Intelligent matching method and system based on procurement demand and execution, and storage medium

Abstract

The application discloses an intelligent matching method and system based on procurement demand and execution, and a storage medium, relates to the technical field of information intelligent matching, and comprises the following steps: constructing a dynamic data correlation graph and synchronizing data in real time; constructing a global task scheduling model according to the synchronized data, so as to dynamically distribute tasks based on load; generating a demand sheet according to the synchronized data, and dynamically analyzing and completing semantics for the demand sheet; according to the demand sheet dynamic analysis result, continuing to perform supplier dynamic matching and priority sorting; for the supplier matching demand process exception, difference tracing and hierarchical early warning are performed; and real-time monitoring is performed on dynamic indexes, and automatic collaborative iteration is performed. The technical problem that the intelligent degree of the existing demand procurement matching is not high is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of information intelligent matching, in particular to an intelligent matching method and system based on procurement demand and execution, and a storage medium. BACKGROUND

[0002] At present, with the increasing complexity of enterprise procurement management needs, data-driven intelligent matching has become the key to improving procurement efficiency. However, in the procurement management scenario, the existing technology has room for improvement in many aspects. For example, in terms of data processing, demand lists, procurement records and warehouse data are stored in different databases, and the data synchronization frequency is low (such as batch synchronization every hour), which can easily lead to information lag, and full data repeated comparison increases the consumption of computing power, which is difficult to meet the real-time and efficiency requirements.

[0003] At the same time, in the demand analysis link, traditional fuzzy expressions and implicit conditions have high requirements for traditional rule engines, making it difficult to accurately extract and convert key information into structured data, and in the supplier matching process, the supplier evaluation dimension is relatively single, and dynamic factors such as real-time capacity and delivery flexibility are not fully considered, affecting the optimization effect of the matching result, in addition, in terms of exception handling, the existing technology can usually only prompt surface differences such as quantity inconsistency, and it is difficult to trace the difference root cause such as demand change, supplier omission or warehousing error, resulting in time-consuming and laborious overall problem troubleshooting, low efficiency, and the early warning response mechanism also has a lag, and different severity of abnormalities cannot be handled in a hierarchical manner, thereby affecting the problem solving efficiency. SUMMARY

[0004] Therefore, the present application provides an intelligent matching method and system based on procurement demand and execution, and a storage medium, to solve the above technical problems in the prior art for demand procurement matching.

[0005] In order to achieve the above purpose, the present application provides the following technical solutions:

[0006] An intelligent matching method based on procurement demand and execution, comprising the following steps:

[0007] Constructing a dynamic data correlation graph and synchronizing data in real time and accurately;

[0008] Constructing a global task scheduling model according to the synchronized data to dynamically allocate tasks based on load;

[0009] Generating a demand list according to the synchronized data, and dynamically analyzing and completing the semantics of the demand list;

[0010] According to the dynamic analysis result of the demand list, continue to perform dynamic matching of suppliers and priority sorting;

[0011] Differential tracing and hierarchical early warning are conducted for supplier matching demand process abnormalities;

[0012] Real-time monitoring is conducted for dynamic indexes, and automatic iteration is cooperated.

[0013] On the basis of the above technical solutions, the present application is further described as follows:

[0014] As a further scheme of the present application,

[0015] The dynamic data correlation graph is constructed, and real-time accurate synchronization data is performed, specifically including:

[0016] A graph database is constructed, and the real-time correlation relationship of the demand order-purchase order-supplier-warehouse-area is stored in the graph database to form a dynamic data correlation graph; wherein the demand order is taken as the starting core node, and is connected with the purchase order, the supplier, the warehouse and the area through the correlation edges, and the edge connection represents the correlation relationship between entities, such as the 'creation relationship' between the demand order and the purchase order, the'supply relationship' between the purchase order and the supplier, etc., when a certain demand order involves a specific commodity category, the related supplier and its current available quantity information are quickly located based on the dynamic data correlation graph through the correlation edge;

[0017] A distributed data gateway is configured, and the distributed data gateway is connected with different independent subsystems, including a demand management system, a procurement system, a warehousing system and a supplier system;

[0018] The demand order, the procurement record and the warehousing data multi-source heterogeneous data are collected through the distributed data gateway;

[0019] A CDC (Change Data Capture) model is further introduced, and the multi-source heterogeneous data change in each independent subsystem database is captured in real time by means of the CDC (Change Data Capture) model;

[0020] Specifically, the distributed data gateway listens to the binary log file of each independent subsystem database in real time based on the CDC model, and the data change events related to the log file are captured, and then the data real-time synchronization mechanism is triggered to effectively reduce the information lag delay;

[0021] The incremental data identification rule is constructed based on the real-time synchronization data information, so as to add a unique identifier and a version number to each data, and the CDC model only synchronizes the data items with version number upgrade change; when the current data is created for the first time, the version number is the initial value, and when the data is changed, the version number is automatically upgraded and changed, and the CDC model only compares the current version number with the last synchronized version number, if the data version number does not change, it is not synchronized and transmitted, if the version number is upgraded, only the data items with version number upgrade change are synchronized and transmitted.

[0022] As a further aspect of the present invention

[0023] The step of constructing a global task scheduling model based on synchronization data to dynamically allocate tasks based on load specifically includes:

[0024] A global task scheduling model is constructed, which loads the CPU and memory usage status information of cluster nodes (servers) in real time, and establishes a heartbeat monitoring mechanism with a frequency of once per second.

[0025] Nodes (servers) with a load rate of <30% are designated as idle nodes. The global task scheduling model automatically assigns a task package to an idle node based on the main task queue generated from the synchronized data. The mapping relationship between the task package and the corresponding node is recorded in real time in the "task-node" subgraph of the dynamic data association graph. When the node's parsing times out (>30 seconds) or fails, the subtask is automatically returned to the global task scheduler to be re-entered into the main task queue. At the same time, the "status" attribute of the corresponding task package node in the dynamic data association graph is updated. This ensures that the synchronized data can be reasonably distributed according to the node load, avoids data accumulation, reduces information lag, and guarantees the real-time efficiency of subsequent needs.

[0026] As a further aspect of the present invention

[0027] The process of generating a demand order based on synchronized data, and dynamically parsing and semantically completing the demand order, specifically includes:

[0028] A BERT pre-trained model is constructed, and a demand semantic understanding model is trained based on the BERT pre-trained model and a corpus of historical purchase orders. The demand semantic understanding model is input with demand order text generated by the CDC model synchronously, and the demand semantic understanding model parses and outputs explicit and implicit information in the demand order text. Explicit information includes clearly structured fields such as demand category, quantity, and region, while implicit information includes non-explicit intent fields. Combining current time series, historical collaboration, and special regulations data, an intent inference submodule is further constructed to complete the implicit information, and the overall parsing result after completion is verified for compliance. The expression for further verifying the compliance of the parsing is as follows:

[0029] S=α·P+β·H+γ·C

[0030] In the formula: S is the score for parsing the requirement form text; P is the score for the completeness of explicit information, which is based on the required and optional items in the standard requirement form and compared with the actual extracted explicit information. With 100 points as the benchmark, 30 points are deducted when a required item is missing and 10 points are deducted when an optional item is missing.

[0031] H is the implicit information completion accuracy score, that is, according to the current time sequence, historical cooperation and special provision data, compared with the actual extracted implicit information, then optional trigger human verification process, finally based on 100 points, when there are missing or incorrect items in the comparison result, then deduct 10 points for each item;

[0032] C is the supplier real-time capacity adaptability score, which is combined with historical cooperation data and supplier real-time capacity data to optimize the automatic splitting ratio of demand quantity, that is, when the demand quantity exceeds the capacity of a single supplier, other suppliers are automatically introduced to split the demand quantity in proportion, and the splitting is based on the current adjustable capacity proportion of different suppliers;

[0033] The expression for continuing to calculate the supplier real-time capacity adaptability score C is:

[0034] The supplier real-time capacity adaptability score C = total demand coverage score × 25% + single supplier capacity adaptation score × 30% + delivery time coordination score × 25% + risk dispersion score × 20%

[0035] Among them, the total demand coverage score, that is, the sum of the quantities borne by each supplier must equal the total demand, otherwise it is 0 points; the single supplier capacity adaptation score, that is, the quantity allocated to each supplier must be within its real-time adjustable capacity, otherwise it is 0 points; the delivery time coordination score, that is, the delivery time of each supplier must meet the delivery time limit of the total demand, otherwise it is 0 points;

[0036] The risk dispersion score avoids excessive dependence on a single supplier, and reduces the overall risk caused by sudden capacity failure of a certain supplier by proportional dispersion. Taking the largest supplier as the benchmark, if the allocation of the largest supplier accounts for ≤60% → 100 points, if it accounts for 60%-80% → 50 points, and if it accounts for >80% → 0 points.

