Sunlight procurement demand and supply intelligent matching method for energy-saving reconstruction industry chain

Abstract

This invention discloses a method for intelligent matching of supply and demand in transparent procurement for the energy-saving renovation industry chain, specifically involving the field of procurement management technology. The method includes steps S1: constructing a digital twin profile of a trusted data space; S2: constructing a dynamic feature vector on the demand side; S3: constructing a static feature vector on the supplier side; S4: calculating the supply-demand matching degree based on multi-dimensional feature cross-validation; S5: ranking suppliers in transparent procurement; and S6: quantifying supplier stability fluctuations. This invention calculates the comprehensive matching degree by cross-validating the dynamic feature vector on the demand side with the static feature vector on the supplier side. Based on the procurement cycle, it extracts the historical average matching degree and calculates the stability fluctuation deviation of the current matching degree. By comparing this deviation with a preset threshold to trigger an abnormal fluctuation marker, it reconstructs a dual-criteria decision-making framework of matching degree value and temporal volatility, thereby coupling procurement demand with the matching algorithm and transforming subjective judgment into objective determination.

Description

Technical Field

[0001] This invention relates to the field of procurement management technology, and more specifically, to a transparent procurement supply and demand intelligent matching method for the energy-saving renovation industry chain. Background Technology

[0002] Energy-saving renovation projects are characterized by strong non-standard attributes, specialized technical approaches, long contract periods, and high compliance requirements. Projects funded by public funds or state-owned assets, in particular, have mandatory requirements for transparent procurement processes and risk control. Under the "dual carbon" target (carbon emissions), the building and industrial energy-saving renovation industry chain is growing rapidly. Energy-saving renovation projects typically involve transparent procurement between energy-consuming units (demanders) and energy service companies (suppliers). Energy-consuming units publish procurement needs containing static information such as building age, building envelope, and electromechanical equipment parameters; suppliers publish supply information including energy-saving technologies (such as HVAC renovation, lighting replacement, and waste heat recovery) and historical case studies.

[0003] Existing supply and demand matching methods mostly employ text matching techniques based on keyword extraction, TF-IDF (term frequency-inverse document frequency) algorithms, or traditional semantic models. However, the field of energy-saving renovation has strong engineering and physical attributes. It is not a matter of calculating the matching degree between "demand text" and "supply text," but rather a physical and logical matching between "system energy efficiency baseline state" and "the renovation potential of specific technologies."

[0004] However, it still has some shortcomings in actual use. For example, the existing matching system only focuses on the similarity of the supplier's solution text and the demand text in terms of vocabulary or semantics. However, in the field of energy-saving renovation, even if the supplier's solution text perfectly contains all the technical terms required by the demander, if the power distribution required by the equipment exceeds the transformer capacity of the demand workshop, or if the equipment size exceeds the reserved installation space of the workshop, the solution is still completely infeasible in actual engineering. This leads to the demander finding that it cannot be implemented during the later site survey, resulting in a huge waste of time and communication costs. Existing technologies typically provide only a single matching score after generating the sorted list, without considering the reasonableness and stability of that score. They lack a penetrating early warning mechanism for abnormal fluctuations in supplier stability and historical behavior, making it difficult to meet the risk warning needs of the procurement process. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a transparent procurement supply and demand intelligent matching method for the energy-saving renovation industry chain, which is used to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a transparent procurement supply and demand intelligent matching method for the energy-saving renovation industry chain, comprising the following steps: Step S1: Construct a digital twin profile of the trusted data space: In the trusted execution environment of the blockchain, construct digital twins of the demand node and the supply node.

[0007] Step S2: Construct dynamic feature vectors for the demand side: Obtain the procurement demand text published by the demand side, and through deep analysis and intelligent expansion of demand features, map it into dynamic feature vectors for the demand side that are adapted to intelligent supply and demand matching.

[0008] Step S3: Construct supplier static feature vector: Obtain the supplier's solution profile, and through in-depth analysis and intelligent expansion of capability characteristics, map it into a supplier static feature vector.

[0009] Step S4: Calculation of supply and demand matching degree based on multi-dimensional feature intersection: Based on the dynamic feature vector of the demand side and the static feature vector of the supplier, the supply and demand matching degree is calculated based on the multi-dimensional feature intersection to determine the comprehensive matching degree between the demand side and the supplier.

[0010] Step S5: Sorting Suppliers for Transparent Procurement: Sort the calculated comprehensive matching degree of each supplier in descending order to generate a list of candidate suppliers.

[0011] Step S6: Supplier stability fluctuation quantification: Based on the demander's procurement demand release cycle, calculate whether there are abnormal fluctuations in supplier stability during the current period, and output the final candidate supplier list.

