AI-Based Intelligent Dynamic Transaction Matching Method, Device, Medium and Equipment for Electronic Components

Abstract

The present disclosure relates to an AI-based intelligent dynamic trading matching method, device, medium and equipment for electronic components, including: in response to a trading matching request, dynamically determining a matching target trading party from the trading sellers of the electronic components corresponding to the sold electronic component names according to the electronic component names; constructing a trading weighted directed graph with the trading matching intermediary as the intermediate node, the matching target trading party as the source point, and the trading matching request as the sink point according to the matching request quantity carried in the trading matching request and the sold quantity of the matching target trading party; determining multiple matching paths by finding the maximum tradable volume of the augmented path according to the capacity of the edges in each path from each source point to the sink point through the intermediate node in the trading weighted directed graph; determining an intelligent dynamic trading matching result according to the maximum tradable volume in the multiple matching paths, and displaying the introduction information of the electronic components in response to the selected browsing operation on the intelligent dynamic trading matching result.

Description

Technical Field

[0001] The present invention relates to the cross - technical field of e - commerce and artificial intelligence, and particularly relates to an intelligent dynamic trading matching method, device, medium and equipment for electronic components based on AI. Background Art

[0002] Traditional B2B platforms rely on manual inquiry, with low price transparency (for example, the price difference of a certain capacitor model reaches 30% in different channels), low matching efficiency, and an average of 5 - 7 days of manual negotiation for a deal. During special periods, the delivery cycle fluctuates, resulting in 60% of orders needing to be rematched. Summary of the Invention

[0003] The purpose of the present invention is to provide an intelligent dynamic trading matching method, device, medium and equipment for electronic components based on AI, so as to solve the technical problem of low trading matching efficiency of electronic components in related scenarios.

[0004] The first aspect of the present disclosure provides an intelligent dynamic trading matching method for electronic components based on AI. The method is applied to a trading matching intermediary, and the method includes:

[0005] In response to receiving a trading matching request, according to the name of the electronic component carried in the trading matching request, dynamically determine a matching target trading party for electronic component trading matching from the trading sellers selling the electronic component corresponding to the name of the electronic component;

[0006] According to the matching request quantity carried in the trading matching request and the selling quantities of each of the matching target trading parties, construct a weighted directed trading graph with the trading matching intermediary as the intermediate node, the matching target trading party as the source point, and the trading matching request as the sink point, where the capacity of the edge in the weighted directed trading graph is determined according to the selling quantity and the matching request quantity;

[0007] According to the capacity of the edges in each path from each source point to the sink point through the intermediate node in the weighted directed trading graph, determine multiple matching paths for the matching target trading party to trade with the trading request party through the intermediate node by finding the maximum tradable quantity of the augmenting path;

[0008] According to the maximum tradable quantity in the multiple matching paths, determine an intelligent dynamic trading matching result for the trading matching request, and in response to a selected browsing operation on the displayed intelligent dynamic trading matching result, display the introduction information of the electronic component by the matching target trading party corresponding to the selected intelligent dynamic trading matching result.

[0009] In a possible implementation manner, based on the matching request quantity carried in the transaction matching request and the selling quantities of each of the matching target trading parties, a weighted directed graph of transactions is constructed with the transaction matching intermediary as the intermediate node, the matching target trading parties as the source points, and the transaction matching request as the sink point, including:

[0010] Dynamically determine the credibility coefficients of each of the matching target trading parties according to the historical performance fulfillment rates of each of the matching target trading parties for the selling quantities;

[0011] According to the transaction quotes of the matching target trading parties for the electronic components and the expected price corresponding to the transaction matching request, determine the price-sensitive edge weights from each of the source points to the intermediate node in the weighted directed graph of transactions;

[0012] According to the committed delivery cycle of the matching target trading parties for the electronic components, the historical overdue days corresponding to the matching target trading parties, and the expected delivery cycle corresponding to the transaction matching request, determine the delivery-time-sensitive edge weights from each of the source points to the intermediate node in the weighted directed graph of transactions;