[0037] α, β, γ are weight coefficients corresponding to the items, wherein α = 0.4, β = 0.3, γ = 0.3;

[0038] When the demand single text analysis score S is ≥80 points, it is a qualified demand analysis, otherwise the non-full-mark explicit information integrity score P and / or implicit information integrity score H and / or supplier real-time capacity adaptability score C are re-adjusted and analyzed.

[0039] As a further scheme of the present application,

[0040] The demand single is generated according to the synchronization data, and dynamic analysis and semantic completion are performed on the demand single, and specifically further includes:

[0041] A dynamic graph embedding model is constructed to observe whether there is a region with significant demand fluctuation in real time and map it as a graph structure vertex of the dynamic graph embedding model, and continue to embed the three-dimensional coordinate distance static attribute and real-time capacity load rate dynamic attribute based on the dynamic graph embedding model, and according to the dynamic graph embedding model, the quantitative results of "static distance + dynamic load rate" are fused with the analysis results of the demand semantic understanding model according to each 50% weighting, and then the overall allocation proportion is recalculated.

[0042] As a further scheme of the present application,

[0043] According to the demand single dynamic analysis result, the supplier dynamic matching and its priority sorting are continued, specifically including:

[0044] In order to avoid the drawbacks of traditional time sorting matching of procurement demand and suppliers, a supplier dynamic capability matrix and a demand-supplier matching score model are constructed to update the matched supplier capability in real time through the supplier dynamic capability matrix, and to sort the matching degree through the demand-supplier matching score model, and finally through intelligent algorithm dynamic optimization, to ensure that high value and urgent demand is satisfied first;

[0045] Based on the supplier dynamic capability matrix, a dynamic archive package is constructed for each supplier, and the content of the above dynamic archive package includes:

[0046] Instant supply amount S1: the number of goods that the supplier can immediately allocate at present;

[0047] Short-term supply amount S2: the number of goods that the supplier can quickly allocate within a certain period of time;

[0048] Capacity elasticity coefficient E: reflecting the ability of the supplier to respond to demand fluctuations, E=(actual maximum capacity / regular standard capacity)-1, the larger the value, the better the capacity elasticity coefficient E;

[0049] Regional distribution cost D: the comprehensive cost of the supplier supplying goods to the target area;

[0050] Cooperation adaptation degree F: based on historical cooperation data, F=(on-time delivery times / total cooperation times)×0.6+(quality qualified times / total cooperation times)×0.4;

[0051] Continue to establish the following relationship for the demand-supplier matching score model:

[0052] M=M1×0.4+M2×0.6

[0053] In the formula, M is the comprehensive matching score, the higher the comprehensive matching score, the higher the matching degree of the supplier and the demand;

[0054] The basic matching score M1: M1= category matching degree Q x 0.5 + cooperation adaptation degree F x 0.5, wherein the category matching degree Q is determined according to the matching degree of the supply category and the demand category, and is 0 or 1;

[0055] The dynamic matching score M2: M2= supply capacity score A x 0.3 + cost benefit score B x 0.4 + time efficiency satisfaction score T x 0.2 + capacity elasticity score E' x 0.1;

[0056] The supply capacity score A: A= (S1 / demand quantity) x 50 + (S2 / demand quantity) x 50, if the result exceeds 100, it is calculated as 100 points;

[0057] The cost benefit score B: B= 100- (regional distribution cost D / highest acceptable cost) x 100;

[0058] The time efficiency satisfaction score T: T= (actual delivery time / demand delivery time) x 100, if the actual delivery time is greater than the demand delivery time, then T=0;

[0059] The capacity elasticity score E': E'= E x 100, the capacity elasticity coefficient is converted into a score of 0-100 points, and the exceeding part is calculated as 100 points;

[0060] Thus, through multi-dimensional calculation, the comprehensive matching score of the supplier and the demand is obtained, and the higher the comprehensive matching score, the higher the matching priority of the supplier and the demand, and then the priority is sorted;

[0061] In order to cope with real-time changes, a dynamic adjustment mechanism is further set as follows:

[0062] The supply amount adjustment coefficient K1: when the supply amount S1 of the supplier changes by more than 20%, K1= adjusted S1 / unadjusted S1, and at this time the dynamic matching score M2 is updated as M2 x K1;

[0063] The cost adjustment coefficient K2: when the regional distribution cost D changes by more than 15%, K2= unadjusted D / adjusted D, and the dynamic matching score M2 is updated as M2 x K2;

[0064] Emergency demand adjustment: for emergency demand, the comprehensive matching score M is multiplied by an emergency coefficient U, U is valued at 1.2-2.0 according to the demand emergency degree, and the higher the emergency degree, the larger U is;

[0065] When there is a high-value order amount that exceeds a preset threshold, the comprehensive matching score M is calculated, the weight of the cooperation adaptation degree F is increased to 0.6, the weight of the category matching degree Q is reduced to 0.4, and a capacity elasticity coefficient E is introduced for adjustment, that is, M1= (category matching degree Q x 0.4 + cooperation adaptation degree F x 0.6) x (1 + E x 0.2), so as to preferentially guarantee the stability of the high-value order and the ability to respond to fluctuations.

[0066] As a further scheme of the present application,

[0067] The difference tracing and early warning for the process exception of the supplier matching demand specifically comprises:

[0068] A difference feature library is constructed, and difference factor data of each node is collected and quantified in real time based on key nodes of the whole procurement process, the key nodes including a demand node, an order node, a production node, a logistics node and a warehousing node, and the difference factor calculation formula and meaning corresponding to each key node are as follows:

[0069] ① The demand node:

[0070] Demand change frequency: Fv = change frequency / original plan period, used to measure the frequency of demand change;

[0071] Demand description completeness: Cd = complete item number / total item number × 100%, reflecting the detailed degree of demand description, and the difference value ΔCd is 100% - Cd;

[0072] ② The order node:

[0073] Order response timeliness: Tr = supplier confirmation time - order placement time, reflecting the response speed of the supplier to the order;

[0074] Price deviation rate: Rp = (actual unit price - benchmark unit price) / benchmark unit price × 100%, measuring the deviation degree of the actual procurement price from the benchmark price;

[0075] ③ The production node:

[0076] Production scheduling achievement rate: Pc = actual production scheduling quantity / planned production scheduling quantity × 100%, evaluating the completion of the production scheduling plan, and the difference value is 100% - Pc;

[0077] Material complete rate: Cm = complete material item number / total material item number × 100%, indicating the completeness of the materials required for production, and the difference value is 100% - Cm;

[0078] ④ The logistics node:

[0079] Transportation damage rate: Dt = damaged quantity / total transportation quantity × 100%, reflecting the commodity damage condition in the logistics transportation process;

[0080] In-transit delay rate: Dd = (actual in-transit time - standard in-transit time) / standard in-transit time × 100%, reflecting the punctuality of the logistics transportation;

[0081] ⑤ The warehousing node:

[0082] Quantity coincidence rate: Qi = actual warehouse-in quantity / order quantity × 100%, used to judge the consistency of actual warehouse-in quantity and order quantity, and the difference value is 100%-Qi;

[0083] Quality pass rate: Qq = qualified quantity / sampling quantity × 100%, used to evaluate the quality of the warehouse-in goods, and the difference value is 100%-Qq;

[0084] Specification matching rate: Qs = matching specification item number / total specification item number × 100%, used to check the matching degree of warehouse-in goods specification and demand specification, and the difference value ΔQs is 100%-Qs;

[0085] Warehouse timeliness: Ti = actual warehouse-in time - planned warehouse-in time, used to measure the timeliness of warehouse-in goods;

[0086] Build a propagation matrix to calculate the cumulative influence index;

[0087] Based on the difference factors extracted in the above steps, a propagation matrix Mij is further constructed, Mij is the influence coefficient of the ith node to the jth node, for example, M15 corresponds to the influence of ΔCd to ΔQs, and its initial value is determined by the correlation analysis of ΔCd and ΔQs, Dd and Ti factors in historical collaborative data, such as M15 = 0.4;

[0088] Get node sensitivity Sj: The influence degree of each node difference factor in the above steps to the historical data of other node difference factors is trained, such as warehouse-in node S5 = 0.9;

[0089] Calculate single-node propagation intensity Si:

[0090] Si = Σ (node i difference value × Mij × Sj)

[0091] For example, the propagation intensity of demand node to warehouse-in node S1 = ΔCd × M15 × S5;

[0092] Get each node weight Wi: According to the historical data influence range setting of each node difference factor in the above steps, meet ΣWi = 1, for example, warehouse-in node W5 = 0.3;

[0093] Further calculate the cumulative influence index Tc:

[0094] Tc = Σ (Si × Wi)

[0095] The cumulative influence index Tc is the core index to measure the comprehensive influence of each node difference on the whole procurement process, and its value reflects the severity of the difference influence, which is used as the reference basis for difference tracing;

[0096] Generate real-time dynamic early warning threshold, calculate early warning level and push, specifically:

[0097] Directly correlate the node difference factor historical data extracted in the above steps, take the 90th percentile value of the same kind of difference in the past 12 months as the basic early warning line Tbase, the 90th percentile value refers to that in a group of data, 90% of the data is less than or equal to the value, and only 10% of the data is greater than the value;

[0098] Further determine the actual risk coefficient R of the current supply chain, and the specific formula is as follows:

[0099] R = 1 + (0.4 x Dd + 0.3 x ACm + 0.3 x Rp) / 100

[0100] In the formula, the three node difference factors in step v are directly used, including:

[0101] In-transit delay rate Dd: weight 40%, used to reflect logistics risk;

[0102] Material set rate difference ACm: weight 30%, used to reflect production risk;

[0103] Price deviation rate Rp: weight 30%, used to reflect cost risk;

[0104] Then the calculation formula of the final early warning threshold Tfinal is as follows:

[0105] Tfinal = Tbase x R

[0106] When any difference value in the above steps is greater than or equal to Tfinal, or the cumulative influence index Tc is greater than or equal to Tfinal, an early warning is triggered, specifically including:

[0107] When any difference value in the above steps is greater than or equal to 1.5 x Tfinal, and Tc is greater than or equal to 1.5 x Tfinal, or any difference value in the above steps is greater than or equal to 2 x Tfinal, or Tc is greater than or equal to 2 x Tfinal, a first-level early warning is triggered;

[0108] When Tfinal is less than any difference value in the above steps and less than 1.5 x Tfinal, and Tfinal is less than Tc and less than 1.5 x Tfinal, or any difference value in the above steps is greater than or equal to 1.5 x Tfinal, but Tc is less than Tfinal, or Tc is greater than or equal to 1.5 x Tfinal, but any difference value in the above steps is less than Tfinal, a second-level early warning is triggered.

[0109] As a further scheme of the present application,

[0110] The real-time monitoring of the dynamic indicators is automatically coordinated and iterated, specifically including:

[0111] A three-dimensional dynamic indicator system of "efficiency-accuracy-timeliness" is constructed;

[0112] Real-time data of the above indexes are automatically collected in a timely manner to form a dynamic trend chart, wherein the horizontal axis is time, the vertical axis is index value, a reference value line and an actual data value line are marked, the average values of the indexes of the last three monitoring periods are compared with the reference value, and accidental fluctuations are avoided through multi-data point monitoring to trigger false operation;

[0113] The model early warning process is triggered when the threshold is lower than the threshold value;

[0114] The normal fluctuation deviation is assumed to exist, if the window average values of the index values of the last three monitoring periods are compared with the reference value and are within the normal fluctuation deviation, the early warning is not triggered, so as to avoid accidental fluctuations;

[0115] The fluctuation trajectory archive is automatically generated through the above-mentioned triggered model early warning process, and the average values of the last three windows are continuously recorded, if the average values are continuously maintained in the normal fluctuation deviation interval, the stable fluctuation band is marked;

[0116] Based on the triggered model early warning and the stable fluctuation band as optimization basis, automatic collaborative iteration is carried out to reduce the model early warning triggering rate, and the normal fluctuation deviation value is updated according to the business scene change.

[0117] An intelligent matching system according to the intelligent matching method based on procurement demand and execution, the intelligent matching system comprises:

[0118] A data synchronization module is configured to construct a dynamic data correlation graph and synchronize data in real time accurately;

[0119] A task dynamic allocation module is configured to construct a global task scheduling model according to the synchronized data, and dynamically allocate tasks based on load;

[0120] A demand dynamic analysis module is configured to generate a demand sheet according to the synchronized data, and dynamically analyze and complete semantics of the demand sheet;

[0121] A supplier dynamic matching module is configured to continue dynamic matching of suppliers and priority sorting of the suppliers according to the dynamic analysis result of the demand sheet;

[0122] A difference traceability early warning module is configured to trace differences and give graded early warning for abnormal supplier matching demand process;

[0123] A dynamic index monitoring iteration module is configured to monitor dynamic indexes in real time and automatically perform collaborative iteration.

[0124] A computer readable storage medium has computer program instructions stored thereon, and the computer program instructions are executed by a processor to implement the method.

[0125] The present application has the following beneficial effects:

[0126] The method and system can realize real-time and accurate synchronization of data, improve data processing efficiency and high concurrency support capability by constructing a dynamic data correlation graph, collecting multi-source heterogeneous data in combination with a distributed data gateway, capturing data changes in real time with the help of a CDC model, and reducing redundant transmission with the aid of incremental data identification rules; at the same time, by constructing a global task scheduling model, real-time monitoring of node load and dynamic allocation of tasks are realized, efficient task allocation based on load is realized, data accumulation and information lag are avoided, and a demand semantic understanding model is constructed by a BERT pre-training model, combined with an intent reasoning sub-module to complete implicit information and eligibility checking, supplemented by a dynamic graph embedding model to fuse the real-time state of the region, which effectively realizes accurate analysis of demand orders and optimization of allocation ratio, so that demand processing is more in line with actual scenarios, in addition, a supplier dynamic capability matrix and a demand-supplier matching score model can be constructed, multi-dimensional scoring and dynamic adjustment mechanism are adopted, accurate dynamic matching and reasonable priority sorting of suppliers are realized, high-value and urgent demands are prioritized, procurement efficiency is improved, and by constructing a difference feature library to quantify difference factors at key nodes in the whole process, combined with a propagation matrix to calculate cumulative influence index, a dynamic early warning threshold and a hierarchical early warning mechanism are set, difference tracing and hierarchical early warning of procurement process abnormalities are realized, so that risks can be responded to more timely. BRIEF DESCRIPTION OF DRAWINGS

[0127] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced below. The structure, proportion, size, etc. shown in the description are only used to cooperate with the content disclosed in the description, so that those skilled in the art can understand and read. Any modification of structure, change of proportion relationship or adjustment of size, which does not affect the effects and purposes that can be achieved by the present application, should still fall within the scope of the technical content disclosed by the present application.

[0128] Figure 1 The overall flowchart of the intelligent matching method based on procurement demand and execution provided by the embodiment of the present application.

[0129] Figure 2 The architecture principle schematic diagram of the intelligent matching system based on procurement demand and execution provided by the embodiment of the present application.

[0130] Figure 3 The electronic device entity structure schematic diagram of the embodiment of the present application.

[0131] In the drawings, the component list represented by each reference numeral is as follows:

[0132] Data synchronization module 10; task dynamic allocation module 20; demand dynamic analysis module 30; supplier dynamic matching module 40; difference traceability early warning module 50; dynamic index monitoring iteration module 60;

[0133] Electronic device 70: processor 701, memory 702, internal bus 703. DETAILED DESCRIPTION

[0134] The following embodiments of the application are described by specific examples, and those skilled in the art can easily understand other advantages and effects of the application from the disclosure. Obviously, the described embodiments are part of the embodiments of the application, not all. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0135] The terms such as "upper", "lower", "left", "right", "middle" and the like referred to in the specification are only for the convenience of clear description, not to limit the scope of the application which can be implemented, and the change or adjustment of the relative relationship is also considered as the scope of the application which can be implemented without substantial change of technical content.

[0136] As shown in Figure 1 The embodiments of the application provide a procurement demand and execution intelligent matching method, which realizes real-time and accurate synchronization of data, improves data processing efficiency and high concurrency support capability by constructing a dynamic data correlation graph, collecting multi-source heterogeneous data in combination with a distributed data gateway, and capturing data changes in real time with the help of a CDC model, supplemented by incremental data identification rules to reduce redundant transmission; At the same time, by constructing a global task scheduling model, real-time monitoring of node load and dynamic allocation of tasks are realized, efficient task allocation based on load is realized, data accumulation and information lag are avoided, and a demand semantic understanding model is constructed by a BERT pre-training model, combined with an intent reasoning sub-module to complete implicit information and eligibility checking, supplemented by a dynamic graph embedding model to fuse the real-time state of the region, which effectively realizes accurate analysis and distribution ratio optimization of demand single, and makes the demand processing more in line with the actual scene. In addition, a supplier dynamic capability matrix and a demand-supplier matching score model can be constructed to realize accurate dynamic matching and reasonable priority sorting of suppliers by using a multi-dimensional scoring and dynamic adjustment mechanism, which guarantees high-value and urgent demand to be satisfied first, improves procurement efficiency, and quantifies difference factors of key nodes in the whole process by constructing a difference feature library, calculates the cumulative influence index combined with a propagation matrix, sets a dynamic early warning threshold and a hierarchical early warning mechanism, realizes difference traceability and hierarchical early warning of procurement process abnormalities, so as to respond to risks more timely. Specifically, the following steps are included:

[0137] S1: Construct a dynamic data correlation graph and synchronize data in real time and accurately;

[0138] The specific process is:

[0139] A graph database is constructed, and real-time association relationships of "demand order-purchase order-supplier-warehouse-plot" are stored in the graph database to form a dynamic data association graph. The demand order is taken as a starting core node, and is connected with a purchase order, a supplier, a warehouse, and a plot through association edges, and the edges represent the association relationships between entities, such as a "creation relationship" between a demand order and a purchase order, a "supply relationship" between a purchase order and a supplier, and the like. When a demand order involves a specific commodity category, the relevant supplier and the current available quantity information can be quickly located through the association edges based on the dynamic data association graph.