[0012] Preferably, step S1: constructing a digital twin profile of a trusted data space is implemented as follows: All participants (such as buyers and suppliers) join a consortium blockchain network as nodes, and all data is timestamped and digitally signed, forming a trusted and transparent data space; By using the Internet of Things (IoT) gateway, the real-time energy consumption data, equipment operating parameters, building information model, historical production capacity data, and encrypted credit assessment reports of the demand side are combined to generate a dynamically updated digital twin of the demand side. Supplier node digital twin integration technology solution library, equipment performance degradation model, dynamic profit sharing strategy matrix, historical project energy saving verification report, user evaluation and performance record storage.

[0013] Preferably, step S2, constructing the dynamic feature vector on the demand side, is implemented as follows: Step S21: Obtain the procurement request text published by the demander, call a natural language processing tool to perform syntactic structure analysis on the procurement request text, and identify the action words and noun components associated with the action words in the text; Extract the noun phrases that are directly affected or changed by the action words and use them as the core energy consumption phrases; Extract noun phrases or prepositional phrases from the text that modify or limit the execution context of the action word, and use them as limiting conditional phrases; Step S22: Use the core energy consumption phrase and the limiting condition phrase as search terms, and traverse and compare them in all node names of the pre-constructed energy-saving renovation knowledge graph; The text overlap is calculated by determining the proportion of the number of characters that are the same between the search term and each node name to the total number of characters in the search term. The node with the highest text overlap is defined as the target energy efficiency indicator entity corresponding to the core energy consumption phrase and the target built environment entity corresponding to the limiting condition phrase. Step S23: Using the target energy efficiency index entity and the target built environment entity as starting nodes, extract adjacent nodes that have a one-way connection line with the starting node in the knowledge graph; Remove non-technical nodes representing standard numbers and release dates from adjacent nodes, and retain nodes whose connecting line labels are implementation means and point to the target energy efficiency index entity, and define them as associated energy-saving technology entities; Retain nodes whose connection line labels contain devices and are connected to the associated energy-saving technology entities, define them as associated energy-consuming device entities, and summarize the aforementioned retained entities to form a preliminary extended entity pool; Step S24: Traverse each associated energy-saving technology entity and associated energy-consuming equipment entity in the preliminary extended entity pool. In the knowledge graph, query whether there is a connection line with conflicting labels between the entities in the preliminary extended entity pool and the target building environment entity. If there is, delete the corresponding associated energy-saving technology entity or associated energy-consuming equipment entity from the preliminary extended entity pool, and the remaining entities constitute the purification entity pool. Step S25: Assign a unique digital identifier to each associated energy-saving technology entity and associated energy-consuming equipment entity in the purification entity pool; All entities assigned with numerical identifiers are arranged in the order in which they were extracted in step S23 to generate an ordered array of numerical identifiers, thus obtaining the demand-side dynamic feature vector.

[0014] Preferably, step S3: constructing the supplier's static feature vector is implemented as follows: S31: Obtain the solution files submitted by the supplier in the system, including historical project case texts and technical solution specification texts; According to the pre-set text segmentation rules, the solution file is divided into multiple independent text segments. Each text segment is then checked to see if it contains a pre-set category keyword at the beginning. If it does, the text segment is marked as a core technical equipment block and retained; otherwise, it is determined as a non-technical information block and removed. S32: Retrieve a pre-built energy-saving renovation standard entity dictionary, which stores preset energy-saving technology standard names and energy-consuming equipment standard names; With a fixed length as the step size, the truncation window moves from left to right on the text string of the core block of the technical equipment. It is determined whether the string in the current truncation window is completely consistent with a certain standard name in the entity dictionary. If they are completely consistent, the standard name is extracted and defined as the candidate energy-saving technology entity and candidate energy-consuming equipment entity of the supplier, respectively. The character position occupied by the standard name in the text is marked as a matched state, and the matched characters are skipped in subsequent window movements. S33: Obtain the corresponding equipment parameter ledger data submitted by the supplier. The equipment parameter ledger is a two-dimensional data table, which includes a column for equipment name and at least one column for operating parameters. The candidate energy-consuming device entity extracted in step S32 is used as a search term. A string comparison is performed in the device name column of the device parameter ledger. If the search terms are exactly the same, the device operating parameter value in the same data row as the device name is extracted, and the extracted value is bound to the corresponding candidate energy-consuming device entity. S34: Assign a unique technical digital identifier to each candidate energy-saving technology entity, and assign a unique equipment digital identifier to each candidate energy-consuming equipment entity that is bound to the equipment operating parameter value. All entities assigned with digital identifiers are arranged in the order of their first appearance in the core block of the technical equipment described in step S31, generating a one-dimensional ordered array of digital identifiers to obtain a supplier static feature vector. The values ​​of the equipment operating parameters that have a mapping and binding relationship with the equipment digital identifiers are synchronously stored as an auxiliary verification dataset for this vector.

