An intelligent matching and recommendation method and system for automobile parts supply and demand based on a knowledge graph

Abstract

The application provides a kind of based on knowledge graph's automobile parts supply and demand intelligent matching recommendation method and system, belongs to artificial intelligence technical field, the application synchronously obtains the transfer volume upper limit and transfer time limit of logistics node, the historical order data of demand node and supply and demand node knowledge graph, the fluctuation coefficient of historical order is formed demand matrix by statistical analysis and demand saturation, and the transfer volume upper limit and transfer time limit are respectively as capacity constraint and line edge weight to build topological graph. According to demand matrix, determine the procurement quota of each demand node, expand iteration bidding with edge weight as conversion cost, flow arbitration is carried out through transfer volume threshold value and adjusts bidding score, and supply and demand flow direction graph is generated with transaction volume as weight after allocation balance. Extract target path from flow direction graph, calculate recommended value combining transaction volume and transfer time limit and output the final supply and demand matching scheme. The application realizes dynamic flow optimization and intelligent matching recommendation of transport path in automobile parts supply and demand network.

Description

Technical Field

[0001] This application belongs to the field of artificial intelligence technology, and in particular relates to a knowledge graph-based intelligent matching and recommendation method and system for automotive parts supply and demand. Background Technology

[0002] With the deepening of digital transformation in the automotive industry, knowledge graph-based intelligent matching and recommendation methods for automotive parts supply and demand have demonstrated significant application value in improving supply chain collaboration efficiency and optimizing logistics resource allocation. Knowledge graphs can uniformly manage the complex dynamic relationships between industrial entities, such as supplier-purchaser relationships and product-part relationships, providing technical support for semantic fusion and intelligent decision-making of multi-source heterogeneous data. By constructing a supply and demand node topology network and combining it with path optimization algorithms, resource competition and transportation timeliness constraints in large-scale parts distribution scenarios can be effectively addressed.

[0003] Existing technologies have incorporated knowledge graphs by integrating upstream and downstream resources of the supply chain, including equipment information, logistics and warehousing, product and component relationships, and supplier and buyer relationships, providing functions such as knowledge retrieval and visualization analysis, and supply decision analysis. Some technologies employ shortest path algorithms and centrality metrics for supply chain risk assessment and logistics planning, and mine upstream and downstream supply chain information and conduct risk monitoring by constructing enterprise relationship chain graphs. These methods provide a basic framework for modeling the supply and demand network of automotive parts.

[0004] However, existing supply chain management methods generally suffer from fragmented knowledge, lack a systematic characterization of demand fluctuations, and often rely on static demand representations and single-path cost assessments, resulting in insufficient dynamic flow allocation and load balancing capabilities under logistics node capacity constraints. Existing technologies struggle to effectively link procurement, logistics, and manufacturing processes when implementing supply chain risk management and component selection. They also lack a comprehensive recommendation and ranking mechanism for transaction volume and timeliness along transportation routes, making it difficult to achieve coordinated optimization of transportation timeliness and resource utilization in large-scale many-to-many supply and demand scenarios. Therefore, existing technologies suffer from insufficient intelligent matching of supply and demand. Summary of the Invention

[0005] The purpose of this application is to provide a knowledge graph-based intelligent matching and recommendation method and system for automotive parts supply and demand, in order to solve the problem of insufficient intelligent matching of supply and demand in the existing technology.

[0006] To address the aforementioned technical problems, in a first aspect, this application provides a knowledge graph-based intelligent matching and recommendation method for automotive parts supply and demand, comprising: Synchronously acquire the upper limit of transfer volume and transfer time of logistics nodes, historical order data of demand nodes in the target area, and a knowledge graph with supply nodes, logistics nodes, demand nodes as nodes and transportation lines between nodes as edges; The fluctuation coefficient and demand saturation of historical order data within a preset sampling window are calculated to obtain a demand matrix. A topology graph is constructed using the upper limit of transshipment volume as the transshipment volume threshold of logistics nodes in the knowledge graph and transshipment timeliness as the edge weight of the transportation route. The procurement quota of the demand node in the topology graph is determined based on the demand matrix, and the conversion cost of the transportation route is used as the edge weight. Iterative bidding is carried out between the demand node and the supply node. The bidding request is arbitrated using the transshipment volume threshold, and the bidding score of the transportation route is adjusted according to the ratio of the load of the logistics node to the transshipment volume threshold. When the bidding request and the supply node reach a distribution balance, the supply and demand flow graph is constructed with the transaction volume of the transportation route as the weight. The target path is determined based on the supply and demand flow diagram. The recommended value of the target path is calculated based on the transaction volume and transit time. All target paths are combined according to the recommended value to obtain the supply and demand matching scheme.

[0007] Optionally, all target paths are combined according to the recommended values ​​to obtain a supply-demand matching scheme, including: Each target path is sorted in descending order of recommendation value to obtain a set of candidate paths; The transaction volume of each target path in the candidate path set is calculated by summing up until it equals the procurement quota. The target paths that have been summed up in the candidate path set are then determined as the supply and demand matching schemes.

[0008] Optionally, the fluctuation coefficient and demand saturation of historical order data within a preset sampling window are calculated to obtain a demand matrix. A topology graph is then constructed using the upper limit of transshipment volume as the transshipment volume threshold for logistics nodes in the knowledge graph and transshipment timeliness as the edge weight of the transportation route, including: Extract the order volume of historical order data for each unit time period within a preset sampling window, calculate the mean and standard deviation of all order volumes, and obtain the fluctuation coefficient by calculating the ratio of the standard deviation to the mean. The demand saturation level is obtained by calculating the ratio of the mean to the order quantity with the largest value within the preset sampling window; Based on the volatility coefficient and demand saturation, a demand vector is constructed for each demand node, and all demand vectors are combined according to the order of each demand node in the knowledge graph to obtain a demand matrix. A topology graph is constructed by using the upper limit of the transfer volume of each logistics node as the threshold of the transfer volume of the corresponding logistics node in the knowledge graph, and using the transfer timeliness of each logistics node as the edge weight of the corresponding transportation route.

[0009] Optionally, the procurement quotas of demand nodes in the topology graph are determined based on the demand matrix, and the conversion cost of transportation routes is used as the edge weight. Iterative bidding is conducted between demand nodes and supply nodes. A transit volume threshold is used to arbitrate the bidding requests, and the bidding score of the transportation route is adjusted according to the ratio of the load of the logistics node to the transit volume threshold. When the bidding requests and supply nodes reach an allocation balance, a supply and demand flow graph is constructed using the transaction volume of the transportation route as the weight, including: The demand vector corresponding to each demand node is extracted from the demand matrix. The procurement quota of each demand node in the topology graph is obtained by weighting the fluctuation coefficient and demand saturation included in each demand vector. Using conversion cost as the initial value, bidding requests including procurement quotas are sent from each demand node to each supply node, and the flow value generated by each transportation line in response to the bidding request is obtained. The load of each logistics node is obtained by summing the flow values ​​of each transportation line passing through the same logistics node. The flow value is generated by allocating the procurement quota to each transportation line using the bidding score. The bidding score for the corresponding transportation route is adjusted based on the ratio of load capacity to transshipment volume threshold, and the transshipment volume threshold is used as the upper limit constraint for the corresponding load capacity. When the total flow value flowing into each demand node equals the corresponding procurement quota and the flow value on the corresponding transportation route no longer changes, the flow value is determined as the transaction volume, and the transaction volume is used as the weight of the transportation route in the knowledge graph to obtain the supply and demand flow diagram.