[0140] A distributed data gateway is configured, and the distributed data gateway is connected with different independent subsystems, including but not limited to a demand management system, a purchase system, a warehouse system, and a supplier system.

[0141] The distributed data gateway is used to collect multi-source heterogeneous data such as demand orders, purchase records, and warehouse data. Since the data involved in the purchase process is scattered in different databases (such as MySQL, Oracle, and the like) and business systems, the distributed data gateway can effectively realize unified reception and forwarding of data requests of different independent subsystems, and can convert and process different data formats, such as converting a JSON format of a purchase order into an XML format of a warehouse system. In addition, the distributed data gateway can control data access permissions, such as limiting the supplier system to read only the purchase order data related to it, thereby significantly improving the reading efficiency and reducing the redundant steps, and supporting high concurrency scenarios through distributed node deployment to avoid information lag caused by single point failure.

[0142] A CDC (change data capture) model is further introduced, and the multi-source heterogeneous data changes in the databases of the independent subsystems are captured in real time by means of the CDC model.

[0143] Since traditional data synchronization methods, such as full-scan database, consume a large amount of computing power, and the timing scanning database has data lag, it is difficult to obtain the latest data in time, and therefore, the CDC model can capture the "addition, modification" and other data change operations more accurately and in real time, rather than directly connecting to full data.

[0144] Specifically, the distributed data gateway listens to the binary log files (such as the binlog of MySQL) of the databases of the independent subsystems in real time based on the CDC model, and captures the data change events related to the log files, and then triggers a real-time data synchronization mechanism to effectively reduce the information lag delay.

[0145] For example, when the demand department creates a new purchase order, the CDC model captures the change of "new demand order" in real time and synchronizes it to the dynamic data association graph through the distributed data gateway, ensuring that the supplier matching module can immediately and accurately obtain the latest demand; when the warehousing system records "inbound quantity", the change data is synchronized to the difference feature library in real time through the distributed data gateway, thereby supporting subsequent shortage / surplus early warning.

[0146] As a preferred embodiment, in order to further reduce redundant data transmission, incremental data identification rules are built based on real-time synchronized data information to add a unique identifier (GUID) and version number to each piece of data. The CDC model only directly synchronizes data items whose version numbers have been upgraded or changed.

[0147] Specifically, when the current data is first created, the version number is an initial value. When the data changes, the version number is automatically updated and upgraded. The CDC model only compares the current version number with the version number of the last synchronization. If the data version number has not changed, it is not transmitted synchronously. If the version number is upgraded, only the data items with the upgraded version number are transmitted synchronously. Thus, based on the CDC model capturing data changes, the version number is further used to identify whether synchronization is needed, which significantly reduces the large amount of duplicate transmission caused by traditional full synchronization. This is especially useful in procurement scenarios involving a large amount of order, inventory, and supplier data. For example, repeatedly transmitting an unmodified purchase order would waste bandwidth and computing power, ultimately achieving "real-time, accurate, and low-redundancy" data synchronization.

[0148] S2: Construct a global task scheduling model based on synchronization data to dynamically allocate tasks based on load;

[0149] The specific process is as follows:

[0150] To improve data processing efficiency in a multi-node (server) cluster environment, a global task scheduling model is constructed. The CPU and memory usage status information of the cluster nodes (servers) are loaded in real time through the global task scheduling model, and a heartbeat monitoring mechanism with a frequency of once per second is established.

[0151] Nodes (servers) with a load rate of <30% are assumed to be idle nodes. The global task scheduling model automatically assigns a task package to an idle node based on the main task queue generated from the synchronized data. The mapping relationship between the task package and the corresponding node is recorded in real time in the "task-node" subgraph of the dynamic data association graph. When the node's parsing times out (>30 seconds) or fails, the subtask is automatically returned to the global task scheduler to be re-entered into the main task queue. At the same time, the "status" attribute of the corresponding task package node in the dynamic data association graph is updated. This ensures that the synchronized data can be reasonably distributed according to the node load, avoids data accumulation, reduces information lag, and ensures real-time and efficient operation for subsequent needs.

[0152] S3: generating a demand order according to the synchronization data, and dynamically analyzing and semantically completing the demand order;

[0153] The specific process is as follows:

[0154] A BERT pre-training model is constructed, and a demand semantic understanding model is trained based on the BERT pre-training model combined with a corpus of historical purchase orders. The demand semantic understanding model inputs the demand order text generated by the CDC model synchronization data, and outputs the explicit information and implicit information in the demand order text through the demand semantic understanding model. The explicit information includes demand categories, quantities, and explicit structured fields such as regions, and the implicit information includes non-explicit intent fields (such as delivery time limit associated with “urgent” and specific category completion of “seasonal fruits”); further combining current time sequence, historical collaboration and special provision data, an intent reasoning sub-module is constructed to complete the implicit information, and the overall analysis result after completion is subjected to qualification calculation, for example, combining the current season and regional sales data, “urgent procurement” is analyzed as “delivery time limit ≤ 72 hours”, and “seasonal fruits” is completed as “watermelon, peach”;

[0155] The expression for further calculating the analysis qualification is as follows:

[0156] S = a·P + β·H + γ·C

[0157] In the formula: S is the demand order text analysis score; P is the explicit information integrity score, that is, based on the mandatory items and optional items in the standard demand order, compared with the actual extracted explicit information, taking 100 points as the basis, when the mandatory items are missing, deduct 30 points, when the optional items are missing, deduct 10 points;

[0158] H is the implicit information completion accuracy score, that is, based on the current time sequence, historical collaboration and special provision data, compared with the actual extracted implicit information, and then optionally triggering a human verification process, finally taking 100 points as the basis, when there are missing or incorrect extraction items in the comparison result, deduct 10 points for each item;

[0159] C is the supplier real-time capacity adaptability score, which is optimized by combining historical collaboration data and supplier real-time capacity data to calculate the automatic splitting ratio of demand quantity, that is, when the demand quantity exceeds the capacity of a single supplier, other suppliers are automatically introduced to split the demand quantity in proportion, and the splitting basis is the current adjustable capacity proportion of different suppliers, for example, when the explicit delivery time limit is 3 days and the delivery quantity is 100 units of a certain type of equipment, the 3-day adjustable capacity of a single supplier is 60 units, the 3-day adjustable capacity of the second introduced supplier is 30 units, and the third supplier needs to be introduced, and the 3-day adjustable capacity of the third supplier is 30 units, then the automatic splitting ratio of the supplier is 2:1:1, that is, 50:25:25.

[0160] The expression for continuing to calculate the supplier real-time capacity adaptability score C is:

[0161] Supplier real-time capacity adaptability score C = total demand coverage score × 25% + single supplier capacity adaptation score × 30% + delivery time coordination score × 25% + risk dispersion score × 20%

[0162] Wherein, the total demand coverage score, i.e. the sum of the quantities borne by each supplier must equal the total demand, otherwise 0 points; the single supplier capacity adaptation score, i.e. the quantity allocated to each supplier must be within its real-time adjustable capacity range, otherwise 0 points; the delivery time coordination score, i.e. the delivery time of each supplier must all meet the delivery time limit of the total demand, otherwise 0 points;

[0163] The risk dispersion score avoids excessive dependence on a single supplier, reduces the overall risk caused by sudden capacity failure of a certain supplier through proportional dispersion, and takes the largest supplier as the benchmark. If the allocation of the largest supplier accounts for ≤60% of the total, it is 100 points; if it accounts for 60%-80%, it is 50 points; if it accounts for >80%, it is 0 points.