[0015] Preferably, step S4, the calculation of supply and demand matching degree based on multi-dimensional feature intersection, is implemented as follows: S41: Extract the demand-side dynamic feature vector generated in step S25 and the supplier static feature vector generated in step S34; The demand-side dynamic feature vector and the supplier static feature vector are converted into set form, and the numeric identifiers that exist in both sets are extracted to obtain the intersection identifier set. The feature overlap is obtained by the ratio of the number of numeric identifiers contained in the intersection identifier set to the total number of numeric identifiers in the demand-side dynamic feature vector. S42: Extract the auxiliary verification dataset that is stored synchronously with the supplier's static feature vector in step S34, as well as the set of boundary margins entered by the demander when publishing the procurement demand; The set of boundary margins includes the remaining available power distribution capacity and remaining available installation area values ​​entered by the demander when issuing the procurement request. Extract the device operating parameter values ​​associated with the intersection identifier set from the auxiliary verification dataset to obtain the actual required values; The individual space occupancy rate is obtained by using the percentage of the actual demand value to the corresponding category value in the boundary margin set. The overall occupancy rate is calculated by summing and averaging the individual space occupancy rates of all categories. The overall occupancy rate is then converted into a physical reduction value using a preset reduction function. S43: Add the feature overlap degree to the physical reduction value to obtain the final comprehensive matching degree.

[0016] Preferably, step S5: ranking of suppliers in transparent procurement is implemented as follows: The comprehensive matching degree obtained from the supply and demand matching degree calculation step based on multi-dimensional feature intersection is used to sort the comprehensive matching degrees of each supplier in descending order from largest to smallest. The suppliers and their corresponding comprehensive matching degrees are combined and output to generate a candidate supplier list.

[0017] Preferably, step S6: quantifying supplier stability fluctuations is implemented as follows: S61: Obtain the procurement demand release cycle of the demander, divide the release cycle into sub-periods according to equal time, and obtain the comprehensive matching degree of each supplier in the candidate supplier list in each period. Retrieve the historical comprehensive matching scores of the same supplier during the procurement demand release cycle to form a historical comprehensive matching score sequence. ; Calculate the arithmetic mean of the historical comprehensive matching degree sequence to obtain the comprehensive matching degree mean; Calculate the absolute value of the difference between the current period's overall matching degree and the average overall matching degree to obtain the supplier stability fluctuation deviation; S62: Obtain the supplier stability fluctuation deviation and compare it with the preset deviation threshold. If the supplier stability fluctuation deviation is greater than the preset deviation threshold, trigger the supplier stability abnormal fluctuation flag and output the current time period's transparent procurement supplier list with risk high highlight to the demand party; otherwise, it indicates that the supplier has no abnormal fluctuation during the release period, and the supplier is retained. S63: Sort all retained suppliers in descending order of overall matching score to generate the final candidate supplier list.

[0018] The technical effects and advantages of this invention are as follows: 1. This invention provides a transparent procurement supply and demand intelligent matching method for the energy-saving renovation industry chain. By constructing a digital twin profile in a trusted data space, and building digital twins of demand-side nodes and supplier nodes in a blockchain trusted execution environment, it achieves a full-dimensional and tamper-proof digital mirror of both parties, laying a transparent data foundation. It obtains the procurement demand text published by the demand side, and through deep analysis and intelligent expansion of demand features, maps it into a demand-side dynamic feature vector adapted to intelligent supply and demand matching. Through knowledge graph, it automatically expands the procurement demand to the range of technologies / equipment that can solve the demand, realizing reasonable generalization of demand-side features. It obtains the supplier's solution profile, and through deep analysis and intelligent expansion of capability features, maps it into a supplier static feature vector. Based on the demand-side dynamic feature vector and the supplier static feature vector, it calculates the supply and demand matching degree based on multi-dimensional feature cross-validation to determine the comprehensive matching degree between the demand side and the supplier. It not only compares the technical feature vectors in the twins, but also introduces boundary verification to achieve cross-validation of semantic features and physical boundaries. 2. This invention provides a method for intelligent matching of supply and demand in transparent procurement for the energy-saving renovation industry chain. The method calculates the comprehensive matching degree of each supplier and arranges them in descending order. It then outputs a candidate supplier list by combining the suppliers and their corresponding comprehensive matching degrees. Based on the demander's procurement demand release cycle, it calculates the absolute value of the difference between the current comprehensive matching degree and the average comprehensive matching degree to obtain the supplier stability fluctuation deviation. This deviation is compared with a preset deviation threshold. If the supplier stability fluctuation deviation is greater than the preset threshold, an abnormal fluctuation flag is triggered, and a list of transparent procurement suppliers with a high-risk warning is output to the demander for the current period. Conversely, if the deviation is less than the threshold, it indicates that the supplier has no abnormal fluctuations during the release cycle. If a supplier is found to be in violation, that supplier is retained. All retained suppliers are then sorted in descending order of their overall matching degree to generate the final candidate supplier list. This reconstructs the traditional single-criteria sorting that relies solely on the immediate matching degree value into a dual-criteria decision-making framework that comprehensively considers both the magnitude of the matching degree value and its temporal fluctuations. This ensures that the suppliers in the final candidate list not only have solutions that closely match the needs but also demonstrate predictable and stable supply capabilities based on historical behavior. This significantly improves the information richness and structural integrity of the matching decision. By marking abnormal fluctuations and fusing data with the sorted sequence and highlighting them on the front end, deep coupling between audit rules and matching algorithms at the data structure level is achieved, transforming subjective judgments that originally relied on human experience into objective logical judgments. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the intelligent matching method for supply and demand in the transparent procurement of the energy-saving renovation industry chain, as described in this invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please see Figure 1 As shown, this invention provides a transparent procurement supply and demand intelligent matching method for the energy-saving renovation industry chain, including the following steps: Step S1: Constructing a digital twin profile of a trusted data space: In a blockchain trusted execution environment, construct digital twins of demand-side nodes and supply-side nodes.

