switchgear to the load (e.g., an 800kVA transformer is recommended for a 500kW load).
[0048] Step S320: Input the basic power supply scheme into the power supply scheme optimization model to obtain the preliminary power supply scheme. The construction process of the power supply scheme optimization model is a general supervised learning process, which will not be described in detail. It should be noted that the initial network model used in the construction of the power supply scheme optimization model can be, but is not limited to, random forest network model, XGBoost network model, MLP network model, etc., and key parameters (load, voltage, equipment selection, etc.) can be extracted using automatic feature engineering.
[0049] This embodiment is based on the State Grid standards and power industry specifications, combined with the basic data of the expanded power supply scheme, to generate a basic power supply scheme, and introduces machine learning to obtain an initial power supply scheme based on the basic power supply scheme. The whole process is accurate and efficient.
[0050] Example 4: As shown in the appendix Figure 4 As shown, the embodiments of the present invention are further optimizations of the above embodiments, wherein user feedback on the initial power supply scheme is obtained, and a reward mechanism and a near-end strategy optimization algorithm are introduced to further optimize the initial power supply scheme, including: Step S410: Obtain user feedback on the preliminary power supply plan and obtain the corresponding reward value based on the set reward mechanism; The above-mentioned reward mechanism may include, but is not limited to: If the user accepts the initial power supply plan, a positive reward of +1 will be given; If a user adjusts the initial power supply plan, a negative reward of -1 will be given; If the user clearly indicates that the initial power supply plan is not feasible, a substantial negative reward of -5 will be given.
[0051] Step S420: Utilize the near-end policy optimization algorithm and combine it with the reward value to optimize the initial power supply strategy and obtain the optimized power supply scheme.
[0052] The above steps specifically include: training an initial power supply strategy using historical power supply schemes (the initial power supply strategy includes inference rules + machine learning); adjusting the initial power supply strategy in conjunction with a reward mechanism; learning user preferences through multiple rounds of iteration to obtain an optimized recommendation strategy and determine the optimized power supply scheme.
[0053] This embodiment fully considers user feedback on the initial power supply scheme, introduces a reward mechanism and a near-end strategy optimization algorithm to further optimize the initial power supply scheme, and outputs a standardized, economical final power supply scheme that meets user preferences, which can improve the reliability of the power supply scheme and enhance user satisfaction.
[0054] Example 5: As shown in the attached document Figure 5 As shown in the figure, an embodiment of the present invention discloses an artificial intelligence-based power supply scheme generation device, comprising: The interactive analysis unit obtains the power supply application text input by the user, performs text recognition processing to obtain basic data of the power supply scheme, which includes power supply category, voltage level, equipment selection, line design, and construction cost; The scheme generation unit, based on industry standard reasoning and verification of the basic data of the business expansion power supply scheme, generates a basic power supply scheme and inputs the basic power supply scheme into the power supply scheme optimization model to obtain a preliminary power supply scheme. The power supply scheme optimization model is obtained by machine learning using multiple samples. Each sample includes the label data of the historical basic power supply scheme and the corresponding historical implemented power supply scheme. The scheme optimization unit obtains user feedback on the initial power supply scheme and introduces a reward mechanism and a near-end strategy optimization algorithm to further optimize the initial power supply scheme.
[0055] The interactive analysis unit includes: The interactive information acquisition module acquires the user's input of the application text for business expansion power supply; The intent recognition module takes the user's business expansion power supply application text as input into the intent recognition model and obtains the matching power supply scheme category. The intent recognition model is learned by using multiple samples to train the initial network model. Each sample includes the user's historical business expansion power supply application text and the corresponding power supply scheme category label data. The initial network model is a network model formed by combining a BERT pre-trained model and a dynamic attention mechanism. The thinking chain reasoning module takes the user's business expansion power supply application text into the pre-built power supply thinking chain reasoning framework and calls the function to obtain power supply scheme reasoning data. The function call module calls functions, including JSON structured API calls and multi-step function calls.