[0164] α, β, γ are weight coefficients corresponding to the items, wherein α = 0.4, β = 0.3, γ = 0.3;

[0165] When the demand order text analysis score S ≥ 80 points, it is a qualified demand analysis, otherwise the explicit information integrity score P and / or the implicit information integrity score H and / or the supplier real-time capacity adaptability score C are re-adjusted and analyzed;

[0166] For example, in a certain actual application scenario, a demand order text requires the purchase of 100 units of a certain type of equipment, then the demand semantic understanding model first extracts the device model and quantity in the explicit information, but the demand order text does not explicitly specify the delivery time limit, so the demand semantic understanding model automatically completes the delivery time limit as 3 days according to historical collaboration data, and optimizes the automatic splitting ratio of the demand quantity according to the current adjustable capacity ratio of different suppliers, such as 70 units for supplier A and 30 units for supplier B. The above process ensures that the key information of the demand order can be accurately extracted and converted into structured data, turning the original simple demand order of "100 units of X1 equipment" into clear structured data containing "model X1, quantity 100 units, delivery within 3 days, split A supplier 70 units / B supplier 30 units". This provides high-quality input for subsequent supplier matching, avoiding problems such as stockout and delay caused by incomplete information or unreasonable analysis;

[0167] As another preferred embodiment of the present embodiment, a dynamic graph embedding model is constructed to observe in real time whether there is a region with significant demand fluctuation and map it as a graph structure vertex of the dynamic graph embedding model, continue to embed the three-dimensional coordinate distance static attribute and real-time capacity load rate dynamic attribute based on the dynamic graph embedding model, and according to the dynamic graph embedding model, the quantitative results of "static distance + dynamic load rate" are fused with the analysis results of the demand semantic understanding model according to each 50% weighting, and then the overall allocation proportion is recalculated;

[0168] For example, according to the analysis result of the demand semantic understanding model, supplier A allocates 80 units, and supplier B allocates 20 units. At this time, the demand of a certain area increases by 30 units, and requires delivery within 3 days. Both suppliers A and B still have adjustable capacity according to the analysis result of the demand semantic understanding model, so the static attribute of the three-dimensional coordinate distance from the demand area center and the real-time capacity load rate dynamic attribute are further introduced. The static distance of supplier B is closer, so the new demand is preferentially allocated to supplier B, that is, supplier B allocates 50 units. Further, the new overall allocation proportion is 7:3 according to the 50% weighting fusion of the analysis result 8:2 of the demand semantic understanding model. However, when the current load rate of suppliers A and B increases sharply or their adjustable capacity cannot meet the new allocation demand, at this time, the allocation needs to be reduced or avoided, and the quantitative result is calculated according to the real-time capacity proportion of suppliers A and B that can be adjusted, or supplier C is directly introduced, and supplier C allocates 30 units, obtaining a new overall allocation proportion of 8:2:3.

[0169] Through the dynamic graph embedding model, the disadvantages of ignoring the real-time situation of the area by only analyzing the demand single text with BERT are avoided, such as the sharp increase of supplier load rate, the influence of long distance on time efficiency, etc. The real-time state is quantified as data by dynamic graph embedding, and further combined with the demand single text analysis result, so that the overall demand analysis is more flexible and practical, and the delay or resource waste caused by sudden state is reduced.

[0170] S4: According to the demand single dynamic analysis result, continue to perform supplier dynamic matching and priority sorting;

[0171] The specific process is as follows:

[0172] In order to avoid the disadvantages of traditional matching of procurement demand and suppliers only according to time sorting, a supplier dynamic capability matrix and a demand-supplier matching score model are constructed to update the matched supplier capability in real time through the supplier dynamic capability matrix, and to sort the matching degree through the demand-supplier matching score model, and finally to dynamically optimize through intelligent algorithm to ensure that high-value and urgent demand is preferentially met;

[0173] More specifically, a dynamic profile package is constructed for each supplier based on the supplier dynamic capability matrix, and the contents of the dynamic profile package include:

[0174] Instant supply S1: the number of goods that the supplier can immediately allocate at present;

[0175] Short-term supply S2: the number of goods that the supplier can quickly allocate within a certain period of time;

[0176] Elasticity coefficient E of production capacity: reflecting the ability of the supplier to respond to demand fluctuations, E = (actual maximum production capacity / regular standard production capacity)-1, the larger the value, the better the elasticity coefficient E of production capacity;

[0177] Regional distribution cost D: the comprehensive cost of the supplier supplying goods to the target area;

[0178] Cooperation adaptation degree F: based on historical cooperation data, F = (on-time delivery times / total cooperation times) x 0.6 + (quality qualified times / total cooperation times) x 0.4;

[0179] The following relationship is established for the demand-supplier matching score model:

[0180] M = M1 x 0.4 + M2 x 0.6

[0181] In the formula, M is the comprehensive matching score, and the higher the comprehensive matching score, the higher the matching degree of the supplier and the demand;

[0182] Basic matching score M1: M1 = category matching degree Q x 0.5 + cooperation adaptation degree F x 0.5, wherein the category matching degree Q is determined according to the matching degree of the supplier category and the demand category, and takes a value of 0 or 1;

[0183] Dynamic matching score M2: M2 = supply capacity score A x 0.3 + cost benefit score B x 0.4 + time efficiency satisfaction score T x 0.2 + production capacity elasticity score E' x 0.1;

[0184] Supply capacity score A: A = (S1 / quantity of demand) x 50 + (S2 / quantity of demand) x 50, if the result exceeds 100, it is calculated as 100 points;

[0185] Cost benefit score B: B = 100 - (regional distribution cost D / highest acceptable cost) x 100;

[0186] Time efficiency satisfaction score T: T = (actual delivery time / required delivery time) x 100, if the actual delivery time is greater than the required delivery time, then T = 0;

[0187] Production capacity elasticity score E': E' = E x 100, which converts the production capacity elasticity coefficient into a score of 0-100 points, and exceeds 100 points;

[0188] Thus, through multi-dimensional calculation, the comprehensive matching score of the supplier and the demand is obtained, and the higher the comprehensive matching score is, the higher the matching priority of the supplier and the demand is, and then the priority is sorted;

[0189] To cope with real-time changes, a dynamic adjustment mechanism is further provided as follows:

[0190] Supply amount adjustment coefficient K1: when the change rate of the instant supply amount S1 of the supplier exceeds 20%, K1 = adjusted S1 / previous S1, at this time the dynamic matching score M2 is updated as M2 x K1;

[0191] Cost adjustment coefficient K2: when the change rate of the regional distribution cost D exceeds 15%, K2 = previous D / adjusted D, and the dynamic matching score M2 is updated as M2 x K2;

[0192] Emergency demand adjustment: for emergency demand, the comprehensive matching score M needs to be multiplied by an emergency coefficient U, U is valued at 1.2-2.0 according to the demand emergency degree, and the higher the emergency degree is, the larger U is;

[0193] An optional embodiment, when there is a high-value order amount exceeding a preset threshold, the comprehensive matching score M is calculated, the weight of the cooperation adaptation degree F is increased to 0.6, the weight of the category matching degree Q is reduced to 0.4, and the capacity elasticity coefficient E is introduced for adjustment, that is, M1 = (category matching degree Q x 0.4 + cooperation adaptation degree F x 0.6) x (1 + E x 0.2), so as to preferentially guarantee the stability of high-value orders and the ability to respond to fluctuations;

[0194] The above scheme introduces multiple quantitative formulas to make the supplier matching process more scientific and operable, the dynamic capability matrix can ensure the timeliness of data, and the dynamic adjustment mechanism can quickly respond to changes, effectively guaranteeing efficient matching in various scenarios and significantly improving procurement efficiency and satisfaction;

[0195] S5: difference traceability and graded early warning for abnormal supplier matching demand process;

[0196] The specific process is as follows:

[0197] S501: build a difference feature library, and collect and quantify each node difference factor data in real time based on the key nodes of the procurement whole process, the key nodes including demand nodes, order nodes, production nodes, logistics nodes and warehouse nodes, and the difference factor calculation formulas and meanings corresponding to each key node are as follows:

[0198] ① Demand node:

[0199] Demand change frequency: Fv = change frequency / planned period, used to measure the frequency of demand change;

[0200] Requirement description completeness: Cd = complete item number / total item number x 100%, which reflects the detailed degree of requirement description, and the difference value ΔCd is 100%-Cd;

[0201] ② Order node:

[0202] Order response timeliness: Tr = supplier confirmation time - order placement time, which reflects the response speed of the supplier to the order;

[0203] Price deviation rate: Rp = (actual unit price - benchmark unit price) / benchmark unit price x 100%, which measures the deviation degree of actual purchase price from benchmark price;

[0204] ③ Production node:

[0205] Production scheduling achievement rate: Pc = actual production scheduling quantity / planned production scheduling quantity x 100%, which evaluates the completion of production scheduling plan, and the difference value is 100%-Pc;

[0206] Material complete rate: Cm = complete material item number / total material item number x 100%, which indicates the completeness of materials required for production, and the difference value is 100%-Cm;

[0207] ④ Logistics node:

[0208] Transportation damage rate: Dt = damaged quantity / total transportation quantity x 100%, which reflects the damage of goods in the logistics transportation process;

[0209] In-transit delay rate: Dd = (actual in-transit time - standard in-transit time) / standard in-transit time x 100%, which reflects the punctuality of logistics transportation;

[0210] ⑤ Warehouse node:

[0211] Quantity coincidence rate: Qi = actual warehouse quantity / order quantity x 100%, which is used to judge the consistency of actual warehouse quantity and order quantity, and the difference value is 100%-Qi;

[0212] Quality qualification rate: Qq = qualified quantity / sampling quantity x 100%, which evaluates the quality compliance of warehouse goods, and the difference value is 100%-Qq;

[0213] Specification matching rate: Qs = matched specification item number / total specification item number x 100%, which checks the matching degree of warehouse goods specification and demand specification, and the difference value ΔQs is 100%-Qs;