[0013] According to the credibility coefficients of each of the matching target trading parties, the price-sensitive edge weights, the delivery-time-sensitive edge weights, and the product of the smaller value between the selling quantities corresponding to each of the matching target trading parties and the matching request quantity, determine the capacities of the edges from each of the source points corresponding to the matching target trading parties to the intermediate node in the weighted directed graph of transactions;

[0014] According to the matching request quantity carried in the transaction matching request, determine the capacity of the edge from the sink point to the intermediate node in the weighted directed graph of transactions, and construct the weighted directed graph of transactions according to the capacities of the edges from each of the source points corresponding to the matching target trading parties to the intermediate node and the capacity of the edge from the sink point to the intermediate node in the weighted directed graph of transactions.

[0015] In a possible implementation manner, the step of determining the price-sensitive edge weights from each of the source points to the intermediate node in the weighted directed graph of transactions according to the transaction quotes of the matching target trading parties for the electronic components and the expected price of the transaction matching request includes:

[0016] Determine the transaction quote diameter according to the first quote difference between the highest transaction quote and the lowest transaction quote in the transaction quotes of the matching target trading parties for the electronic components;

[0017] Determine the average transaction quote according to the transaction quotes of each of the matching target trading parties for the electronic components and the time offsets of each of the transaction quotes from the current time;

[0018] Determine the trading price offset corresponding to each of the matching target trading parties according to the absolute value of the second price difference between the trading price of each of the matching target trading parties for the electronic component and the average trading price;

[0019] Determine the price-sensitive edge weights from each of the source nodes to the intermediate node in the trading weighted directed graph according to the product of the ratio of the trading price offset corresponding to each of the matching target trading parties to the trading price diameter and the weight attenuation coefficient corresponding to the preset deviation threshold range where the ratio is located, where each of the preset deviation threshold ranges is correspondingly set with the weight attenuation coefficient.

[0020] In a possible implementation manner, the determining the average trading price according to the trading price of each of the matching target trading parties for the electronic component and the time offset of each of the trading prices from the current time includes:

[0021] Determine the adaptive volatility according to the preset volatility and the mean and standard deviation of the trading prices corresponding to each of the matching target trading parties within the historical preset duration from the current time;

[0022] Determine the attenuation weight corresponding to each of the matching target trading parties according to the time offset of the trading price corresponding to each of the matching target trading parties from the current time and the adaptive volatility;

[0023] Determine the attenuated trading price corresponding to each of the matching target trading parties according to the product of the attenuation weight corresponding to each of the matching target trading parties and the trading price of the corresponding electronic component;

[0024] Sum up the attenuated trading prices corresponding to each of the matching target trading parties to determine the total sum of the attenuated trading prices, and sum up the attenuation weights corresponding to each of the matching target trading parties to determine the attenuation coefficient;

[0025] Determine the average trading price according to the ratio of the total sum of the attenuated trading prices to the attenuation coefficient.

[0026] In a possible implementation manner, the determining the delivery-time sensitive edge weights from each of the source nodes to the intermediate node in the trading weighted directed graph according to the committed delivery cycle of the matching target trading party for the electronic component, the historical overdue days corresponding to the matching target trading party, and the expected delivery cycle corresponding to the trading matching request includes:

[0027] Determine the performance attenuation rate corresponding to each of the matching target trading parties according to the historical overdue days corresponding to each of the matching target trading parties;

[0028] Determine the weights of the delivery-time sensitive edges from each source point to the intermediate node in the transaction weighted directed graph according to the performance attenuation rate corresponding to each matching target trading party and the cycle difference between the promised delivery cycle of each matching target trading party for the electronic component and the expected delivery cycle corresponding to the transaction matching request.