[0022] In one possible design, step S1: constructing a digital twin profile of a trusted data space is implemented as follows: All participants (such as buyers and suppliers) join a consortium blockchain network as nodes, and all data is timestamped and digitally signed, forming a trusted and transparent data space; By using the Internet of Things (IoT) gateway, the real-time energy consumption data, equipment operating parameters, building information model, historical production capacity data, and encrypted credit assessment reports of the demand side are combined to generate a dynamically updated digital twin of the demand side. Supplier node digital twin integration technology solution library, equipment performance degradation model, dynamic profit sharing strategy matrix, historical project energy saving verification report, user evaluation and performance record storage.

[0023] Step S2: Construct a dynamic feature vector on the demand side: Obtain the procurement demand text published by the demand side, and through deep analysis and intelligent expansion of demand features, map it into a dynamic feature vector on the demand side that adapts to intelligent matching of supply and demand.

[0024] In one possible design, step S2: constructing the dynamic feature vector on the demand side is implemented as follows: Step S21: Obtain the procurement request text published by the demander, call a natural language processing tool to perform syntactic structure analysis on the procurement request text, and identify the action words and noun components associated with the action words in the text; Extract the noun phrases that are directly affected or changed by the action words and use them as the core energy consumption phrases; Extract noun phrases or prepositional phrases from the text that modify or limit the execution context of the action word, and use them as limiting conditional phrases; Step S22: Use the core energy consumption phrase and the limiting condition phrase as search terms, and traverse and compare them in all node names of the pre-constructed energy-saving renovation knowledge graph; The text overlap is calculated by determining the proportion of the number of characters that are the same between the search term and each node name to the total number of characters in the search term. The node with the highest text overlap is defined as the target energy efficiency indicator entity corresponding to the core energy consumption phrase and the target built environment entity corresponding to the limiting condition phrase. Step S23: Using the target energy efficiency index entity and the target built environment entity as starting nodes, extract adjacent nodes that have a one-way connection line with the starting node in the knowledge graph; Remove non-technical nodes representing standard numbers and release dates from adjacent nodes, and retain nodes whose connecting line labels are implementation means and point to the target energy efficiency index entity, and define them as associated energy-saving technology entities; Retain nodes whose connection line labels contain devices and are connected to the associated energy-saving technology entities, define them as associated energy-consuming device entities, and summarize the aforementioned retained entities to form a preliminary extended entity pool; Step S24: Traverse each associated energy-saving technology entity and associated energy-consuming equipment entity in the preliminary extended entity pool. In the knowledge graph, query whether there is a connection line with conflicting labels between the entities in the preliminary extended entity pool and the target building environment entity. If there is, delete the corresponding associated energy-saving technology entity or associated energy-consuming equipment entity from the preliminary extended entity pool, and the remaining entities constitute the purification entity pool. Step S25: Assign a unique digital identifier to each associated energy-saving technology entity and associated energy-consuming equipment entity in the purification entity pool; All entities assigned with numerical identifiers are arranged in the order in which they were extracted in step S23 to generate an ordered array of numerical identifiers, thus obtaining the demand-side dynamic feature vector.