[0010] Optionally, the method further includes: The total bid value is obtained by calculating the sum of the bid scores corresponding to multiple transportation routes associated with each demand node. The allocation weight of each transportation route is determined by the ratio of the bidding score to the total bidding value for each route, and the product of the procurement quota and the allocation weight is calculated to obtain the flow value of each transportation route.

[0011] Optionally, the bidding score for the corresponding transportation route is adjusted based on the ratio of load capacity to transshipment volume threshold, and the transshipment volume threshold is used as the upper limit constraint for the corresponding load capacity, including: Calculate the ratio of the load capacity of each logistics node to the corresponding transfer volume threshold to obtain the load coefficient, and obtain the adjustment amount of each logistics node based on the product of the load coefficient and the preset step size parameter. The bidding scores for the corresponding transportation routes are negatively corrected using the adjustment amount; During the iterative bidding process, when the load of each logistics node equals the corresponding transshipment volume threshold, the allocation of flow value to the transportation route passing through the corresponding logistics node is stopped, so as to limit the load within the range of the transshipment volume threshold.

[0012] Optionally, the target route is determined based on the supply and demand flow map, and a recommended value for the target route is calculated based on transaction volume and transit time, including: The target path is obtained by extracting the connected sequence of transportation routes that satisfy the condition of transaction volume being greater than zero from the supply and demand flow diagram; The total travel time of the route is obtained by summing the transfer times of each transportation line in each target route, and the average transaction volume of each transportation line in each target route is calculated to obtain the average traffic flow of the route. The recommended value for each target path is obtained based on the ratio of average path traffic to total path time.

[0013] Secondly, this application provides a knowledge graph-based intelligent matching and recommendation system for automotive parts supply and demand, including: The acquisition module is used to synchronously acquire the upper limit of the transfer volume and the transfer time of logistics nodes, the historical order data of demand nodes in the target area, and the knowledge graph with supply nodes, logistics nodes, demand nodes as nodes and transportation lines between nodes as edges. The generation module is used to calculate the fluctuation coefficient and demand saturation of historical order data within a preset sampling window, obtain the demand matrix, and construct a topology graph with the upper limit of the transfer volume as the transfer volume threshold of the logistics node in the knowledge graph and the transfer timeliness as the edge weight of the transportation route. The module is used to determine the procurement quota of the demand nodes in the topology graph based on the demand matrix, and to use the edge weight as the conversion cost of the transportation route. Iterative bidding is carried out between the demand nodes and the supply nodes. The bidding requests are arbitrated using the transshipment volume threshold, and the bidding score of the transportation route is adjusted according to the ratio of the load of the logistics node to the transshipment volume threshold. When the bidding requests and the supply nodes reach an allocation balance, the supply and demand flow graph is constructed with the transaction volume of the transportation route as the weight. The calculation module is used to determine the target path based on the supply and demand flow diagram, calculate the recommended value of the target path based on the transaction volume and transit time, and combine all target paths according to the recommended value to obtain the supply and demand matching scheme.

[0014] Thirdly, this application provides an electronic device, comprising: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the knowledge graph-based intelligent matching and recommendation method for automotive parts supply and demand as described in the first aspect above.

[0015] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps of the knowledge graph-based intelligent matching and recommendation method for automotive parts supply and demand as described in the first aspect above.

[0016] The knowledge graph-based intelligent matching and recommendation method for automotive parts supply and demand provided in this application constructs a demand matrix by calculating the fluctuation coefficient and demand saturation of historical order data, thereby achieving a multi-dimensional quantitative representation of the dynamic characteristics of demand nodes and effectively solving the problem of insufficient systematic characterization of demand fluctuation characteristics in existing technologies. Subsequently, by introducing an iterative bidding mechanism and using a transit volume threshold for traffic arbitration, and dynamically adjusting the bidding score according to the ratio of the logistics node load to the transit volume threshold, adaptive traffic allocation and load balancing under logistics node capacity constraints are achieved, avoiding the problem of local node overload caused by static path planning.

[0017] Finally, by extracting target paths from the supply and demand flow graph and calculating recommendation values ​​based on transaction volume and transit time, a comprehensive evaluation and ranking mechanism for transportation routes was established, enabling data-driven collaborative decision-making in procurement, logistics, and supply. Therefore, this application effectively improves the matching accuracy and path recommendation efficiency in large-scale many-to-many supply and demand scenarios, achieves synergistic optimization of transportation timeliness and logistics resource utilization, and solves the technical problem of insufficient intelligent matching of supply and demand in existing technologies.

[0018] Furthermore, this application determines procurement quotas by extracting demand vectors from the demand matrix and weighting the fluctuation coefficient and demand saturation, directly transforming the dynamic characteristics of demand nodes into quantifiable procurement decision-making criteria, thus overcoming the problem of unreasonable quota setting caused by static demand representation in existing technologies. Based on this, by introducing an iterative bidding mechanism with conversion costs as the initial value, each demand node sends bidding requests to supply nodes according to its procurement quota, and uses bidding scores to allocate traffic across transportation routes, achieving dynamic game-theoretic matching between procurement demand and logistics resources, avoiding the rigidity of traffic allocation caused by single-path cost assessment.

[0019] Subsequently, the load is obtained by summing the traffic flow values ​​of transportation routes passing through the same logistics node, and the bidding score is dynamically adjusted according to the ratio of the load to the transfer volume threshold. At the same time, the transfer volume threshold is used as the upper limit constraint of the load, realizing traffic arbitration and adaptive load balancing under the logistics node capacity constraint, effectively solving the problem of insufficient dynamic traffic allocation and load balancing capabilities under the logistics node capacity constraint in the existing technology.

[0020] Finally, the allocation balance is determined by whether the total flow value flowing into the demand node equals the procurement quota and the flow value no longer changes. The flow value at this point is then used as the transaction volume to construct a supply and demand flow graph, enabling data-driven, interconnected decision-making in the procurement and logistics processes. Therefore, this application effectively improves the rationality of flow allocation and resource utilization efficiency in large-scale many-to-many supply and demand scenarios. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A flowchart illustrating a knowledge graph-based intelligent matching and recommendation method for automotive parts supply and demand, provided in an embodiment of this application; Figure 2 A flowchart illustrating a method for constructing a supply and demand flow graph, provided as an embodiment of this application; Figure 3 A flowchart illustrating a method for generating a supply and demand matching scheme provided in an embodiment of this application; Figure 4 A schematic diagram of the structure of a knowledge graph-based intelligent matching and recommendation system for automotive parts supply and demand, provided in an embodiment of this application; Figure 5 This is a schematic diagram of the hardware structure of an electronic device provided in one embodiment of this application. Detailed Implementation

[0023] Existing knowledge graph-based supply and demand matching methods face significant shortcomings in dynamic adaptability: on the one hand, the use of static demand representation cannot capture the fluctuation patterns and saturation characteristics of historical orders, resulting in procurement quota settings deviating from the actual demand scenario; on the other hand, the use of a single time-efficiency cost to evaluate the value of the path ignores the phenomenon of traffic competition under the capacity constraints of logistics nodes, causing some logistics nodes to be overloaded while other nodes are idle during high-demand periods, resulting in an imbalance in resource utilization.

[0024] This contradiction stems from the lack of a multi-dimensional characterization mechanism for demand fluctuations, the dynamic arbitration capability of logistics node load, and a comprehensive evaluation system for route transaction volume and timeliness in the supply and demand matching process. There is an urgent need for a supply and demand matching method that can realize demand perception, adaptive traffic allocation, and intelligent route recommendation.