[0029] In a possible implementation manner, according to the capacities of the edges in each path from each source point to the sink point via the intermediate node in the transaction weighted directed graph, by finding the maximum tradable volume of the augmenting path, determine multiple matching paths for the matching target trading party to trade with the transaction requester via the intermediate node, including:

[0030] Construct a residual network according to each path from each source point to the sink point via the intermediate node in the transaction weighted directed graph;

[0031] Determine the residual capacities of all edges in the residual network according to breadth-first search or depth-first search;

[0032] According to the capacities of the edges in each path from each source point to the sink point via the intermediate node in the transaction weighted directed graph, by finding the maximum tradable volume of the augmenting path, calculate the bottleneck capacity of a single path where a matching target trading party supplies alone to meet the matching request volume, or the bottleneck capacity of a combined path where multiple matching target trading parties jointly supply to meet the matching request volume;

[0033] Update the residual capacities of all edges in the residual network according to the bottleneck capacity of each single path and / or each combined path. When it is impossible to perform augmenting path search in the residual network, obtain multiple single paths and / or combined paths;

[0034] Determine multiple matching paths for the matching target trading party to trade with the transaction requester via the intermediate node according to multiple single paths and / or combined paths.

[0035] In a possible implementation manner, in response to receiving a transaction matching request, dynamically determine a matching target trading party for electronic component transaction matching from the transaction sellers selling the electronic components corresponding to the electronic component name carried in the transaction matching request, including:

[0036] In response to receiving a transaction matching request, query the transaction sellers registered with the electronic components corresponding to the electronic component name in the transaction matching intermediary according to the electronic component name carried in the transaction matching request;

[0037] Determine the historical view times and view frequencies of each of the transaction sellers on the transaction matching intermediary;

[0038] According to the historical view times and view frequencies of each of the transaction sellers on the transaction matching intermediary, dynamically determine, from the transaction sellers selling the electronic components corresponding to the electronic component name, the matching target transaction parties for electronic component transaction matching.

[0039] A second aspect of the present disclosure provides an AI-based intelligent dynamic transaction matching device for electronic components, which is applied to a transaction matching intermediary. The device includes:

[0040] A first determination module, configured to, in response to receiving a transaction matching request, dynamically determine, from the transaction sellers selling the electronic components corresponding to the electronic component name carried in the transaction matching request, the matching target transaction parties for electronic component transaction matching;

[0041] A second determination module, configured to construct a transaction weighted directed graph with the transaction matching intermediary as an intermediate node, the matching target transaction parties as source points, and the transaction matching request as a sink point according to the matching request quantity carried in the transaction matching request and the sales volumes of each of the matching target transaction parties, wherein the capacity of the edge in the transaction weighted directed graph is determined according to the sales volume and the matching request quantity;

[0042] A third determination module, configured to determine multiple matching paths for the matching target transaction parties to transact with the transaction request party via the intermediate node by finding the maximum tradable volume of the augmenting path according to the capacities of the edges in each path from each of the source points to the sink point in the transaction weighted directed graph;

[0043] A fourth determination module, configured to determine an intelligent dynamic transaction matching result for the transaction matching request according to the maximum tradable volume in the multiple matching paths, and display the introduction information of the electronic components of the matching target transaction party corresponding to the selected intelligent dynamic transaction matching result in response to a selected browsing operation on the displayed intelligent dynamic transaction matching result.

[0044] A third aspect of the present disclosure provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the method described in any item of the first aspect are implemented.

[0045] A fourth aspect of the present disclosure provides an electronic device, including:

[0046] A memory, on which a computer program is stored;

[0047] A processor for executing the computer program in the memory to implement the steps of the method according to any one of the first aspects.

[0048] The above technical solution can at least have the following beneficial effects:

[0049] By responding to transaction requests in real time and dynamically screening target sellers, the inefficiency of traditional traversal matching is avoided, and the matching time complexity is reduced. The transaction problem is transformed into a maximum flow problem of a weighted directed graph, and the augmenting path algorithm is used to quickly calculate the optimal allocation scheme to ensure the splitting and path planning of complex transaction volumes within milliseconds. By constructing a weighted graph from multiple source points (sellers) to a sink point (requestor), large orders can be intelligently split among multiple suppliers to avoid transaction failures caused by insufficient inventory of a single supplier. The capacity of the edge is jointly determined by the real-time sales volume and the request volume to ensure the feasibility of the transaction. The transaction plan is displayed with multiple matching paths, allowing users to intuitively compare different combinations in terms of price, delivery date, etc. After selecting the matching result, the detailed information of the supplier is directly displayed, reducing the user's jump links and improving the decision-making efficiency. At the same time, the construction and solution of the transaction weighted directed graph can be deployed in a distributed computing framework to support real-time computing of a scale of millions of nodes, adapting to the large-scale electronic component trading market. When a certain path fails due to a supplier's temporary out-of-stock, the augmenting path can be quickly recalculated to ensure the continuity of transaction matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The present invention will be further described below with reference to the accompanying drawings.

[0051] Figure 1 is a flowchart of an AI-based intelligent dynamic transaction matching method for electronic components shown according to an embodiment of the specification.

[0052] Figure 2 is a block diagram of an AI-based intelligent dynamic transaction matching device for electronic components shown according to an embodiment of the specification.

[0053] Figure 3 is a block diagram of a display device shown according to an embodiment of the specification. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0054] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments of the present invention belong to the scope of protection of the present invention.

[0055] Such as Figure 1The figure shows a flowchart of an AI-based intelligent dynamic trading matching method for electronic components according to the present invention. The method is applied to a trading matching intermediary. The trading matching intermediary can provide, for example, a website or an application program to terminal devices such as computers and mobile phones through a server. Users can generate trading matching requests through operations such as searching or querying on the website or application program. The method includes:

[0056] In step S11, in response to receiving a trading matching request, according to the name of the electronic component carried in the trading matching request, dynamically determine a matching target trading party for electronic component trading matching from trading sellers selling the electronic component corresponding to the name of the electronic component;

[0057] Among them, dynamic determination can be based on real-time data (such as inventory quantity, supplier score, historical transaction success rate, etc.) and an AI model to screen eligible sellers in real time, rather than relying on static rules. The matching of the electronic component name can be achieved by natural language processing (NLP) or a predefined component classification system to accurately match the component name in the request with the components in the seller database.

[0058] In the embodiments of the present disclosure, parameters such as the name of the electronic component, the matching request quantity, and the quality requirements in the trading matching request are parsed. Furthermore, an AI model (such as a GBDT or a deep learning model) can dynamically rank the sellers according to the following features:

[0059] Inventory matching degree: Whether the current inventory meets the demand quantity.

[0060] Transaction evaluation: Historical transaction scores (such as on-time delivery rate, quality qualification rate).

[0061] Price competitiveness: The ratio of the quoted price to the market price.

[0062] Output: A list of target sellers sorted by priority.

[0063] Illustrative example: The trading matching request is to purchase 1000 "Resistor X". Suppliers A (score 4.9) with inventory ≥ 1000, Supplier B (score 4.8) with inventory 800, and Supplier C (score 4.7) with inventory 500 are screened out. After dynamic ranking, A is preferentially selected. If the inventory of A is insufficient, B and C are matched in turn.

[0064] In step S12, according to the matching request quantity carried in the trading matching request and the selling quantity of each matching target trading party, construct a weighted directed trading graph with the trading matching intermediary as the intermediate node, the matching target trading party as the source point, and the trading matching request as the sink point. The capacity of the edge in the weighted directed trading graph is determined according to the selling quantity and the matching request quantity;

[0065] Among them, each edge in the weighted directed graph has a direction (such as from the supplier to the intermediate party to the requester) and a weight (capacity), representing the upper limit of the transaction flow. The intermediate node is the intermediate party for transaction matching, acting as the hub node in the graph, connecting all source nodes (suppliers) and sink nodes (requesters).

[0066] In the embodiments of the present disclosure, the capacity of the edge from the source node to the intermediate node in the transaction weighted directed graph can be determined according to the current available inventory (i.e., the sold quantity) of the supplier and the matching request quantity. The capacity of the edge from the intermediate node to the sink node in the transaction weighted directed graph is the transaction request quantity (if the request quantity is split, the capacity of this edge is the total request quantity).