[0025] Step S3: Constructing a static feature vector for suppliers: Obtain the supplier's solution profile, and through in-depth analysis and intelligent expansion of capability characteristics, map it into a static feature vector for suppliers; In one possible design, step S3: constructing the supplier's static feature vector is implemented as follows: S31: Obtain the solution files submitted by the supplier in the system, including historical project case texts and technical solution specification texts; According to the pre-set text segmentation rules, the solution file is divided into multiple independent text segments. Each text segment is then checked to see if it contains a pre-set category keyword at the beginning. If it does, the text segment is marked as a core technical equipment block and retained; otherwise, it is determined as a non-technical information block and removed. S32: Retrieve a pre-built energy-saving renovation standard entity dictionary, which stores preset energy-saving technology standard names and energy-consuming equipment standard names; With a fixed length as the step size, the truncation window moves from left to right on the text string of the core block of the technical equipment. It is determined whether the string in the current truncation window is completely consistent with a certain standard name in the entity dictionary. If they are completely consistent, the standard name is extracted and defined as the candidate energy-saving technology entity and candidate energy-consuming equipment entity of the supplier, respectively. The character position occupied by the standard name in the text is marked as a matched state, and the matched characters are skipped in subsequent window movements. S33: Obtain the corresponding equipment parameter ledger data submitted by the supplier. The equipment parameter ledger is a two-dimensional data table, which includes a column for equipment name and at least one column for operating parameters. The candidate energy-consuming device entity extracted in step S32 is used as a search term. A string comparison is performed in the device name column of the device parameter ledger. If the search terms are exactly the same, the device operating parameter value in the same data row as the device name is extracted, and the extracted value is bound to the corresponding candidate energy-consuming device entity. S34: Assign a unique technical digital identifier to each candidate energy-saving technology entity, and assign a unique equipment digital identifier to each candidate energy-consuming equipment entity that is bound to the equipment operating parameter value. All entities assigned with digital identifiers are arranged in the order of their first appearance in the core block of the technical equipment described in step S31, generating a one-dimensional ordered array of digital identifiers to obtain a supplier static feature vector. The values ​​of the equipment operating parameters that have a mapping and binding relationship with the equipment digital identifiers are synchronously stored as an auxiliary verification dataset for this vector.

[0026] Step S4: Calculation of supply and demand matching degree based on multi-dimensional feature intersection: Based on the dynamic feature vector of the demand side and the static feature vector of the supplier, the supply and demand matching degree is calculated based on the multi-dimensional feature intersection to determine the comprehensive matching degree between the demand side and the supplier. In one possible design, step S4, calculating the supply-demand matching degree based on multi-dimensional feature intersection, is implemented as follows: S41: Extract the demand-side dynamic feature vector generated in step S25 and the supplier static feature vector generated in step S34; The demand-side dynamic feature vector and the supplier static feature vector are converted into set form, and the numeric identifiers that exist in both sets are extracted to obtain the intersection identifier set. The feature overlap is obtained by the ratio of the number of numeric identifiers contained in the intersection identifier set to the total number of numeric identifiers in the demand-side dynamic feature vector. S42: Extract the auxiliary verification dataset that is stored synchronously with the supplier's static feature vector in step S34, as well as the set of boundary margins entered by the demander when publishing the procurement demand; The set of boundary margins includes the remaining available power distribution capacity and remaining available installation area values ​​entered by the demander when issuing the procurement request. Extract the device operating parameter values ​​associated with the intersection identifier set from the auxiliary verification dataset to obtain the actual required values; The individual space occupancy rate is obtained by using the percentage of the actual demand value to the corresponding category value in the boundary margin set. The overall occupancy rate is calculated by summing and averaging the individual space occupancy rates of all categories. The overall occupancy rate is then converted into a physical reduction value using a preset reduction function. S43: Add the feature overlap degree to the physical reduction value to obtain the final comprehensive matching degree; Specifically, when the supplier's solution does not exceed the physical boundary, the overall matching degree is equal to the feature overlap degree, and the original semantic matching ranking is maintained; when the supplier's solution exceeds the physical boundary, the overall matching degree is adjusted downwards based on the excess ratio on the base score, and the supplier with the more severe limitation is placed further down in the ranking list. In this embodiment, it should be specifically explained that the feature overlap degree is obtained by measuring the proportion of the number of numeric identifiers contained in the intersection identifier set to the total number of numeric identifiers in the demand-side dynamic feature vector, specifically as follows: The formula for calculating the feature overlap degree is: ,in, Indicates the degree of feature overlap. This indicates the total number of numeric identifiers in the intersection identifier set, i.e., how many technologies or equipment from the demand side are covered by the supplier's solution. This represents the total number of numeric identifiers in the dynamic feature vector of the demand side, that is, the total number of all technologies or equipment required by the demand side; The process involves summing and averaging the individual space occupancy rates of all categories to obtain the overall occupancy rate. This overall occupancy rate is then converted into a physical reduction value using a preset reduction function. Specifically: The physical reduction value The calculation formula is: When the overall occupancy rate is less than or equal to 1 ; When the overall occupancy rate is greater than 1 ; in, Indicates the physical reduction value. Indicates the overall occupancy rate. This represents the penalty weighting coefficient; Specifically, when the overall occupancy rate A value less than or equal to 1 indicates sufficient physical space, and no deduction will be applied. The overall occupancy rate is 0; A value greater than 1 indicates that a physical space violation has occurred, exceeding the specified proportion. The larger the value, the greater the physical reduction value. The larger the negative absolute value, the greater the value.

[0027] The step of adding the feature overlap degree to the physical reduction value to obtain the final comprehensive matching degree is as follows: The formula for calculating the overall matching degree is: ,in, Indicates the overall matching degree. Indicates the degree of feature overlap. Indicates the physical reduction value; Specifically, in the comprehensive matching degree formula, A decimal number between 0 and 1, representing the basic score of the supplier's solution in meeting the requirements at the textual semantic level; A value of 0 or negative indicates a dynamic deduction from the engineering physics level. When the supplier's solution does not exceed the physical boundary, the overall matching degree is equal to the feature overlap degree, and the original semantic matching ranking is maintained. When the supplier's solution exceeds the physical boundary, the overall matching degree is adjusted downwards based on the excess ratio on the base score, and the supplier with the more severe limitation is placed further down in the ranking list.