[0214] Warehouse timeliness: Ti = actual warehouse time - planned warehouse time, which measures the timeliness of goods warehouse;

[0215] S502: Construct a propagation matrix and calculate a cumulative influence index;

[0216] Based on the difference factor extracted in step S501, a propagation matrix Mij is further constructed, Mij is the influence coefficient of the ith node to the jth node, for example, M15 corresponds to the influence of ΔCd to ΔQs, and its initial value is determined by the correlation analysis of ΔCd and ΔQs, Dd and Ti factors in historical collaborative data, such as M15 = 0.4;

[0217] Obtain the node sensitivity Sj: The influence degree of each node difference factor in step S501 to the historical data of other node difference factors is trained, such as the warehouse node S5 = 0.9;

[0218] Calculate the single node propagation intensity Si:

[0219] Si = Σ (node i difference value × Mij × Sj)

[0220] For example, the propagation intensity of the demand node to the warehouse node S1 = ΔCd × M15 × S5;

[0221] Obtain each node weight Wi: According to the historical data influence range setting of each node difference factor in step S501, meet ΣWi = 1, for example, the warehouse node W5 = 0.3;

[0222] Further calculate the cumulative influence index Tc:

[0223] Tc = Σ (Si × Wi)

[0224] The cumulative influence index Tc is the core index for measuring the comprehensive influence degree of each node difference on the whole procurement process, and the value size reflects the severity of the difference influence, which is used as the reference basis for difference tracing;

[0225] S503: Generate real-time dynamic early warning threshold, calculate early warning level and push, specifically:

[0226] Directly associate the node difference factor historical data extracted in step S501, take the 90th percentile value of the same kind of difference in the past 12 months as the basic warning line Tbase, the 90th percentile value means that in a group of data, 90% of the data is less than or equal to the value, and only 10% of the data is greater than the value;

[0227] Taking "quantity difference rate (ΔQi)" as an example: if the quantity difference rate data in the past 12 months is: 2, 3, 5, 5, 6, 7, 8, 9, 10, 12, 15, 20 (unit: %), after sorting these data from small to large, the 90th percentile value is the 11th data (12 months × 90% = 10.8, rounding up to the 11th), that is, 15%, then in the past 12 months, 90% of the quantity difference rate is ≤15%, and only 10% of the cases exceeds 15%;

[0228] Further determine the actual risk coefficient R of the current supply chain, the specific formula is as follows:

[0229] R = 1 + (0.4 x Dd + 0.3 x ACm + 0.3 x Rp) / 100

[0230] In the formula, the three node difference factors in step S501 are directly used, including:

[0231] In-transit delay rate Dd: weight 40%, used to reflect logistics risk;

[0232] Material matching rate difference ACm: weight 30%, used to reflect production risk;

[0233] Price deviation rate Rp: weight 30%, used to reflect cost risk;

[0234] Then the calculation formula of the final early warning threshold Tfinal is as follows:

[0235] Tfinal = Tbase x R

[0236] When any difference value in step S501 is greater than or equal to Tfinal, or the cumulative impact index Tc is greater than or equal to Tfinal, an early warning is triggered immediately, specifically including:

[0237] When any difference value in step S501 is greater than or equal to 1.5 x Tfinal, and Tc is greater than or equal to 1.5 x Tfinal, or any difference value in step S501 is greater than or equal to 2 x Tfinal, or Tc is greater than or equal to 2 x Tfinal, a first-level early warning is triggered;

[0238] When Tfinal is less than any difference value in step S501 and less than 1.5 x Tfinal, and Tfinal is less than Tc and less than 1.5 x Tfinal, or any difference value in step S501 is greater than or equal to 1.5 x Tfinal, but Tc is less than Tfinal, or Tc is greater than or equal to 1.5 x Tfinal, but any difference value in step S501 is less than Tfinal, a second-level early warning is triggered;

[0239] S6: Real-time monitoring of dynamic indicators and automatic iterative coordination;

[0240] The specific process is: a three-dimensional dynamic indicator system of "efficiency-precision-timeliness" is constructed;

[0241] Real-time data of the above indicators are automatically collected at regular intervals to form a dynamic trend chart, in which the horizontal axis is time and the vertical axis is the indicator value, and the reference value line and the actual data value line are marked, the average values of the indicators of the last three monitoring periods are compared with the reference value, and accidental fluctuations are avoided to trigger false operations through multiple data point monitoring;

[0242] When it is lower than the threshold, the model early warning process is triggered, for example: the matching accuracy reference value is 95%, the indicator values of the last three monitoring periods are 94%, 93% and 92% respectively, and the window average value 93% < 95% → early warning is triggered;

[0243] Suppose that there is a normal fluctuation deviation of ±1%, if the index values of the last three monitoring periods are 94%, 96%, and 93%, the window mean value is 94.3% ≈ 95% → no triggering, so as to avoid accidental fluctuations;

[0244] The fluctuation trajectory archive is automatically generated through the above trigger model early warning process, and the mean values of the next three windows are continuously recorded, if they are continuously maintained in the interval of 94%-95%, the stable fluctuation band is marked;

[0245] Based on the trigger model early warning and the stable fluctuation band as the optimization basis, automatic collaborative iteration is carried out to reduce the model early warning triggering rate, and the normal fluctuation deviation value is updated according to the business scene change.

[0246] Please refer to Figure 2 The embodiment of the application further provides an intelligent matching system based on the intelligent matching method based on procurement demand and execution, the intelligent matching system comprising:

[0247] A data synchronization module 10 is configured to construct a dynamic data correlation graph and synchronize data in real time and accurately.

[0248] A task dynamic allocation module 20 is configured to construct a global task scheduling model according to the synchronized data, and dynamically allocate tasks based on load.

[0249] A demand dynamic analysis module 30 is configured to generate a demand sheet according to the synchronized data, and dynamically analyze and complete the semantics of the demand sheet.

[0250] A supplier dynamic matching module 40 is configured to continue dynamic matching of suppliers and priority sorting thereof according to the demand sheet dynamic analysis result.

[0251] A difference tracing early warning module 50 is configured to trace differences and give graded early warnings for abnormal supplier matching demand processes.

[0252] A dynamic index monitoring iteration module 60 is configured to monitor dynamic indexes in real time and automatically perform collaborative iteration.

[0253] Figure 3 The electronic device entity structure schematic diagram of the embodiment of the application is shown in Figure 3 The electronic device 70 comprises a processor 701, a memory 702, and an internal bus 703; wherein the processor 701 and the memory 702 communicate with each other through the internal bus 703;

[0254] The processor 701 is configured to invoke program instructions in the memory 702 to execute the method provided by the above method embodiments, for example, including: constructing a dynamic data correlation graph, and synchronizing data in real time; constructing a global task scheduling model according to the synchronized data, to dynamically allocate tasks based on load; generating a demand sheet according to the synchronized data, and dynamically analyzing and semantically completing the demand sheet; continuing to perform dynamic matching of suppliers and priority sorting of the suppliers according to the dynamic analysis result of the demand sheet; performing difference tracing and hierarchical early warning for abnormal processes of supplier matching demands; and performing real-time monitoring and automatic collaborative iteration for dynamic indicators.

[0255] The embodiment provides a non-transitory computer-readable storage medium storing computer instructions, and the computer instructions cause a computer to execute the method provided by the above method embodiments, for example, including: constructing a dynamic data correlation graph, and synchronizing data in real time; constructing a global task scheduling model according to the synchronized data, to dynamically allocate tasks based on load; generating a demand sheet according to the synchronized data, and dynamically analyzing and semantically completing the demand sheet; continuing to perform dynamic matching of suppliers and priority sorting of the suppliers according to the dynamic analysis result of the demand sheet; performing difference tracing and hierarchical early warning for abnormal processes of supplier matching demands; and performing real-time monitoring and automatic collaborative iteration for dynamic indicators.

[0256] Those skilled in the art can understand that all or part of the steps of the above method embodiments can be completed by program instruction related hardware, and the foregoing program can be stored in a computer readable storage medium, and the program is executed to perform steps including the above method embodiments; and the foregoing storage medium includes ROM, RAM, magnetic disc or optical disc and various storage media that can store program codes.

[0257] The device embodiments described above are only schematic, and units described as separate components can or can not be physically separate, and components shown as units can or can not be physical units, that is, can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0258] Those skilled in the art can clearly understand the implementation of the various embodiments by means of software and necessary general hardware platforms through the description of the above embodiments, and of course, the embodiments can also be implemented by hardware. Based on such understanding, the above technical solutions can be embodied in the form of a software product, and the computer software product can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, and the like, and includes a plurality of instructions to make a computer device (which can be a server, a network device, or the like) execute the method of the various embodiments or some parts of the embodiments.