[0067] For example, to purchase 1000 components, the inventory of supplier A is 800, and the inventory of supplier B is 600.

[0068] Construct a graph: from A to the intermediate party (capacity 800), from B to the intermediate party (capacity 600), from the intermediate party to the requester (capacity 1000). At this time, an algorithm needs to be used to calculate how to allocate the inventory of A and B to meet the demand of 1000.

[0069] In step S13, according to the capacity of the edges in each path from each source node to the sink node through the intermediate node in the transaction weighted directed graph, by finding the maximum tradable volume of the augmenting path, multiple matching paths for the matching target trading party to trade with the transaction requester through the intermediate node are determined;

[0070] Among them, an augmenting path is a path from the source node to the sink node in the weighted directed graph, and the remaining capacity (current capacity - allocated quantity) of each edge on the path is greater than 0. The maximum tradable volume is the minimum remaining capacity (bottleneck value) in the augmenting path, which determines the maximum tradable volume that can be allocated for this path.

[0071] In the embodiments of the present disclosure, the Edmonds-Karp algorithm (the shortest augmenting path algorithm based on BFS) can be adopted, and the time complexity is O(E²V). The iterative process can be as follows: Find the augmenting path: find the shortest path from any source node to the sink node through BFS. Calculate the bottleneck value: the minimum value of the remaining capacity in the path. Update the allocation: allocate the bottleneck value to the edges on the path and reduce the remaining capacity. Repeat the iteration: until no new augmenting path can be found.

[0072] For example, the initial graph: from A to the intermediate party (capacity 800), from B to the intermediate party (capacity 600), from the intermediate party to the requester (capacity 1000).

[0073] The first iteration: find the path from A to the intermediate party to the requester, and the bottleneck value is 800 (the capacity of A). After allocating 800, the remaining capacity of A is 0, and the remaining capacity from the intermediate party to the requester is 200.

[0074] Second iteration: Find the path from B to the middle party to the requesting party, with a bottleneck value of 200 (remaining request volume). After allocating 200, the remaining capacity of B is 400, and the total allocation volume of the requesting party reaches 1000.

[0075] Result: A provides 800 pieces, B provides 200 pieces, and the matching is completed.

[0076] In step S14, according to the maximum tradable volume in the multiple matching paths, determine the intelligent dynamic trading matching result for the trading matching request, and in response to the selected browsing operation for the displayed intelligent dynamic trading matching result, display the introduction information of the matching target trading party corresponding to the selected intelligent dynamic trading matching result for the electronic components.

[0077] Among them, the intelligent dynamic trading matching result is the output multi-path allocation scheme (such as supplier combination and allocation volume). The selected browsing operation is the user's interaction behavior with the matching result (such as clicking to view details).

[0078] In the embodiments of the present disclosure, convert the multi-path allocation scheme into structured data (such as JSON), including supplier ID, allocation volume, price, etc.

[0079] Front-end display: Matching result list: Display multiple feasible solutions according to priority (such as "Solution 1: A + B, total cost X; Solution 2: Only B, total cost Y"). After the user clicks on the solution, obtain the detailed introduction of the supplier through API calls (such as quality inspection reports, logistics information).

[0080] The above technical solution avoids the inefficiency of traditional traversal matching by responding to trading requests in real time and dynamically screening target sellers, reducing the matching time complexity. Convert the trading problem into a maximum flow problem of a weighted directed graph, and use the augmenting path algorithm to quickly calculate the optimal allocation scheme to ensure the splitting and path planning of complex trading volumes within milliseconds. By constructing a weighted graph from multiple source points (sellers) to a sink point (requesting party), large orders can be intelligently split among multiple suppliers to avoid transaction failures caused by insufficient inventory of a single supplier. The capacity of the edge is jointly determined by the real-time selling volume and the request volume to ensure transaction feasibility. Display the trading solutions with multiple matching paths, allowing users to intuitively compare dimensions such as price and delivery date of different combinations. Directly display the detailed supplier information after selecting the matching result, reducing the user's jump link and improving the decision-making efficiency. At the same time, the construction and solution of the trading weighted directed graph can be deployed in a distributed computing framework to support real-time computing of a scale of millions of nodes, adapting to the large-scale electronic component trading market. When a certain path fails due to a supplier's temporary out-of-stock, the augmenting path can be quickly recalculated to ensure the continuity of transaction matching.