[0028] Step S5: Supplier ranking in transparent procurement: Sort the calculated comprehensive matching degree of each supplier in descending order to generate a candidate supplier list.

[0029] In one possible design, step S5: ranking suppliers for transparent procurement is implemented as follows: The comprehensive matching degree obtained from the supply and demand matching degree calculation step based on multi-dimensional feature intersection is used to sort the comprehensive matching degrees of each supplier in descending order from largest to smallest. The suppliers and their corresponding comprehensive matching degrees are combined and output to generate a candidate supplier list.

[0030] Step S6: Supplier stability fluctuation quantification: Based on the demander's procurement demand release cycle, calculate whether there are abnormal fluctuations in supplier stability during the current period, and output the final candidate supplier list.

[0031] In one possible design, step S6: quantification of supplier stability fluctuations is implemented as follows: S61: Obtain the procurement demand release cycle of the demander, divide the release cycle into sub-periods according to equal time, and obtain the comprehensive matching degree of each supplier in the candidate supplier list in each period. Retrieve the historical comprehensive matching scores of the same supplier during the procurement demand release cycle to form a historical comprehensive matching score sequence. ; Calculate the arithmetic mean of the historical comprehensive matching degree sequence to obtain the comprehensive matching degree mean; Calculate the absolute value of the difference between the current period's overall matching degree and the average overall matching degree to obtain the supplier stability fluctuation deviation; S62: Obtain the supplier stability fluctuation deviation and compare it with the preset deviation threshold. If the supplier stability fluctuation deviation is greater than the preset deviation threshold, trigger the supplier stability abnormal fluctuation flag and output the current time period's transparent procurement supplier list with risk high highlight to the demand party; otherwise, it indicates that the supplier has no abnormal fluctuation during the release period, and the supplier is retained. S63: Sort all retained suppliers in descending order of overall matching score to generate the final candidate supplier list.

[0032] The formula for calculating the arithmetic mean is: ,in, This represents the average overall matching score. This represents the overall matching degree of the i-th sub-time period. This indicates the number of sub-time periods.

[0033] In this embodiment, it should be specifically explained that the present invention constructs a digital twin profile in a trusted data space. In a blockchain trusted execution environment, it constructs digital twins of demand-side nodes and digital twins of supplier nodes, realizing a full-dimensional and tamper-proof digital mirror of both supply and demand sides, laying a transparent data foundation. It obtains the procurement demand text published by the demand side, and through deep analysis and intelligent expansion of demand features, maps it into a dynamic feature vector of the demand side that adapts to intelligent supply and demand matching. Through knowledge graph, it automatically expands the procurement demand to the range of technologies / equipment that can solve the demand, realizing reasonable generalization of demand-side features. It obtains the solution profile of the supplier, and through deep analysis and intelligent expansion of capability features, maps it into a static feature vector of the supplier. Based on the dynamic feature vector of the demand side and the static feature vector of the supplier, it calculates the supply and demand matching degree based on the cross-dimensional feature of supply and demand, and determines the comprehensive matching degree between the demand side and the supplier. It not only compares the technical feature vectors in the twin, but also introduces boundary verification to realize the cross-verification of semantic features and physical boundaries. This invention sorts the calculated comprehensive matching scores of each supplier in descending order, and outputs the suppliers and their corresponding comprehensive matching scores to generate a candidate supplier list. Based on the demander's procurement demand release cycle, it calculates the absolute value of the difference between the current comprehensive matching score and the average comprehensive matching score to obtain the supplier stability fluctuation deviation, and compares it with a preset deviation threshold. If the supplier stability fluctuation deviation is greater than the preset deviation threshold, it triggers an abnormal fluctuation flag for the supplier stability, and outputs a list of transparent procurement suppliers with high risk warnings for the current period to the demander. Conversely, if the deviation is less than the preset threshold, it indicates that the supplier has no abnormal fluctuations during the release cycle, and the supplier is retained. The retained supplier list is sorted in descending order of comprehensive matching degree to generate the final candidate supplier list. The traditional single-criteria sorting that only relies on the real-time matching degree value is reconstructed into a dual-criteria decision-making framework that comprehensively considers the magnitude of the matching degree value and the temporal fluctuation of the matching degree. This ensures that the suppliers in the final output candidate list not only have solutions that are highly aligned with the needs, but also demonstrate predictable and stable supply capabilities in their historical behavior. This significantly improves the information richness and structural integrity of the matching decision. By marking abnormal fluctuations and performing data fusion with the sorted sequence and highlighting on the front end, deep coupling between audit rules and matching algorithms is achieved at the data structure level, transforming the subjective judgment that originally relied on human experience into objective logical judgment.