[0259] Although the present application has been described in detail by the general description and specific embodiments above, some modifications or improvements can be made on the basis of the present application, which is obvious to those skilled in the art. Therefore, these modifications or improvements made on the basis of not deviating from the spirit of the present application are within the scope of the present application.

Claims

1. An intelligent matching method based on procurement needs and execution, characterized in that, Includes the following steps: Construct dynamic data association maps and synchronize data accurately in real time; A global task scheduling model is built based on synchronized data to dynamically allocate tasks based on load; A requirement order is generated based on the synchronized data, and the requirement order is dynamically parsed and semantically completed. Based on the dynamic parsing results of the demand order, continue to perform dynamic matching and priority ranking of suppliers; Implement differential source tracing and tiered early warning for anomalies in the supplier matching process; Real-time monitoring of dynamic indicators and automatic collaborative iteration; The construction of a dynamic data association map and the real-time accurate data synchronization specifically include: A graph database is constructed, and the real-time relationships between "demand order - purchase order - supplier - warehouse - region" are stored in the graph database to form a dynamic data relationship graph. The demand order is used as the starting core node, and relationship edges are established with the purchase order, supplier, warehouse and region respectively. The edge connection represents the relationship between entities. When a demand order involves a specific product category, the relevant supplier and its current availability information can be quickly located based on the dynamic data relationship graph through the relationship edges. Configure a distributed data gateway, which is associated with and connected to different independent subsystems, including the demand management system, procurement system, warehousing system, and supplier system. Collect multi-source heterogeneous data such as demand orders, purchase records, and warehouse data through a distributed data gateway; Furthermore, the CDC model is introduced. The CDC model is used to capture change data and capture the changes of multi-source heterogeneous data in the databases of each independent subsystem in real time. Specifically, the distributed data gateway monitors the binary log files of each independent subsystem database in real time based on the CDC model, and captures the data change events related to the log files, thereby triggering a real-time data synchronization mechanism to effectively reduce information lag and delay. Continue to build incremental data identification rules based on real-time synchronized data information to add a unique identifier and version number to each piece of data. The CDC model only directly synchronizes data items whose version numbers have been upgraded. When the current data is first created, the version number is an initial value. When the data changes, the version number is automatically synchronized and upgraded. The CDC model only compares the current version number with the last synchronized version number. If the data version number has not changed, it is not synchronized and transmitted. If the version number is upgraded, only the data items whose version number has been upgraded and transmitted are synchronized and transmitted. The process of generating a demand order based on synchronized data, and dynamically parsing and semantically completing the demand order, specifically includes: A BERT pre-trained model is constructed, and a demand semantic understanding model is trained based on the BERT pre-trained model and a corpus of historical purchase orders. The demand semantic understanding model is input with demand order text generated by the CDC model synchronously, and the demand semantic understanding model parses and outputs explicit and implicit information in the demand order text. Explicit information includes clearly structured fields such as demand category, quantity, and region, while implicit information includes non-explicit intent fields. Combining current time series, historical collaboration, and special regulations data, an intent inference submodule is further constructed to complete the implicit information, and the overall parsing result after completion is verified for compliance. The expression for further verifying the compliance of the parsing is as follows: S = α·P + β·H + γ·C In the formula: S is the score for parsing the requirement form text; P is the score for the completeness of explicit information, which is based on the required and optional items in the standard requirement form and compared with the actual extracted explicit information. With 100 points as the benchmark, 30 points are deducted when a required item is missing and 10 points are deducted when an optional item is missing. H represents the accuracy score for completing implicit information. This score is based on the current timeline, historical collaboration, and specially specified data, and is compared with the actual extracted implicit information. A manual verification process can be triggered afterward. The final score is based on 100 points. If there are missing or incorrect extracted items in the comparison results, 10 points will be deducted for each item. C represents the supplier's real-time capacity adaptability score. Combining historical collaboration data with the supplier's real-time capacity data, the automatic splitting ratio of the demand quantity is optimized and calculated. That is, when the demand quantity exceeds the capacity of a single supplier, other suppliers are automatically introduced to split the demand quantity proportionally. The splitting is based on the current available capacity ratio of different suppliers. The expression for calculating the supplier's real-time capacity adaptability score C is as follows: Supplier real-time capacity adaptability score C = Total demand coverage score × 25% + Single supplier capacity adaptability score × 30% + Delivery time coordination score × 25% + Risk diversification score × 20% Among them, the total demand coverage score means that the sum of the quantities undertaken by each supplier must equal the total demand, otherwise it is 0 points; the single supplier capacity matching score means that the quantity allocated to each supplier must be within its real-time adjustable capacity, otherwise it is 0 points; and the delivery time coordination score means that the delivery time of each supplier must fully meet the delivery time of the total demand, otherwise it is 0 points. Risk diversification score: Avoid over-reliance on a single supplier. Diversify the risk by reducing the overall risk caused by a sudden production failure of a single supplier. The score is based on the supplier with the largest allocation. If the largest supplier's allocation is ≤60%, it is 100 points; if it is 60%-80%, it is 50 points; if it is >80%, it is 0 points. α, β, and γ are the weighting coefficients of the corresponding terms, where α = 0.4, β = 0.3, and γ = 0.

3. When the text parsing score S of the demand order is greater than or equal to 80, it is considered a qualified demand parsing. Otherwise, the parsing will be readjusted for the explicit information completeness score P and / or implicit information completeness score H and / or supplier real-time capacity adaptability score C, which are not full marks.

2. The intelligent matching method based on procurement needs and execution as described in claim 1, characterized in that, The step of constructing a global task scheduling model based on synchronization data to dynamically allocate tasks based on load specifically includes: A global task scheduling model is constructed, which loads the CPU and memory usage status information of the cluster nodes in real time. The cluster nodes are servers, and a heartbeat monitoring mechanism with a frequency of once per second is established. Nodes with a load rate of <30% are designated as idle nodes, and each node represents a server. The global task scheduling model automatically assigns a task package to an idle node based on the main task queue generated from the synchronized data. The mapping relationship between this task package and the corresponding node is recorded in real time in the "task-node" subgraph of the dynamic data association graph. When the node's parsing times out or fails (with a timeout exceeding 30 seconds), the subtask is automatically returned to the global task scheduler to be re-entered into the main task queue. Simultaneously, the "status" attribute of the corresponding task package node in the dynamic data association graph is updated. This ensures that synchronized data can be reasonably distributed according to node load, avoiding data accumulation, reducing information lag, and guaranteeing real-time and efficient processing of subsequent needs.

3. The intelligent matching method based on procurement needs and execution as described in claim 2, characterized in that, The process of generating a demand order based on synchronized data, and dynamically parsing and semantically completing the demand order, specifically includes: A dynamic graph embedding model is constructed to observe in real time whether there are areas with significant demand fluctuations, and these areas are mapped to the vertices of the graph structure of the dynamic graph embedding model. The static attribute of three-dimensional coordinate distance and the dynamic attribute of real-time capacity load rate are then embedded based on the dynamic graph embedding model. The quantification results of "static distance + dynamic load rate" are then fused with the parsing results of the demand semantic understanding model at a weight of 50% each, and the overall allocation ratio is recalculated.