[0056] The scheme generation unit includes: The rule-based reasoning module, based on State Grid standards and power industry specifications, constructs a knowledge graph-based reasoning engine. The reasoning engine is used to verify the basic data of the business expansion power supply scheme and generate a basic power supply scheme. The model optimization module takes the basic power supply scheme as input to the power supply scheme optimization model to obtain the preliminary power supply scheme. The power supply scheme optimization model is obtained by learning the initial network model using multiple samples. Each sample includes the label data of the historical basic power supply scheme and the corresponding historical implemented power supply scheme. The initial network model includes the random forest network model, the XGBoost network model, and the MLP network model.
[0057] The scheme optimization unit includes: The reward analysis module obtains user feedback on the initial power supply plan and generates corresponding reward values based on the set reward mechanism. The strategy optimization module uses a near-end strategy optimization algorithm and combines reward values to optimize the initial power supply strategy and obtain an optimized power supply scheme.
[0058] Example 6: This embodiment of the invention discloses an electronic device, including a processor and a memory. The memory stores a computer program, which is loaded and executed by the processor to implement an artificial intelligence-based power supply scheme generation method.
[0059] The processor described above can be a central processing unit (CPU), a general-purpose processor, a digital signal processor (DSP), an ASIC, an FPGA, or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. It can also be a combination that implements computational functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc. The memory can include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, portable hard drives, magnetic disks, or optical disks.
[0060] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.
[0061] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0062] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0063] The above content is only a specific embodiment of this application, which has strong adaptability and implementation effect. However, the protection scope of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the protection scope of this application. Therefore, equivalent changes made in accordance with the claims of this application are still within the scope of this application.
Claims
1. A method for generating power supply solutions for business expansion based on artificial intelligence, characterized in that, include: The system obtains the user's input application text for power supply expansion, performs text recognition processing to obtain basic data for the power supply scheme, which includes power supply category, voltage level, equipment selection, line design, and construction cost. Based on industry standard reasoning and verification of basic data for business expansion power supply schemes, a basic power supply scheme is generated and input into the power supply scheme optimization model to obtain a preliminary power supply scheme. The power supply scheme optimization model is obtained by machine learning using multiple samples. Each sample includes historical basic power supply schemes and corresponding historical implemented power supply scheme label data. We obtain user feedback on the initial power supply plan and introduce a reward mechanism and a near-end strategy optimization algorithm to further optimize the initial power supply plan.
2. The method for generating a business expansion power supply scheme based on artificial intelligence according to claim 1, characterized in that, Obtain the user-inputted power supply application text, perform text recognition processing to obtain basic power supply scheme data, including: Obtain the user-inputted application text for power supply expansion; Input the user's business expansion power supply application text into the intent recognition model to obtain the matching power supply scheme category. The intent recognition model is obtained by learning the initial network model using multiple samples. Each sample includes the user's historical business expansion power supply application text and the corresponding power supply scheme category label data. The initial network model is a network model formed by combining a BERT pre-trained model and a dynamic attention mechanism. Input the user's application text for power supply expansion into the pre-built power supply thought chain reasoning framework, and call the function to obtain power supply scheme reasoning data.
3. The method for generating a business expansion power supply scheme based on artificial intelligence according to claim 2, characterized in that, The construction process of the power supply thought chain reasoning framework includes: A power supply reasoning process is constructed, which establishes a reasoning chain in the order of voltage level, equipment selection, line design, and cost estimation. The reasoning process yields power supply scheme reasoning data, and the formal expression of the reasoning chain is as follows: Voltage level → Equipment selection → Circuit design → Cost estimation.