[0081] In a possible implementation manner, in step S12, a weighted directed graph of transactions is constructed with the transaction matching intermediary as the intermediate node, the matching target trading parties as the source points, and the transaction matching request as the sink point according to the matching request quantity carried in the transaction matching request and the sales volumes of the respective matching target trading parties, including:

[0082] In step S121, the credibility coefficients of the respective matching target trading parties are dynamically determined according to the historical performance fulfillment rates of the respective matching target trading parties for the sales volume;

[0083] Among them, the historical performance fulfillment rate is the proportion of transactions (on-time delivery, qualified quality) completed by the supplier in the past. The performance fulfillment rate can be mapped to a coefficient in the [0, 1] interval, which is used to adjust the priority of the supplier in transaction matching.

[0084] In the embodiments of the present disclosure, the performance fulfillment records of the supplier in the past N transactions (such as the most recent 100 orders) are obtained in real time, and the historical performance fulfillment rate is calculated. The performance fulfillment rate is converted into a credibility coefficient through a non-linear function (such as Sigmoid) to amplify the advantages of a high performance fulfillment rate. For example, the credibility coefficient = 1 / (1 + e −k⋅(履约率−θ) )

[0085] In step S122, according to the transaction quotes of the matching target trading parties for the electronic components and the expected price corresponding to the transaction matching request, the price-sensitive edge weights from each source point to the intermediate node in the weighted directed graph of transactions are determined;

[0086] Among them, the price-sensitive edge weight is used to reflect the deviation degree between the supplier's quote and the requestor's expected price. For example, according to the difference between the transaction quote of the matching target trading party for the electronic component and the expected price corresponding to the transaction matching request, dividing by the expected price and then multiplying by the price sensitivity coefficient to obtain a price parameter value, and then determining the natural exponential function according to the price parameter value to obtain the price-sensitive edge weight.

[0087] In step S123, according to the committed delivery cycle of the matching target trading party for the electronic component, the historical overdue days corresponding to the matching target trading party, and the expected delivery cycle corresponding to the transaction matching request, the delivery-time-sensitive edge weights from each source point to the intermediate node in the weighted directed graph of transactions are determined;

[0088] Among them, the delivery-time-sensitive edge weight is used to comprehensively reflect the matching degree between the supplier's committed delivery cycle, historical overdue days and the requestor's expected cycle. For example, the delivery-time-sensitive edge weight can be determined according to the following formula:

[0089] Delivery-time deviation ΔT i Calculation: ΔT = committed delivery cycle t i- Expected cycle T;

[0090] Delivery risk A i = β × T i + γ × Number of days overdue in history, where β is the delivery deviation coefficient and γ is the penalty coefficient for historical overdue.

[0091] Convert the delivery risk A through the Softmax function i into a weight in the interval [0, 1]:

[0092] Weight of the delivery-sensitive edge: , where n is the number of matching target trading parties.

[0093] In step S124, according to the credibility coefficient of each of the matching target trading parties, the price-sensitive edge weight, the delivery-sensitive edge weight, and the product of the smaller value of the sales volume and the matching request volume corresponding to each of the matching target trading parties, determine the capacity of the edge from the source point to the intermediate node corresponding to each of the matching target trading parties in the transaction weighted directed graph;

[0094] Among them, the edge capacity is the maximum tradable quantity from the source point (supplier) to the intermediate node (transaction matching intermediary) in the weighted directed graph, and factors such as reliability, price, and delivery need to be considered comprehensively.

[0095] In the embodiment of the present disclosure, the edge capacity = credibility coefficient × price-sensitive edge weight × delivery-sensitive edge weight × min(sales volume, matching request volume).