[0025] To address the aforementioned issues, this application proposes a knowledge graph-based intelligent matching and recommendation method for automotive parts supply and demand. Its core lies in a three-layer collaborative mechanism—demand matrix construction, iterative bidding flow allocation, and supply-demand flow graph recommendation—to achieve dynamic balance and intelligent decision-making in the supply-demand network. Specifically, a demand matrix is ​​first constructed by calculating the fluctuation coefficient and demand saturation of historical orders within a preset sampling window, quantifying the dynamic characteristics of demand nodes into calculable procurement quotas.

[0026] Subsequently, using transit time as the conversion cost, an iterative bidding mechanism is introduced between demand and supply nodes. Traffic is arbitrated using transit volume thresholds, and the bidding score for each transportation route is dynamically adjusted based on the ratio of logistics node load to the transit volume threshold. After the allocation is balanced, a supply and demand flow diagram is generated using transaction volume as the weight. Finally, target paths are extracted from the flow diagram, and recommended values ​​are calculated by combining transaction volume and transit time, and matching solutions are output according to priority.

[0027] This method abandons the traditional static demand allocation model. Through the deep integration of demand fluctuation perception, load adaptive arbitration and multi-dimensional path evaluation, it ensures that the procurement quota accurately reflects the demand characteristics, the logistics resources are dynamically and evenly allocated, and the recommended path comprehensively weighs timeliness and flow. It solves the problem of insufficient intelligent matching of supply and demand caused by insufficient demand characterization, rigid flow allocation and single path evaluation in the existing technology, and significantly improves the matching accuracy and resource utilization efficiency in large-scale many-to-many supply and demand scenarios.

[0028] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0029] To address the problems of existing technologies, embodiments of this application provide a knowledge graph-based intelligent matching and recommendation method, apparatus, device, computer storage medium, and computer program product for automotive parts supply and demand. The knowledge graph-based intelligent matching and recommendation method for automotive parts supply and demand provided in this application embodiment will be described below first.

[0030] Figure 1 This illustration shows a flowchart of a knowledge graph-based intelligent matching and recommendation method for automotive parts supply and demand, provided in one embodiment of this application. Figure 1 As shown, the method includes: S101. Synchronously obtain the upper limit of the transfer volume and the transfer time of the logistics node, the historical order data of the demand node in the target area, and the knowledge graph with supply nodes, logistics nodes, demand nodes as nodes and transportation lines between nodes as edges.

[0031] A logistics node refers to a logistics distribution center or transit warehouse that undertakes functions such as parts transfer, sorting, and warehousing. The maximum transfer volume refers to the maximum number of parts that a logistics node can handle per unit of time. Transfer timeliness refers to the time cost required for parts to travel from one node to another via a transportation route.

[0032] Demand nodes refer to the purchasers or end-users of automotive parts. Historical order data refers to the records of parts purchase orders generated by demand nodes over a past period, including information such as order time, order quantity, and part type. Supply nodes refer to the suppliers or manufacturers of automotive parts.

[0033] A knowledge graph is a supply and demand network topology represented in the form of a directed graph structure. Nodes represent supply nodes, logistics nodes, and demand nodes. Directed edges represent transportation routes between nodes and their connections. The direction of the edges represents the flow direction of components, i.e., from supply nodes to logistics nodes and from logistics nodes to demand nodes.

[0034] Knowledge graphs are based on directed triples Head node, transport relationship, tail node The connection information between nodes is stored in a format that includes transportation route identifiers and their attributes such as transit time. The target area refers to a specific geographical area or administrative region where supply and demand matching is required. A transportation route refers to a logistics channel or delivery route connecting two nodes.

[0035] Specifically, firstly, the daily processing capacity configuration parameters of each logistics node are queried through the data interface of the logistics management platform to obtain the upper limit of the transfer volume of each logistics node, and the transfer timeliness data is obtained by statistically analyzing the average time consumption of each transportation route in historical transportation records. Next, historical order data of each demand node in the target area is extracted from the order management platform. This data covers the order timestamps and order quantities within a preset time window.

[0036] Simultaneously, a pre-built knowledge graph is obtained from the supply chain knowledge base. The knowledge graph construction process is as follows: using supplier files, logistics distribution center files, and purchaser files from the supply chain management system as data sources, supply node sets are extracted respectively. Logistics node set and demand node set Based on historical delivery records in the logistics operation platform, the connections between nodes that actually involve logistics transactions are extracted, and directed edges are constructed in the directions of supply node → logistics node and logistics node → demand node, forming directed triples. Transportation relations and Transportation relations The above-mentioned node and edge relationships are stored in a graph database such as Neo4j to form a knowledge graph with triples as the basic unit.

[0037] For example, suppose a regional automotive parts supply network includes a set of supply nodes. Logistics node set and demand node set ,in , , These represent the number of supply nodes, logistics nodes, and demand nodes, respectively.

[0038] Obtain logistics nodes through data collection The maximum transshipment volume is From the supply node To logistics nodes The transit time for the transportation route is From logistics nodes To demand node The transit time for the transportation route is Demand Node Historical order data within the preset sampling window is ,in Indicates the number of time periods within the sampling window. Indicates the first Order volume for a given period Knowledge graphs use directed triples. , Indicates the node connection relationship, where Indicates from the supply node Pointing to logistics nodes Directed transport routes, Indicates from logistics nodes Pointing to the demand node directional transport routes.

[0039] S102. Calculate the fluctuation coefficient and demand saturation of historical order data within the preset sampling window to obtain the demand matrix. Then, construct a topology graph with the upper limit of the transfer volume as the transfer volume threshold of the logistics node in the knowledge graph and the transfer timeliness as the edge weight of the transportation route.

[0040] Optionally, step S102 calculates the fluctuation coefficient and demand saturation of historical order data within a preset sampling window to obtain a demand matrix. The process of constructing a topology graph, using the upper limit of transshipment volume as the transshipment volume threshold for logistics nodes in the knowledge graph and transshipment timeliness as the edge weight of the transportation route, can specifically include: S1021. Extract the order volume of historical order data for each unit time period within the preset sampling window, calculate the mean and standard deviation of all order volumes, and obtain the fluctuation coefficient by calculating the ratio of the standard deviation to the mean.

[0041] A preset sampling window refers to the time range used to statistically analyze historical order data, including the window start time, window end time, and the method of dividing the unit time period within the window. The size of the preset sampling window is determined based on business needs and the sufficiency of historical data: for components with relatively stable demand fluctuations, a 90-day rolling window is recommended to improve statistical stability; for components with significant seasonal demand fluctuations, a 365-day window is recommended to cover the complete seasonal cycle.

[0042] A unit of time period refers to the smallest time unit for data sampling within a preset sampling window, and can be a time granularity such as hour, day, or week. The volatility coefficient is a measure of the dispersion of historical order data within the preset sampling window, and is used to quantify the volatility characteristics of order volume at demand nodes.

[0043] Specifically, the historical order data first obtained from step S101 Extract the order volume for each time period within the preset sampling window, where... This represents the total number of time periods within the sampling window. Then, the sum of the order volumes for all time periods is divided by the total number of time periods to obtain the mean. .

[0044] Then, the standard deviation is obtained by calculating the square root of the sum of squares of the deviations of order volume from the mean for each time period. Since the historical order data is a finite sample, the sample standard deviation calculation method is used. ,when When there is only one time period within the sampling window, it is impossible to calculate the sample standard deviation. In this case, let Finally, the fluctuation coefficient is obtained by calculating the ratio of the standard deviation to the mean. .