4. The method for generating an industrial expansion power supply scheme based on artificial intelligence according to claim 1, 2, or 3, characterized in that, Based on the rule-based reasoning of the business expansion power supply scheme, the basic data of the business expansion power supply scheme are verified, a basic power supply scheme is generated, and the basic power supply scheme is input into the power supply scheme optimization model to obtain a preliminary power supply scheme, including: Based on State Grid standards and power industry specifications, a knowledge graph-based reasoning engine is constructed. The reasoning engine is used to verify the basic data of the business expansion power supply scheme and generate a basic power supply scheme. The basic power supply scheme is input into the power supply scheme optimization model to obtain the preliminary power supply scheme. The power supply scheme optimization model is obtained by learning the initial network model using multiple samples. Each sample includes the label data of the historical basic power supply scheme and the corresponding historical implemented power supply scheme. The initial network model includes the random forest network model, the XGBoost network model, and the MLP network model.
5. The method for generating an industrial expansion power supply scheme based on artificial intelligence according to any one of claims 1 to 4, characterized in that, Obtain user feedback on the initial power supply plan, and introduce a reward mechanism and a near-end policy optimization algorithm to further optimize the initial power supply plan, including: Obtain user feedback on the initial power supply plan and obtain corresponding reward values based on the set reward mechanism; By using a near-end policy optimization algorithm and combining it with reward values to optimize the initial power supply strategy, an optimized power supply scheme is obtained.
6. An artificial intelligence-based power supply scheme generation device applying the method described in any one of claims 1 to 5, characterized in that, include: The interactive analysis unit obtains the power supply application text input by the user, performs text recognition processing to obtain basic data of the power supply scheme, which includes power supply category, voltage level, equipment selection, line design, and construction cost; The scheme generation unit, based on industry standard reasoning and verification of the basic data of the business expansion power supply scheme, generates a basic power supply scheme and inputs the basic power supply scheme into the power supply scheme optimization model to obtain a preliminary power supply scheme. The power supply scheme optimization model is obtained by machine learning using multiple samples. Each sample includes the label data of the historical basic power supply scheme and the corresponding historical implemented power supply scheme. The scheme optimization unit obtains user feedback on the initial power supply scheme and introduces a reward mechanism and a near-end strategy optimization algorithm to further optimize the initial power supply scheme.
7. The artificial intelligence-based power supply scheme generation device according to claim 6, characterized in that, Interactive analysis unit, including: The interactive information acquisition module acquires the user's input of the application text for business expansion power supply; The intent recognition module takes the user's business expansion power supply application text as input into the intent recognition model and obtains the matching power supply scheme category. The intent recognition model is learned by using multiple samples to train the initial network model. Each sample includes the user's historical business expansion power supply application text and the corresponding power supply scheme category label data. The initial network model is a network model formed by combining a BERT pre-trained model and a dynamic attention mechanism. The thinking chain reasoning module takes the user's business expansion power supply application text into the pre-built power supply thinking chain reasoning framework and calls the function to obtain power supply scheme reasoning data. The function call module calls functions, including JSON structured API calls and multi-step function calls.
8. The artificial intelligence-based power supply scheme generation device according to claim 6 or 7, characterized in that, The solution generation unit includes: The rule-based reasoning module, based on State Grid standards and power industry specifications, constructs a knowledge graph-based reasoning engine. The reasoning engine is used to verify the basic data of the business expansion power supply scheme and generate a basic power supply scheme. The model optimization module takes the basic power supply scheme as input to the power supply scheme optimization model to obtain the preliminary power supply scheme. The power supply scheme optimization model is obtained by learning the initial network model using multiple samples. Each sample includes the label data of the historical basic power supply scheme and the corresponding historical implemented power supply scheme. The initial network model includes the random forest network model, the XGBoost network model, and the MLP network model.
9. The artificial intelligence-based power supply scheme generation device according to claim 6 or 7, characterized in that, The scheme optimization unit includes: The reward analysis module obtains user feedback on the initial power supply plan and generates corresponding reward values based on the set reward mechanism. The strategy optimization module uses a near-end strategy optimization algorithm and combines reward values to optimize the initial power supply strategy and obtain an optimized power supply scheme.
10. An electronic device, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program, which is loaded and executed by the processor to implement the steps of the method as claimed in any one of claims 1 to 5.