[0096] In step S125, according to the matching request volume carried in the transaction matching request, determine the capacity of the edge from the sink point to the intermediate node in the transaction weighted directed graph, and construct the transaction weighted directed graph according to the capacity of the edge from the source point to the intermediate node corresponding to each of the matching target trading parties and the capacity of the edge from the sink point to the intermediate node in the transaction weighted directed graph.

[0097] In the embodiment of the present disclosure, the sink point is used to represent the transaction matching requester and needs to receive traffic from the intermediate node (matching intermediary). The transaction matching problem can be transformed into a network flow problem, and optimal matching is achieved through edge capacity constraints.

[0098] In the embodiment of the present disclosure, the determination of the capacity of the sink point edge: the edge capacity from the sink point to the intermediate node is equal to the matching request volume. Graph structure construction: From the source point to the intermediate node: The capacity of each edge is the value calculated in step S124. From the intermediate node to the sink point: The edge capacity is the matching request volume. Find the maximum feasible flow from all source points to the sink point through the maximum flow algorithm (such as Ford-Fulkerson), which is the matching scheme.

[0099] The above technical solution converts the historical performance, price competitiveness, and delivery reliability of suppliers into computable weights. The credibility coefficient and the weights jointly determine the matching order of suppliers, avoiding the priority of suppliers with "low price but high risk". When constructing a weighted directed graph, cost, delivery time, and reliability are comprehensively considered instead of a single indicator. Furthermore, reliability, price, delivery time, and supply / demand are comprehensively converted into edge capacities, avoiding single-indicator decision-making. The capacity of the edge changes with the real-time status of the supplier (such as inventory, quotation), supporting dynamic adjustment. The global optimal matching solution is found through the maximum flow algorithm instead of the local optimal. The edge capacity of suppliers with high price / delivery risk is relatively low, reducing the probability of being selected.

[0100] In a possible implementation manner, in step S122, determining the price-sensitive edge weights from each source point to the intermediate node in the transaction weighted directed graph according to the transaction quotation of the matching target trading party for the electronic component and the expected price of the transaction matching request includes:

[0101] In step S1221, according to the first quotation difference between the highest transaction quotation and the lowest transaction quotation in the transaction quotation of the matching target trading party for the electronic component, determine the transaction quotation diameter;

[0102] Among them, the transaction quotation diameter (Price Diameter) is the difference between the maximum value and the minimum value in all the quotations of the target transaction sellers, reflecting the discreteness of the quotations and measuring the quotation fluctuation range.

[0103] In step S1222, according to the transaction quotations of each matching target trading party for the electronic component and the time offset of each transaction quotation from the current time, determine the average transaction quotation;

[0104] Among them, the time offset (Time Offset) is the time difference between the current time and the quotation time (such as hours / days), used to measure the timeliness of the quotation. The average transaction quotation is weighted according to the time offset, and the newer the quotation, the higher the weight.

[0105]

[0106] Among them, V is the average transaction quotation calculated this time, v i is the transaction quotation of the i-th matching target trading party, and △t i is the time offset of the i-th matching target trading party. is the attenuation coefficient, used to control the influence speed of the time weight and reduce the weight of outdated quotations.

[0107] In step S1223, according to the absolute value of the second quotation difference between the transaction quotations of each of the matching target trading parties for the electronic component and the average transaction quotation, determine the transaction quotation offset corresponding to each of the matching target trading parties;

[0108] Among them, the quotation offset (Price Offset) is the absolute difference between a single quotation and the average transaction quotation, reflecting the degree of outlier of the quotation.

[0109] In step S1224, according to the product of the ratio of the transaction quotation offset corresponding to each of the matching target trading parties and the transaction quotation diameter and the weight attenuation coefficient corresponding to the preset deviation threshold range where the ratio is located, determine the price-sensitive edge weights from each of the source points to the intermediate node in the transaction weighted directed graph, where each of the preset deviation threshold ranges is correspondingly set with the weight attenuation coefficient.