4. The intelligent matching method based on procurement needs and execution as described in claim 3, characterized in that, The step of dynamically matching and prioritizing suppliers based on the dynamic parsing results of the demand order specifically includes: To avoid the drawbacks of traditional methods that only sort and match procurement needs and suppliers by time, a dynamic supplier capability matrix and a demand-supplier matching scoring model are constructed in a distributed manner. The dynamic supplier capability matrix updates the capabilities of matched suppliers in real time, and the demand-supplier matching scoring model sorts the matching degree. Finally, intelligent algorithms are used to dynamically optimize and ensure that high-value and urgent needs are prioritized. Based on the supplier dynamic capability matrix, a dynamic profile package is constructed for each supplier. The contents of the dynamic profile package include: Real-time supply S1: The quantity of goods that the supplier can immediately allocate; Short-term supply S2: The quantity of goods that a supplier can quickly allocate within a specific time period; Capacity elasticity coefficient E: reflects the supplier's ability to cope with demand fluctuations. E = (actual maximum capacity / conventional standard capacity) - 1. The larger the value, the better the capacity elasticity coefficient E. Regional delivery cost D: The total cost for the supplier to deliver goods to the target area; Cooperation fit F: Based on historical cooperation data, F = (Number of on-time deliveries / Total number of cooperations) × 0.6 + (Number of quality-compliant deliveries / Total number of cooperations) × 0.4; Continuing with the demand-supplier matching scoring model, the following relationship is established: M = M1 × 0.4 + M2 × 0.6 In the formula, M is the comprehensive matching score. The higher the comprehensive matching score, the higher the matching degree between the supplier and the demand. Basic matching score M1: M1 = Category fit Q × 0.5 + Cooperation adaptability F × 0.5, where the category fit Q is 0 or 1 depending on the degree of matching between the supplier's category and the demand category; Dynamic matching score M2: M2 = Supply capacity score A × 0.3 + Cost-effectiveness score B × 0.4 + Timeliness satisfaction score T × 0.2 + Capacity elasticity score E' × 0.1; Supply capacity score A: A = (S1 / demand quantity) × 50 + (S2 / demand quantity) × 50. If the result exceeds 100, it is calculated as 100. Cost-benefit analysis B: B = 100 - (Regional delivery cost D / Maximum acceptable cost) × 100; Timeliness requirement T: T = (actual delivery time / required delivery time) × 100. If the actual delivery time is longer than the required delivery time, then T = 0. Capacity elasticity score E': E' = E × 100, converting the capacity elasticity coefficient into a score of 0-100, with any value exceeding 100 being counted as 100. Therefore, through multi-dimensional calculation, a comprehensive matching score between suppliers and demands is obtained. The higher the comprehensive matching score, the higher the matching priority between suppliers and demands, and thus their priority is ranked. To address real-time changes, a dynamic adjustment mechanism is further implemented as follows: Supply adjustment coefficient K1: When the change rate of the supplier's real-time supply S1 exceeds 20%, K1 = adjusted S1 / original S1, and at this time the dynamic matching M2 is updated to M2 × K1; Cost adjustment coefficient K2: When the change rate of regional delivery cost D exceeds 15%, K2 = D before adjustment / D after adjustment, and the dynamic matching M2 is updated to M2 × K2; Emergency Request Adjustment: For emergency requests, the comprehensive matching score M needs to be multiplied by the urgency coefficient U. U ranges from 1.2 to 2.0 depending on the urgency of the request. The higher the urgency, the larger U is. When a high-value order exceeds a preset threshold in the comprehensive matching score M calculation, the weight of the cooperation fit degree F is increased to 0.6, while the weight of the category fit degree Q is reduced to 0.4, and the capacity elasticity coefficient E is introduced for adjustment, i.e., M1 = (category fit degree Q × 0.4 + cooperation fit degree F × 0.6) × (1 + E × 0.2), in order to prioritize the stability of high-value orders and the ability to cope with fluctuations.

5. The intelligent matching method based on procurement needs and execution as described in claim 4, characterized in that, The aforementioned method for tracing discrepancies and providing tiered early warnings for anomalies in the supplier matching process specifically includes: A difference feature library is constructed, and difference factor data for each key node in the entire procurement process is collected and quantified in real time. These key nodes include demand nodes, order nodes, production nodes, logistics nodes, and warehousing nodes. The calculation formulas and meanings of the difference factors corresponding to each key node are as follows: ① Demand Node: Requirement change frequency: Fv = number of changes / original planned cycle, used to measure the frequency of requirement changes; Completeness of requirement description: Cd = number of complete items / total number of items × 100%, which reflects the level of detail in the requirement description. The difference value ΔCd is 100% - Cd. ② Order Node: Order response time: Tr = Supplier confirmation time - Order placement time, reflecting the supplier's response speed to the order; Price deviation rate: Rp = (actual unit price - benchmark unit price) / benchmark unit price × 100%, which measures the degree of deviation between the actual purchase price and the benchmark price; ③ Production nodes: Production scheduling achievement rate: Pc = actual production scheduling output / planned production scheduling output × 100%, which is used to evaluate the completion of the production scheduling plan. The difference is 100% - Pc. Material completeness rate: Cm = number of complete material items / total number of material items × 100%, which represents the completeness of the materials required for production. The difference value is 100% - Cm; ④ Logistics nodes: Transportation damage rate: Dt = number of damaged items / total transportation volume × 100%, which reflects the damage to goods during the logistics transportation process; Delay rate in transit: Dd = (actual transit time - standard transit time) / standard transit time × 100%, reflecting the timeliness of logistics transportation; ⑤ Inbound node: Quantity matching rate: Qi = actual inbound quantity / order quantity × 100%, used to judge the consistency between the actual inbound quantity and the order quantity, and the difference value is 100% - Qi; Quality pass rate: Qq = number of qualified products / number of samples × 100%, used to assess the quality compliance of goods entering the warehouse, and the difference is 100% - Qq; Specification matching rate: Qs = number of matching specifications / total number of specifications × 100%. This checks the degree of matching between the specifications of the goods entering the warehouse and the required specifications. The difference value ΔQs is 100% - Qs. Warehouse entry timeliness: Ti = actual entry time - planned entry time, which measures the timeliness of goods entering the warehouse; Construct a propagation matrix and calculate the cumulative impact index; Based on the difference factors extracted in the above steps, a propagation matrix Mij is further constructed, where Mij is the influence coefficient of the i-th node on the j-th node; Obtain the node sensitivity Sj: It is obtained by training on the degree to which the difference factors of each node are affected by the historical data of the difference factors of other nodes in the above steps; Calculate the propagation intensity Si at a single node: Si = Σ (Difference value of node i × Mij × Sj) Obtain the weight Wi of each node: Set the weight Wi based on the historical data influence range of the difference factors of each node in the above steps, and satisfy ΣWi=1; Further calculate the cumulative impact index Tc: Tc = Σ(Si × Wi) The cumulative impact index Tc is the core indicator for measuring the degree of impact of differences at each node on the entire procurement process. Its value reflects the severity of the impact of the difference, and it serves as a reference basis for tracing the source of the difference. Generate real-time dynamic early warning thresholds, calculate early warning levels, and push them out, specifically as follows: By directly linking the historical data of the node difference factors extracted in the above steps, the 90th percentile value of the same type of difference in the past 12 months is taken as the basic warning line T base. The 90th percentile value means that in a set of data, 90% of the data are less than or equal to the value, and only 10% of the data are greater than the value. Further determine the actual risk coefficient R of the current supply chain. The specific formula is as follows: R = 1 + (0.4×Dd + 0.3×ΔCm + 0.3×Rp) / 100 The three node difference factors directly adopted in the formula include: In-transit delay rate Dd: with a weight of 40%, used to reflect logistics risks; Material completeness rate difference ΔCm: with a weight of 30%, used to reflect production risks; Price deviation rate Rp: with a weight of 30%, used to reflect cost risks; Then the calculation formula for the final warning threshold T_final is as follows: T_final = T_base × R When any of the difference values in the above steps ≥ T_final, or the cumulative impact index Tc ≥ T_final, an alarm is immediately triggered, specifically including: When any of the difference values in the above steps ≥ 1.5×T_final, and Tc ≥ 1.5×T_final, or any of the difference values in the above steps ≥ 2×T_final, or Tc ≥ 2×T_final, a first-level alarm is triggered; When T_final ≤ any of the difference values in the above steps < 1.5×T_final, and T_final ≤ Tc < 1.5×T_final, or any of the difference values in the above steps ≥ 1.5×T_final, but Tc < T_final, or Tc ≥ 1.5×T_final, but any of the difference values in the above steps < T_final, a second-level alarm is triggered.

6. The intelligent matching method based on procurement requirements and execution according to claim 5, wherein The real-time monitoring of dynamic indicators and automatic collaborative iteration specifically include: Construct a three-dimensional dynamic indicator system of "efficiency - accuracy - timeliness"; Regularly and automatically collect the real-time data of the above indicators to form a dynamic trend graph, where the horizontal axis is time and the vertical axis is the indicator value. At the same time, mark the reference value line and the actual data value line, and compare the average value of the indicators in the recent 3 monitoring cycles with the reference value. Avoid false operations triggered by accidental fluctuations through multi-data point monitoring; Trigger the model warning process when it is lower than the threshold; Suppose there is a normal fluctuation deviation. If the window average value formed by the indicator values in the recent 3 monitoring cycles is within the normal fluctuation deviation range compared with the reference value, no warning is triggered to avoid accidental fluctuations; Automatically generate a fluctuation trajectory file through the above trigger model warning process, and continuously record the average values of the subsequent 3 windows. If it continues to maintain within the normal fluctuation deviation range, it is marked as a stable fluctuation band; Based on the above trigger model warning and stable fluctuation band as the optimization basis for automatic collaborative iteration to reduce the model warning trigger rate, and update the normal fluctuation deviation value according to the changes in the business scenario.

7. An intelligent matching system based on the intelligent matching method for procurement needs and execution as described in claim 6, characterized in that, The intelligent matching system includes: A data synchronization module, used to construct a dynamic data association map and synchronize data accurately in real time; A task dynamic allocation module, used to construct a global task scheduling model based on the synchronized data to dynamically allocate tasks based on the load; A demand dynamic analysis module, used to generate a demand list based on the synchronized data and perform dynamic analysis and semantic completion on the demand list; A supplier dynamic matching module, used to continue with supplier dynamic matching and its priority ranking based on the dynamic analysis results of the demand list; A difference traceability warning module, used to perform difference traceability and hierarchical warning for abnormalities in the process of supplier matching requirements; The dynamic indicator monitoring and iteration module is used for real-time monitoring of dynamic indicators and automatic collaborative iteration.

8. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1-6.

Images