[0045] S1022. The demand saturation is obtained by calculating the ratio of the mean to the order quantity with the largest value within the preset sampling window.

[0046] Demand saturation refers to the ratio of the average demand level to the peak demand level of historical order data within a preset sampling window, and is used to quantify the degree to which the order volume at a demand node approaches the maximum carrying capacity.

[0047] Specifically, firstly, the total number of orders within the preset sampling window... Find the order quantity with the largest value. Next, the mean obtained in step S1021 is used. The demand saturation level is obtained by calculating the ratio of the mean to the maximum order quantity. .

[0048] S1023. Based on the fluctuation coefficient and demand saturation, construct the demand vector for each demand node, and combine all demand vectors according to the order of each demand node in the knowledge graph to obtain the demand matrix.

[0049] A demand vector is a two-dimensional feature vector formed by combining the volatility coefficient and demand saturation of each demand node, used to represent the demand characteristics of that node. A demand matrix is ​​a matrix structure formed by vertically combining the demand vectors of all demand nodes according to their order in the knowledge graph, used to uniformly manage the demand characteristics of all demand nodes within a target area.

[0050] Specifically, firstly, based on the fluctuation coefficient calculated in step S1021... The demand saturation calculated in step S1022 Build demand nodes demand vector ,in This represents the vector transpose. Next, it follows the order of the requirement nodes in the knowledge graph. Stack the demand vectors of all demand nodes row by row to obtain Demand Matrix :

[0051] The demand matrix The Row corresponding to demand node The demand vector has the first column as the fluctuation coefficient of all demand nodes and the second column as the demand saturation of all demand nodes.

[0052] S1024. Construct a topology graph by using the upper limit of the transfer volume of each logistics node as the threshold of the transfer volume of the corresponding logistics node in the knowledge graph, and using the transfer timeliness of each logistics node as the edge weight of the corresponding transportation route.

[0053] The transshipment volume threshold refers to the upper limit constraint on the flow capacity set for each logistics node in the topology graph, used to limit the maximum flow of parts that the logistics node can handle during the supply and demand matching process. Edge weight refers to the numerical attribute assigned to the transportation routes in the topology graph, used to quantify the time-delivery cost or path cost of the transportation routes. The topology graph is a weighted constrained network graph formed by adding the transshipment volume threshold and edge weight attributes to a knowledge graph.

[0054] Specifically, each logistics node is first obtained from step S101. upper limit of transshipment volume This is used as the transfer volume threshold for the corresponding logistics node in the topology graph. Next, the transit time for each transportation route is obtained from step S101, including the transit time from the supply node. To logistics nodes Transit time And from logistics nodes To demand node Transit time These are used as edge weights for the corresponding transportation routes in the topology graph. and .

[0055] Finally, based on the node set and edge connections of the knowledge graph, combined with the transfer volume threshold and edge weights, a directed topological graph is constructed. ,in Represents a set of nodes, including the set of supply nodes. Logistics node set and demand node set ; This represents a set of directed edges, including all transportation routes, with directions from supply nodes to logistics nodes and from logistics nodes to demand nodes; each logistics node... Includes transit volume threshold attribute Each directed edge Attached edge weight attribute or .

[0056] This embodiment achieves deep feature extraction and network structure enhancement of historical order data of demand nodes. It not only overcomes the problem of insufficient feature characterization caused by static demand representation in the prior art, but also realizes unified management and structured expression of the features of all demand nodes in the target area, which significantly improves the accuracy of supply and demand matching and the rationality of resource allocation.

[0057] S103. Determine the procurement quota of the demand nodes in the topology graph based on the demand matrix, and use the edge weight as the conversion cost of the transportation route. Iterative bidding is carried out between the demand nodes and the supply nodes. The bidding requests are arbitrated using the transshipment volume threshold, and the bidding score of the transportation route is adjusted according to the ratio of the load of the logistics node to the transshipment volume threshold. When the bidding requests and the supply nodes reach a distribution balance, the supply and demand flow graph is constructed with the transaction volume of the transportation route as the weight.

[0058] Optionally, step S103 determines the procurement quota of the demand nodes in the topology graph based on the demand matrix, and uses the edge weight as the conversion cost of the transportation route. Iterative bidding is conducted between the demand nodes and the supply nodes. The bidding requests are arbitrated using the transshipment volume threshold, and the bidding score of the transportation route is adjusted according to the ratio of the load of the logistics node to the transshipment volume threshold. When the bidding requests and the supply nodes reach an allocation balance, the process of constructing a supply and demand flow graph with the transaction volume of the transportation route as the weight can specifically include: Figure 2 A flowchart illustrating a method for constructing a supply and demand flow graph according to an embodiment of this application is shown. Figure 2 As shown, its core lies in conducting multi-round iterative bidding based on the supply and demand relationship and edge weights of nodes within the topology graph. First, the unique demand vector of each demand node is extracted from the demand matrix. By weighting the fluctuation coefficient and demand saturation included in this vector, the procurement quota of each demand node in the topology graph is calculated.

[0059] Subsequently, using the transportation route conversion cost represented by the edge weight as the initial benchmark, each demand node initiates a bidding request carrying a procurement quota to the corresponding supply node. During this interaction, the procurement quota is allocated at the road network level according to the predetermined bidding score, thereby generating the initial flow value on each transportation route. To achieve reasonable arbitration of global flow, the flow values ​​of each route flowing through the same logistics node are summed to obtain the actual load of that logistics node. Based on the ratio of this load to a preset transshipment volume threshold, the bidding score of the corresponding transportation route is dynamically adjusted, and the transshipment volume threshold is strictly used as the physical upper limit constraint of the node's load.

[0060] The aforementioned score-based flow allocation and arbitration mechanism will continue to iterate until the total flow into each demand node is precisely equal to its previously proposed procurement quota and the flow allocation on each transportation route reaches an absolute steady state and no longer changes. Once this balance condition is met, it is determined that the bidding side and the supply side have reached an allocation balance. At this point, the solidified flow value is officially confirmed as the transaction volume of the corresponding route, and this transaction volume is used as the actual weight of the transportation route in the knowledge graph, ultimately constructing a supply and demand flow graph that reflects the true supply and demand mapping relationship. This method includes: S1031. Extract the demand vector corresponding to each demand node from the demand matrix. By weighting the fluctuation coefficient and demand saturation included in each demand vector, obtain the procurement quota of each demand node in the topology graph.

[0061] The procurement quota refers to the quantity index of parts to be procured, calculated based on the fluctuation coefficient and demand saturation of the demand node, and is used to determine the total amount of parts that the demand node needs to acquire in the supply and demand matching process.

[0062] Specifically, firstly, the demand matrix constructed in step S102... Extract each requirement node Corresponding demand vector ,in Indicates the volatility coefficient. This represents the demand saturation level. Next, a weighted sum of the fluctuation coefficient and the demand saturation level in the demand vector is calculated to obtain the demand node. Procurement quota ,in and The weighted coefficients of volatility coefficient and demand saturation, respectively, satisfy the following conditions: Finally, the procurement quota is calculated for each demand node to obtain a procurement quota set. .