[0034] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A transparent procurement and intelligent supply-demand matching method for the energy-saving renovation industry chain, characterized in that: Includes the following steps: Step S1: Constructing a digital twin profile of a trusted data space: In a blockchain trusted execution environment, construct digital twins of demand-side nodes and supply-side nodes; Step S2: Construct dynamic feature vectors for the demand side: Obtain the procurement demand text published by the demand side, and through deep analysis and intelligent expansion of demand features, map it into dynamic feature vectors for the demand side that are adapted to intelligent matching of supply and demand. Step S3: Construct supplier static feature vector: Obtain supplier solution profiles and map them into supplier static feature vectors through deep analysis and intelligent expansion of capability characteristics; Step S4: Calculation of supply and demand matching degree based on multi-dimensional feature intersection: Based on the dynamic feature vector of the demand side and the static feature vector of the supplier, the supply and demand matching degree is calculated based on the multi-dimensional feature intersection to determine the comprehensive matching degree between the demand side and the supplier. Step S5: Supplier Ranking in Transparent Procurement: Sort the calculated comprehensive matching scores of each supplier in descending order to generate a candidate supplier list; Step S6: Supplier stability fluctuation quantification: Based on the demander's procurement demand release cycle, calculate whether there are abnormal fluctuations in supplier stability during the current period, and output the final candidate supplier list.

2. The intelligent supply and demand matching method for transparent procurement in the energy-saving renovation industry chain as described in claim 1, characterized in that: The implementation method of step S1: constructing a digital twin profile of a trusted data space is as follows: All participants (such as buyers and suppliers) join a consortium blockchain network as nodes, and all data is timestamped and digitally signed, forming a trusted and transparent data space; By using the Internet of Things (IoT) gateway, the real-time energy consumption data, equipment operating parameters, building information model, historical production capacity data, and encrypted credit assessment reports of the demand side are combined to generate a dynamically updated digital twin of the demand side. Supplier node digital twin integration technology solution library, equipment performance degradation model, dynamic profit sharing strategy matrix, historical project energy saving verification report, user evaluation and performance record storage.

3. The intelligent supply and demand matching method for transparent procurement in the energy-saving renovation industry chain as described in claim 1, characterized in that: The implementation method for step S2: constructing the dynamic feature vector on the demand side is as follows: Step S21: Obtain the procurement request text published by the demander, call a natural language processing tool to perform syntactic structure analysis on the procurement request text, and identify the action words and noun components associated with the action words in the text; Extract the noun phrases that are directly affected or changed by the action words and use them as the core energy consumption phrases; Extract noun phrases or prepositional phrases from the text that modify or limit the execution context of the action word, and use them as limiting conditional phrases; Step S22: Use the core energy consumption phrase and the limiting condition phrase as search terms, and traverse and compare them in all node names of the pre-constructed energy-saving renovation knowledge graph; The text overlap is calculated by determining the proportion of the number of characters that are the same between the search term and each node name to the total number of characters in the search term. The node with the highest text overlap is defined as the target energy efficiency indicator entity corresponding to the core energy consumption phrase and the target built environment entity corresponding to the limiting condition phrase. Step S23: Using the target energy efficiency index entity and the target built environment entity as starting nodes, extract adjacent nodes that have a one-way connection line with the starting node in the knowledge graph; Remove non-technical nodes representing standard numbers and release dates from adjacent nodes, and retain nodes whose connecting line labels are implementation means and point to the target energy efficiency index entity, and define them as associated energy-saving technology entities; Retain nodes whose connection line labels contain devices and are connected to the associated energy-saving technology entities, define them as associated energy-consuming device entities, and summarize the aforementioned retained entities to form a preliminary extended entity pool; Step S24: Traverse each associated energy-saving technology entity and associated energy-consuming equipment entity in the preliminary extended entity pool. In the knowledge graph, query whether there is a connection line with conflicting labels between the entities in the preliminary extended entity pool and the target building environment entity. If there is, delete the corresponding associated energy-saving technology entity or associated energy-consuming equipment entity from the preliminary extended entity pool, and the remaining entities constitute the purification entity pool. Step S25: Assign a unique digital identifier to each associated energy-saving technology entity and associated energy-consuming equipment entity in the purification entity pool; All entities assigned with numerical identifiers are arranged in the order in which they were extracted in step S23 to generate an ordered array of numerical identifiers, thus obtaining the demand-side dynamic feature vector.