[0063] Among them, the weighting coefficient and The method for determining the weighting is as follows: Based on historical supply and demand matching data, the weights of the impact of demand fluctuations on procurement failures and the impact of demand saturation on resource waste are statistically analyzed. The optimal weighting coefficients are then obtained by minimizing the procurement cost objective function. Specifically, a grid search method is used to satisfy... Under the constraints, traverse Within the interval, candidate values ​​are selected in steps of 0.1. Historical data on procurement quotas are calculated for each candidate coefficient group, and the matching effect is evaluated. The coefficient combination that minimizes the total cost is selected as the final value. and The value of is recommended. In standard scenarios, when demand fluctuates significantly, the following is suggested: , Recommended when demand is relatively stable. , .

[0064] S1032. Using the conversion cost as the initial value, send bidding requests, including procurement quotas, to each supply node through each demand node, and obtain the flow value generated by each transportation line when responding to the bidding request. By summing the flow values ​​of each transportation line passing through the same logistics node, the load of each logistics node is obtained. The flow value is generated by allocating the procurement quota to each transportation line using the bidding score.

[0065] Switching cost refers to the initial path cost when a demand node sends a bid request to a supply node via a transportation route, and it is equal to the edge weight of that transportation route. A bid request is a resource allocation application sent by a demand node to a supply node, which includes information on the procurement quota and switching cost.

[0066] Traffic volume refers to the quantity of parts traffic allocated to each transport route during the iterative bidding process. Load capacity refers to the sum of the traffic volumes of all transport routes passing through a logistics node in the current iterative bidding round. Bidding score refers to the competitiveness score of a transport route during the bidding process and is used to determine the proportion of traffic allocation among multiple transport routes.

[0067] Specifically, firstly, the topology graph constructed in step S102... Extract the edge weights of each transportation route. This is used as the initial value of the conversion cost. ,in The weights of the transit time edges for each transportation route in the topology graph are uniformly represented, and will be used in subsequent embodiments. To unify the symbol representation of the edge weights of the corresponding transportation routes, the symbol differences between supply nodes to logistics nodes and logistics nodes to demand nodes are no longer distinguished separately.

[0068] Next, the switching costs are normalized, making... ,in This represents the maximum cost of switching all transportation routes. If the transit time data is abnormal, the transit route should be marked as invalid and excluded from the bidding process. For valid transit routes, the reciprocal of the normalized conversion cost should be used as the initial bidding score. .

[0069] Then, each demand node Send procurement quotas to the supply nodes connected to it The bidding request. In responding to the bidding request, the procurement quota is allocated based on the bidding score, generating the flow value for each transportation route. Finally, for each logistics node... The load capacity of a logistics node is obtained by summing the flow rates of all transport routes passing through it. The summation range is all nodes that pass through the logistics nodes. The transportation routes.

[0070] S1033. Adjust the bidding score of the corresponding transportation route according to the ratio of load to transfer volume threshold, and use the transfer volume threshold as the upper limit constraint of the corresponding load.

[0071] Specifically, first calculate each logistics node. load and the corresponding transfer volume threshold The ratio of the two values ​​yields the load factor. Next, based on the load factor and the preset step size parameters... The product of these two factors yields the adjustment amount. Then, the adjustment amount is used to adjust the logistics nodes. The bidding scores for the transportation routes will be adjusted accordingly. Finally, the transfer volume threshold will be set. As a logistics node load The upper limit constraint is When the load reaches the transfer volume threshold, the allocation of traffic to the transportation lines passing through the logistics node will stop.

[0072] S1034. When the total amount of flow value flowing into each demand node is equal to the corresponding procurement quota and the flow value on the corresponding transportation route no longer changes, the flow value is determined as the transaction volume, and the transaction volume is used as the weight of the transportation route in the knowledge graph to obtain the supply and demand flow diagram.

[0073] Allocation equilibrium refers to the convergence condition during the iterative bidding process where the total traffic acquired by each demand node satisfies its procurement quota and the traffic allocation of each transportation route reaches a stable state and no longer changes. Transaction volume refers to the final determined quantity of component traffic on each transportation route when allocation equilibrium is achieved. The supply and demand flow graph is a weighted directed graph reflecting the supply and demand allocation results, formed by adding a transaction volume attribute to each transportation route on the basis of the original directed topology of the knowledge graph. The original attributes such as transit time remain unchanged, and the transaction volume is stored in parallel on the corresponding transportation route as a newly added weight attribute.

[0074] Specifically, firstly, during each round of iterative bidding, the flow of demand to each node is... The total traffic volume obtained by the demand node is obtained by summing the traffic volume values ​​of all transportation routes. Next, determine whether the total traffic volume equals the procurement quota. ,Right now Simultaneously record the flow rate of each transportation route in the current round. Compared with the previous round of traffic value Calculate the change in flow rate When all demand nodes are satisfied And all transportation routes meet the requirements. At that time, among them The preset traffic convergence threshold is used to determine whether a distribution balance has been achieved.

[0075] Flow convergence threshold The determination method is as follows: based on the average procurement quota of the demand nodes. Calculate the relative convergence threshold and set... ,in This represents the relative convergence coefficient. By analyzing the distribution characteristics of traffic changes in historical iteration data, it was found that a relatively small relative convergence coefficient leads to excessive iterations, affecting real-time performance, while a relatively large coefficient results in decreased matching accuracy. Different... The optimal relative convergence coefficient is determined by the number of iterations corresponding to the value and the matching error. ,Right now For example, when the average purchase quota is 1000 units, the flow convergence threshold is set to... Item.

[0076] Then, the flow rate values ​​of each transportation route at this time are... Determined as trading volume Finally, based on the node set and directed edge connections of the knowledge graph, transaction volume is added as an additional weight attribute to each transportation route, building upon the original weights of transshipment timeliness edges, to construct a supply and demand flow graph. ,in Represents a set of nodes. Represents the set of directed edges, where each directed edge... It carries two weighted attributes: transit time attribute. and transaction volume attributes .

[0077] This embodiment implements dynamic traffic optimization and adaptive load balancing based on an iterative bidding mechanism. It not only accurately translates demand characteristics into procurement decision-making criteria, effectively avoiding logistics node overload and resource waste, but also achieves synergistic optimization of procurement demand, logistics capacity, and supply capacity. This overcomes the shortcomings of static traffic allocation and load balancing capabilities in existing technologies, significantly improving the dynamic adaptability of supply and demand matching and resource utilization efficiency.

[0078] Optionally, the method further includes: The total bid value is obtained by summing the bid scores of multiple transport routes associated with each demand node. The allocation weight of each transport route is determined based on the ratio of its bid score to the total bid value. The product of the procurement quota and the allocation weight is then calculated to obtain the flow rate of each transport route.

[0079] The bid sum refers to the total bid score of all transport routes associated with a certain demand node, and is used to normalize the traffic allocation weight. The allocation weight refers to the proportion of traffic allocated to a certain transport route among all competing routes, and is determined by the ratio of the bid score to the bid sum.

[0080] Specifically, firstly, regarding the demand nodes For multiple related transportation routes, the sum of the bidding scores for each route is calculated to obtain the total bidding value. The summation range includes all data arriving at the demand node from the supply node via logistics nodes. The transportation routes. Then, based on the bidding points for each transportation route... With bidding and value The ratio determines the allocation weights. Then, calculate the procurement quota. With weight allocation The product of these two values ​​yields the flow rate of the transport route. .

[0081] This embodiment implements a traffic normalization allocation mechanism based on bidding scores. This not only ensures a fair competition mechanism where paths with high bidding scores receive more traffic, but also guarantees the reasonable distribution of procurement quotas among multiple transportation routes for demand nodes. It effectively solves the problem of unclear traffic allocation weights in multi-path competition scenarios, improving the transparency and controllability of traffic allocation.

[0082] Optionally, step S1033, which adjusts the bidding score of the corresponding transportation route based on the ratio of load capacity to transshipment volume threshold and uses the transshipment volume threshold as the upper limit constraint of the corresponding load capacity, may specifically include: S10331. Calculate the ratio of the load of each logistics node to the corresponding transfer volume threshold to obtain the load coefficient, and obtain the adjustment amount of each logistics node based on the product of the load coefficient and the preset step size parameter.

[0083] The load factor refers to the proportion of the current load of a logistics node to its transit volume threshold, and is used to quantify the load saturation level of the logistics node. The preset step size parameter is an adjustment factor used to control the magnitude of bid score correction during iterative bidding, and can be configured according to the supply and demand network size and traffic fluctuation characteristics. The adjustment amount refers to the bid score correction value calculated based on the load factor and the preset step size parameter, and is used to adjust the competitiveness of transportation routes passing through the logistics node. The preset step size parameters are shown in Table 1 below: Table 1: Preset Step Size Parameter Comparison Table

[0084] As shown in Table 1, the recommended values ​​of the preset step size parameter are displayed for different supply and demand network sizes. The larger the network size, the smaller the corresponding step size parameter should be to ensure the stability of iterative convergence.

[0085] Preset step size parameters The determination method is as follows: Simulation experiments are used to statistically analyze the convergence speed and load balancing effect of iterative bidding under different network sizes. When the step size parameter is too large, it leads to oscillations in the bidding score and non-convergence; when the step size parameter is too small, it leads to excessively slow iterative convergence. A binary search method is then used... Find the optimal step size within the interval that minimizes both the average number of iterations and the load variance.

[0086] For small-scale networks, due to the small number of nodes and strong system fault tolerance, a larger step size of 0.15 can be used to accelerate convergence. For large-scale networks, due to the high complexity of traffic allocation, a smaller step size of 0.03 is needed to ensure stability. When the total number of nodes is at the scale boundary, linear interpolation is used to determine the step size parameter. For example, when the total number of nodes is 50... .

[0087] Specifically, the load of the logistics nodes is first calculated from step S1032. The transfer volume threshold constructed in step S102 In the process, each logistics node is calculated. The load factor is obtained by dividing the load amount by the corresponding transfer volume threshold. Next, based on the load factor... With preset step size parameters The product of these values ​​yields the adjustment amount for each logistics node. .

[0088] S10332. Use the adjustment amount to negatively correct the bidding score of the corresponding transportation route.

[0089] Negative correction refers to subtracting the bidding score of a transportation route by using the adjustment amount as a negative increment, and is used to reduce the competitiveness of transportation routes corresponding to high-load logistics nodes.

[0090] Specifically, the first step is to identify the logistics nodes. All transportation routes, including those from supply nodes to logistics nodes and those from logistics nodes to demand nodes. Next, for each route passing through a logistics node... The transportation route, using the adjustment amount calculated in step S10331. Current bidding points for this transport route Perform negative correction, that is Finally, the revised bidding points will be... As part of the bidding score for this transportation route in the next round of iterative bidding, traffic will be diverted to logistics nodes with lower loads.

[0091] S10333. During the iterative bidding process, when the load of each logistics node is equal to the corresponding transshipment volume threshold, the allocation of flow value to the transportation line passing through the corresponding logistics node is stopped, so as to limit the load within the range of the transshipment volume threshold.

[0092] Specifically, during each round of iterative bidding, each logistics node is monitored in real time. load With transfer volume threshold The relationship. Then, when the load of a certain logistics node is detected to equal the corresponding transshipment volume threshold, i.e. When this occurs, the capacity constraint mechanism is triggered, stopping all shipments passing through that logistics node. New flow values ​​are assigned to the transport routes. Then, the transshipment volume threshold is set. As a logistics node load The hard upper limit constraint is Finally, for logistics nodes that have reached their capacity limits, their associated transportation routes will no longer accept traffic allocation in subsequent iterations, and traffic will be automatically redirected to other logistics nodes and their associated paths that are not yet fully loaded.

[0093] This embodiment implements an adaptive bidding score adjustment mechanism based on load feedback and strict capacity constraint management. It not only enables precise linkage between the bidding score adjustment and the load status of logistics nodes, but also effectively avoids local node overload and ensures that the load of all logistics nodes is strictly controlled within capacity limits. This overcomes the lack of dynamic load control capabilities for logistics nodes in existing technologies, significantly improving the load balance and traffic distribution stability of the supply and demand network.

[0094] S104. Determine the target path based on the supply and demand flow diagram, calculate the recommended value of the target path based on the transaction volume and transit time, and combine all target paths according to the recommended value to obtain a supply and demand matching scheme.

[0095] Figure 3 A flowchart illustrating a method for generating supply and demand matching schemes according to an embodiment of this application is shown. Figure 3 As shown, the method includes: Optionally, step S104, which involves determining the target path based on the supply and demand flow diagram and calculating the recommended value of the target path based on transaction volume and transit time, may specifically include: S1041. Extract the connected sequence of transportation routes that satisfy the condition of transaction volume being greater than zero from the supply and demand flow diagram to obtain the target path.

[0096] A connected sequence refers to an ordered combination of transportation routes from a supply node, through logistics nodes, to a demand node, where adjacent transportation routes in the sequence are directly connected in the topology graph. A target path refers to a connected sequence in the supply and demand flow graph where the transaction volume is greater than zero, representing an effective supply and demand path that actually carries the flow of components.

[0097] Specifically, firstly, starting with the supply and demand flow diagram constructed in step S103... The system identifies all transportation routes with a transaction volume greater than zero, i.e., those that meet the following criteria. The set of transportation routes. Next, a graph traversal algorithm is used to perform path search on the supply and demand flow graph, starting from each supply node. Starting from the beginning, use either depth-first search or breadth-first search to find the logistics nodes. Finally reaching the demand node All connected sequences.

[0098] Then, for each connected sequence, verify whether adjacent transportation routes in the sequence are directly connected in the topology graph and whether the transaction volume is greater than zero. Connected sequences that meet the conditions are determined as the target path. ,in This represents the path number. Finally, all target paths that meet the criteria are summarized to obtain the target path set. ,in This represents the total number of target paths.

[0099] S1042. By summing the transfer times corresponding to each transportation line in each target path, the total path time is obtained, and the average transaction volume corresponding to each transportation line in each target path is calculated to obtain the average path flow.

[0100] Total route time refers to the sum of the transit times of all transportation routes in the target route, and is used to quantify the total time cost required to transport components from the supply node to the demand node along the route. Average route flow refers to the arithmetic mean of the transaction volume of all transportation routes in the target route, and is used to represent the average transportation capacity of the route.

[0101] Specifically, firstly for each target path Extract all transportation routes included in the route and their corresponding transit times. Then, analyze the route... The total route time is obtained by summing the transit times of each transportation route. The summation range is the target path. All transportation routes included. This refers to the transit time of the transportation route.

[0102] Then, extract the path. Transaction volume of various transportation routes in China Calculate the arithmetic mean of all transaction volumes to obtain the average path flow. ,in Indicates the target path The total number of transportation routes included, with the summation range being the number of paths. This includes all transportation routes. Finally, the total path time and average path traffic are calculated for each target path to obtain the set of total path times. and path average flow set .

[0103] S1043. Based on the ratio of average path traffic to total path time, obtain the recommended value for each target path.