4. The intelligent supply and demand matching method for transparent procurement in the energy-saving renovation industry chain as described in claim 1, characterized in that: The implementation method for step S3: constructing the supplier's static feature vector is as follows: S31: Obtain the solution files submitted by the supplier in the system, including historical project case texts and technical solution specification texts; According to the pre-set text segmentation rules, the solution file is divided into multiple independent text segments. Each text segment is then checked to see if it contains a pre-set category keyword at the beginning. If it does, the text segment is marked as a core technical equipment block and retained; otherwise, it is determined as a non-technical information block and removed. S32: Retrieve a pre-built energy-saving renovation standard entity dictionary, which stores preset energy-saving technology standard names and energy-consuming equipment standard names; With a fixed length as the step size, the truncation window moves from left to right on the text string of the core block of the technical equipment. It is determined whether the string in the current truncation window is completely consistent with a certain standard name in the entity dictionary. If they are completely consistent, the standard name is extracted and defined as the candidate energy-saving technology entity and candidate energy-consuming equipment entity of the supplier, respectively. The character position occupied by the standard name in the text is marked as a matched state, and the matched characters are skipped in subsequent window movements. S33: Obtain the corresponding equipment parameter ledger data submitted by the supplier. The equipment parameter ledger is a two-dimensional data table, which includes a column for equipment name and at least one column for operating parameters. The candidate energy-consuming device entity extracted in step S32 is used as a search term. A string comparison is performed in the device name column of the device parameter ledger. If the search terms are exactly the same, the device operating parameter value in the same data row as the device name is extracted, and the extracted value is bound to the corresponding candidate energy-consuming device entity. S34: Assign a unique technical digital identifier to each candidate energy-saving technology entity, and assign a unique equipment digital identifier to each candidate energy-consuming equipment entity that is bound to the equipment operating parameter value. All entities assigned with digital identifiers are arranged in the order of their first appearance in the core block of the technical equipment described in step S31, generating a one-dimensional ordered array of digital identifiers to obtain a supplier static feature vector. The values ​​of the equipment operating parameters that have a mapping and binding relationship with the equipment digital identifiers are synchronously stored as an auxiliary verification dataset for this vector.

5. The intelligent supply and demand matching method for transparent procurement in the energy-saving renovation industry chain as described in claim 1, characterized in that: The implementation method of step S4, which calculates the supply and demand matching degree based on multi-dimensional feature intersection, is as follows: S41: Extract the demand-side dynamic feature vector generated in step S25 and the supplier static feature vector generated in step S34; The demand-side dynamic feature vector and the supplier static feature vector are converted into set form, and the numeric identifiers that exist in both sets are extracted to obtain the intersection identifier set. The feature overlap is obtained by the ratio of the number of numeric identifiers contained in the intersection identifier set to the total number of numeric identifiers in the demand-side dynamic feature vector. S42: Extract the auxiliary verification dataset that is stored synchronously with the supplier's static feature vector in step S34, as well as the set of boundary margins entered by the demander when publishing the procurement demand; The set of boundary margins includes the remaining available power distribution capacity and remaining available installation area values ​​entered by the demander when issuing the procurement request. Extract the device operating parameter values ​​associated with the intersection identifier set from the auxiliary verification dataset to obtain the actual required values; The individual space occupancy rate is obtained by using the percentage of the actual demand value to the corresponding category value in the boundary margin set. The overall occupancy rate is calculated by summing and averaging the individual space occupancy rates of all categories. The overall occupancy rate is then converted into a physical reduction value using a preset reduction function. S43: Add the feature overlap degree to the physical reduction value to obtain the final comprehensive matching degree.

6. The intelligent supply and demand matching method for transparent procurement in the energy-saving renovation industry chain according to claim 1, characterized in that: The implementation method for step S5: ranking suppliers in the transparent procurement process is as follows: The comprehensive matching degree obtained from the supply and demand matching degree calculation step based on multi-dimensional feature intersection is used to sort the comprehensive matching degrees of each supplier in descending order from largest to smallest. The suppliers and their corresponding comprehensive matching degrees are combined and output to generate a candidate supplier list.

7. The intelligent supply and demand matching method for transparent procurement in the energy-saving renovation industry chain according to claim 1, characterized in that: The implementation method for step S6: quantifying supplier stability fluctuations is as follows: S61: Obtain the procurement demand release cycle of the demander, divide the release cycle into sub-periods according to equal time, and obtain the comprehensive matching degree of each supplier in the candidate supplier list in each period. Retrieve the historical comprehensive matching scores of the same supplier during the procurement demand release cycle to form a historical comprehensive matching score sequence. ; Calculate the arithmetic mean of the historical comprehensive matching degree sequence to obtain the comprehensive matching degree mean; Calculate the absolute value of the difference between the current period's overall matching degree and the average overall matching degree to obtain the supplier stability fluctuation deviation; S62: Obtain the supplier stability fluctuation deviation and compare it with the preset deviation threshold. If the supplier stability fluctuation deviation is greater than the preset deviation threshold, trigger the supplier stability abnormal fluctuation flag and output the list of transparent procurement suppliers with risk high-highlighting prompts for the current period to the demand party. Conversely, if the supplier does not exhibit abnormal fluctuations during the release cycle, then the supplier should be retained. S63: Sort all retained suppliers in descending order of overall matching score to generate the final candidate supplier list.

Images