[0104] The recommended value is a path priority score calculated by the ratio of average path traffic to total path travel time, used to comprehensively evaluate the transportation efficiency and traffic capacity of a target path. Specifically, firstly, for each target path... Extract the path average flow calculated in step S1042. Total path time Next, the ratio of average path traffic to total path time is calculated to obtain the recommended value for the target path. Finally, a recommendation value is calculated for each target path, resulting in a set of recommendation values. .

[0105] This embodiment implements a target path priority evaluation mechanism based on the fusion of transaction volume and transit time. It not only avoids interference from invalid paths but also achieves dual-dimensional feature extraction of timeliness and traffic volume. It overcomes the problem in existing technologies of lacking a comprehensive recommendation and ranking mechanism for transaction volume and timeliness in transportation paths, significantly improving the decision-making quality and implementation efficiency of supply and demand matching solutions.

[0106] Optionally, the process of combining all target paths according to the recommended values ​​to obtain a supply-demand matching scheme in step S104 may specifically include: S1044. Sort each target path in descending order of recommendation value to obtain a set of candidate paths.

[0107] The candidate path set refers to an ordered sequence of target paths arranged in descending order of recommendation value, used to select supply-demand matching paths according to priority. Specifically, the recommendation value set first calculated in step S1043... In the first step, all recommended values ​​are sorted in descending order. Then, the order of the target paths is adjusted based on the sorting result, placing the target path with the highest recommended value first, the second highest second, and so on. Finally, a set of candidate paths is obtained. ,in Indicates the sorted order of the first... The target path of the bit and satisfying .

[0108] S1045. Calculate the transaction volume of each target path in the candidate path set in turn until it equals the procurement quota, and determine the target path that has been accumulated in the candidate path set as the supply and demand matching scheme.

[0109] A supply-demand matching scheme refers to the combination of target paths selected from the candidate path set in order of priority, where the total transaction volume of all selected paths equals the procurement quota of the demand node.

[0110] Specifically, the procurement quota for the demand node is first calculated from step S1031. Extract the target total purchase volume from the candidate path set. The first target path First, the transaction volume of each transportation route in each target path is extracted sequentially, and these transaction volumes are accumulated to obtain the cumulative transaction volume. ,in This indicates the current number of target paths that have been accumulated. Indicates the target path The transaction volume of the transportation route with the lowest transaction volume. Then, determine the cumulative transaction volume. Is it equal to the procurement quota? ,when Stop accumulating when the time is right. Finally, combine the accumulated paths from the candidate path set. Target Path Determined as a supply and demand matching solution .

[0111] This embodiment implements a path combination optimization mechanism based on recommended value priority and a precise quota matching strategy. It not only ensures the priority utilization of high-quality paths but also achieves precise matching between priority-based path combinations and procurement needs, avoiding excessive resource consumption by inefficient paths. It overcomes the lack of path priority ranking and precise quota control mechanisms in existing technologies, significantly improving the resource utilization efficiency and scientific basis of supply and demand matching schemes.

[0112] Secondly, this application provides a knowledge graph-based intelligent matching and recommendation system for automotive parts supply and demand, including: Figure 4 This application provides a schematic diagram of a specific implementation of a knowledge graph-based intelligent matching and recommendation system for automotive parts supply and demand, with reference to... Figure 4 The system may include: The acquisition module 410 is used to synchronously acquire the upper limit of the transfer volume and the transfer time of the logistics node, the historical order data of the demand node in the target area, and the knowledge graph with supply nodes, logistics nodes, and demand nodes as nodes and the transportation lines between nodes as edges. The generation module 420 is used to calculate the fluctuation coefficient and demand saturation of historical order data within a preset sampling window, obtain the demand matrix, and construct a topology graph with the upper limit of the transfer volume as the transfer volume threshold of the logistics node in the knowledge graph and the transfer timeliness as the edge weight of the transportation route. Module 430 is used to determine the procurement quota of the demand nodes in the topology graph based on the demand matrix, and to use the edge weight as the conversion cost of the transportation route. Iterative bidding is carried out between the demand nodes and the supply nodes. The bidding requests are arbitrated using the transshipment volume threshold, and the bidding score of the transportation route is adjusted according to the ratio of the load of the logistics node to the transshipment volume threshold. When the bidding requests and the supply nodes reach an allocation balance, the supply and demand flow graph is constructed with the transaction volume of the transportation route as the weight. The calculation module 440 is used to determine the target path based on the supply and demand flow diagram, calculate the recommended value of the target path based on the transaction volume and transit time, and combine all target paths according to the recommended value to obtain the supply and demand matching scheme.

[0113] The knowledge graph-based intelligent matching and recommendation system for automotive parts supply and demand in this application is used to implement the aforementioned knowledge graph-based intelligent matching and recommendation method for automotive parts supply and demand. Therefore, the specific implementation of the knowledge graph-based intelligent matching and recommendation system for automotive parts supply and demand can be found in the embodiment section of the knowledge graph-based intelligent matching and recommendation method for automotive parts supply and demand described above. The specific implementation can be referred to the description of the corresponding embodiments, and will not be repeated here.

[0114] Figure 5 A schematic diagram of the hardware structure of an electronic device provided in one embodiment of this application is shown.

[0115] The electronic device may include a processor 510 and a memory 520 storing computer program instructions.

[0116] Specifically, the processor 510 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0117] Memory 520 may include mass storage for data or instructions. For example, and not limitingly, memory 520 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 520 may include removable or non-removable (or fixed) media. Where appropriate, memory 520 may be internal or external to the integrated gateway disaster recovery device. In a particular embodiment, memory 520 is non-volatile solid-state memory.

[0118] Memory may include read-only memory (ROM), random access memory (RAM), disk storage media devices, optical storage media devices, flash memory devices, and electrical, optical, or other physical / tangible memory storage devices. Therefore, typically, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to the first aspect of this disclosure.

[0119] The processor 510 reads and executes computer program instructions stored in the memory 520 to implement any of the knowledge graph-based intelligent matching and recommendation methods for automotive parts supply and demand in the above embodiments.

[0120] In one example, the electronic device may also include a communication interface 530 and a bus 540. Wherein, such as Figure 5 As shown, the processor 510, memory 520, and communication interface 530 are connected through bus 540 and complete communication with each other.

[0121] The communication interface 530 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0122] Bus 540 includes hardware, software, or both, that couples components of an online data traffic metering device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 540 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, any suitable bus or interconnect is contemplated herein.

[0123] The electronic device can execute the knowledge graph-based intelligent matching and recommendation method for automotive parts supply and demand in the embodiments of this application, thereby realizing the knowledge graph-based intelligent matching and recommendation method for automotive parts supply and demand described in conjunction with the accompanying drawings.

[0124] Furthermore, in conjunction with the knowledge graph-based intelligent matching and recommendation method for automotive parts supply and demand in the above embodiments, this application embodiment can provide a computer-readable storage medium for implementation. This computer-readable storage medium stores computer program instructions; when executed by a processor, these computer program instructions implement any of the knowledge graph-based intelligent matching and recommendation methods for automotive parts supply and demand in the above embodiments.

[0125] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0126] The functional blocks shown in the above-described structural diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0127] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0128] The aspects of this application have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by dedicated hardware performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0129] The above provides a detailed description of the intelligent matching and recommendation method and system for automotive parts supply and demand based on knowledge graphs provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are only for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make several improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A knowledge graph-based intelligent matching and recommendation method for automotive parts supply and demand, characterized in that, include: Synchronously acquire the upper limit of transfer volume and transfer time of logistics nodes, historical order data of demand nodes in the target area, and a knowledge graph with supply nodes, logistics nodes, demand nodes as nodes and transportation lines between nodes as edges; The fluctuation coefficient and demand saturation of the historical order data within a preset sampling window are calculated to obtain a demand matrix. A topology graph is constructed using the upper limit of the transshipment volume as the transshipment volume threshold of the logistics node in the knowledge graph and the transshipment timeliness as the edge weight of the transportation route. The procurement quota of the demand node in the topology graph is determined based on the demand matrix, and the conversion cost of the transportation route is taken as the edge weight. Iterative bidding is carried out between the demand node and the supply node. The bidding request is arbitrated using the transshipment volume threshold, and the bidding score of the transportation route is adjusted according to the ratio of the load of the logistics node to the transshipment volume threshold. When the bidding request and the supply node reach a distribution balance, the supply and demand flow graph is constructed with the transaction volume of the transportation route as the weight. The target path is determined based on the supply and demand flow diagram. The recommended value of the target path is calculated based on the transaction volume and the transit time. All the target paths are combined according to the recommended value to obtain a supply and demand matching scheme.

2. The method according to claim 1, characterized in that, The step of combining all the target paths according to the recommended values ​​to obtain a supply-demand matching scheme includes: Each target path is sorted in descending order of the recommended value to obtain a set of candidate paths; The transaction volume of each target path in the candidate path set is calculated by accumulating the transaction volume until it equals the procurement quota. The target paths that have been accumulated in the candidate path set are then determined as the supply and demand matching scheme.

3. The method according to claim 1, characterized in that, The calculation of the fluctuation coefficient and demand saturation of the historical order data within a preset sampling window yields a demand matrix. A topology graph is then constructed using the upper limit of the transshipment volume as the transshipment volume threshold for logistics nodes in the knowledge graph and the transshipment timeliness as the edge weight of the transportation route. This includes: Extract the order volume of the historical order data for each unit time period within a preset sampling window, calculate the mean and standard deviation of all the order volumes, and obtain the fluctuation coefficient by calculating the ratio of the standard deviation to the mean; The demand saturation is obtained by calculating the ratio of the mean value to the order quantity with the largest value within the preset sampling window; Based on the volatility coefficient and the demand saturation, a demand vector for each demand node is constructed, and all demand vectors are combined according to the order of each demand node in the knowledge graph to obtain the demand matrix. The topology graph is constructed by using the upper limit of the transshipment volume of each logistics node as the threshold of the transshipment volume of the corresponding logistics node in the knowledge graph, and using the transshipment timeliness of each logistics node as the edge weight of the corresponding transportation route.

4. The method according to claim 1, characterized in that, Based on the demand matrix, the procurement quotas of the demand nodes in the topology graph are determined, and the edge weights are used as the conversion costs of the transportation routes. Iterative bidding is conducted between the demand nodes and the supply nodes. A transit volume threshold is used to arbitrate the bidding requests, and the bidding score of the transportation routes is adjusted according to the ratio of the load capacity of the logistics nodes to the transit volume threshold. When a distribution balance is reached between the bidding requests and the supply nodes, a supply and demand flow graph is constructed using the transaction volume of the transportation routes as the weight, including: The demand vector corresponding to each demand node is extracted from the demand matrix. The procurement quota of each demand node in the topology graph is obtained by weighting the fluctuation coefficient and demand saturation included in each demand vector. Using the conversion cost as the initial value, bidding requests including the procurement quota are sent from each of the demand nodes to each of the supply nodes, and the flow value generated by each of the transportation lines in response to the bidding request is obtained. The load of each of the logistics nodes is obtained by summing the flow values ​​of each of the transportation lines passing through the same logistics node. The flow value is generated by allocating the procurement quota to each of the transportation lines using the bidding score. The bidding score for the corresponding transportation route is adjusted based on the ratio of the load to the transfer volume threshold, and the transfer volume threshold is used as the upper limit constraint for the corresponding load. When the total number of flow values ​​flowing into each demand node is equal to the corresponding procurement quota and the flow value on the corresponding transportation route no longer changes, the flow value is determined as the transaction volume, and the transaction volume is used as the weight of the transportation route in the knowledge graph to obtain the supply and demand flow diagram.

5. The method according to claim 4, characterized in that, The method further includes: The bidding sum is obtained by calculating the sum of the bidding scores corresponding to the multiple transportation routes associated with each of the demand nodes; The allocation weight of each transportation route is determined based on the ratio of the bidding score to the bidding sum for each transportation route, and the product of the procurement quota and the allocation weight is calculated to obtain the flow value of each transportation route.

6. The method according to claim 4, characterized in that, The step of adjusting the bidding score of the corresponding transportation route based on the ratio of the load capacity to the transshipment volume threshold, and using the transshipment volume threshold as the upper limit constraint of the corresponding load capacity, includes: Calculate the ratio of the load of each logistics node to the corresponding transfer volume threshold to obtain the load coefficient, and obtain the adjustment amount of each logistics node based on the product of the load coefficient and the preset step size parameter. The bidding score for the corresponding transportation route is negatively corrected using the adjustment amount. During the iterative bidding process, when the load of each logistics node is equal to the corresponding transshipment volume threshold, the allocation of the flow value to the transportation line passing through the corresponding logistics node is stopped, so as to limit the load to the range of the transshipment volume threshold.

7. The method according to claim 1, characterized in that, The step of determining the target path based on the supply and demand flow diagram and calculating the recommended value of the target path based on the transaction volume and the transit time includes: The target path is obtained by extracting the connected sequence of transportation routes that satisfy the condition that the transaction volume is greater than zero from the supply and demand flow diagram; The total path time is obtained by summing the transit times corresponding to each of the transport lines in each target path, and the average transaction volume corresponding to each of the transport lines in each target path is calculated to obtain the average path flow. The recommended value for each target path is obtained based on the ratio of the average traffic to the total time taken for the path.

8. A knowledge graph-based intelligent matching and recommendation system for automotive parts supply and demand, characterized in that, include: The acquisition module is used to synchronously acquire the upper limit of the transfer volume and the transfer time of logistics nodes, the historical order data of demand nodes in the target area, and the knowledge graph with supply nodes, logistics nodes, demand nodes as nodes and transportation lines between nodes as edges. The generation module is used to calculate the fluctuation coefficient and demand saturation of the historical order data within a preset sampling window to obtain a demand matrix, and to construct a topology graph using the upper limit of the transshipment volume as the transshipment volume threshold of the logistics node in the knowledge graph and the transshipment timeliness as the edge weight of the transportation route. The construction module is used to determine the procurement quota of the demand node in the topology graph according to the demand matrix, and use the edge weight as the conversion cost of the transportation line. Iterative bidding is carried out between the demand node and the supply node. The bidding request is arbitrated using the transshipment volume threshold, and the bidding score of the transportation line is adjusted according to the ratio of the load of the logistics node to the transshipment volume threshold. When the bidding request and the supply node reach an allocation balance, the supply and demand flow graph is constructed with the transaction volume of the transportation line as the weight. The calculation module is used to determine the target path based on the supply and demand flow diagram, calculate the recommended value of the target path based on the transaction volume and the transit time, and combine all the target paths according to the recommended value to obtain a supply and demand matching scheme.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the knowledge graph-based intelligent matching and recommendation method for automotive parts supply and demand as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, enables the implementation of the knowledge graph-based intelligent matching and recommendation method for automotive parts supply and demand as described in any one of claims 1 to 7.

Images