[0110] Among them, the deviation threshold range is a preset offset ratio interval (such as 0-0.3, 0.3-0.6, 0.6-1.0), corresponding to different weight attenuation coefficients. The weight attenuation coefficient is a coefficient used to reduce the edge weight according to the degree of deviation.

[0111] In the embodiments of the present disclosure, the relative deviation degree of the quotation is quantified by the ratio, and this deviation is converted into an edge weight by using the threshold range and the attenuation coefficient. The larger the ratio, the farther the quotation deviates from the average, but the attenuation coefficient may reduce its influence to prevent extreme values from overly affecting the weight. Specific quotation data can be set, the diameter, offset, and ratio are calculated, and then the attenuation coefficient is determined according to the preset threshold range, and finally the price-sensitive edge weight is obtained.

[0112] By presetting multiple deviation threshold ranges, the continuous deviation ratio is divided into discrete risk levels, which is convenient for taking differential processing for quotations with different deviation degrees. According to the risk level interval where the quotation deviation ratio is located, the corresponding attenuation coefficient is selected and multiplied by the deviation ratio to obtain the final edge weight. The farther the deviation (the higher the risk), the greater the edge weight, but the attenuation coefficient will limit its growth rate to avoid extreme quotations from overly affecting the matching result.

[0113] The above technical solution combines time decay and offset to avoid a single quotation or outdated data from dominating the weight. The higher the edge weight of the supplier with a larger deviation from the average quotation, and the influence of extreme quotations is limited by the attenuation coefficient to prevent the weight from being too large or too small. Defining parameters such as the quotation diameter and offset is convenient for debugging and strategy adjustment.

[0114] In a possible implementation manner, in step S1222, determining the average transaction price according to the transaction quotes of each of the matching target trading parties for the electronic component and the time offset of each of the transaction quotes from the current time includes:

[0115] In step S12221, determine an adaptive volatility according to a preset volatility and the mean and standard deviation of the transaction quotes corresponding to each of the matching target trading parties within a historical preset duration from the current time;

[0116] Among them, volatility: an index measuring the degree of price fluctuation, usually represented by the standard deviation. Adaptive Volatility: a volatility parameter dynamically adjusted according to the mean, standard deviation, and preset volatility of historical quotes.

[0117] In the embodiments of the present disclosure, by combining the volatility of historical quotes and preset parameters, the volatility can adapt to market changes in different time periods. Calculate the mean (μ) and standard deviation (σ) of historical quotes. If the current price fluctuation (σ) exceeds the preset volatility threshold, the adaptive volatility is equal to σ; otherwise, it is equal to the preset value.

[0118] In step S12222, determine the decay weight corresponding to each of the matching target trading parties according to the time offset of the transaction quote corresponding to each of the matching target trading parties from the current time and the adaptive volatility;

[0119] Among them, Time Offset: the time difference between the current time and the quote time (such as hours / days). Decay Weight: a weight based on the time offset and adaptive volatility, used to reduce the impact of outdated quotes.

[0120] In the embodiments of the present disclosure, Decay Weight = e -自适应波动率×时间偏移量 , the larger the time offset (the older the quote), or the higher the adaptive volatility (the more unstable the market), the smaller the decay weight.

[0121] In step S12223, determine the decay transaction quote corresponding to each of the matching target trading parties according to the product of the decay weight corresponding to each of the matching target trading parties and the transaction quote of the corresponding electronic component;

[0122] Among them, the decay transaction quote is the product of the original quote and the decay weight, reflecting the current reference value of the quote. The decay weight of an outdated quote is lower, and its decay transaction quote is also lower, reducing its impact on the average value.

[0123] In step S12224, based on the decayed transaction quotes corresponding to each of the matched target trading parties, sum them up to determine the total decayed transaction quote, and based on the decayed weights corresponding to each of the matched target trading parties, sum them up to determine the decay coefficient;

[0124] Among them, the total decayed transaction quote is the accumulation of all decayed transaction quotes, which is used for subsequent weighted averaging. The total decay coefficient is the accumulation of all decayed weights to ensure weight normalization.

[0125] In step S12225, based on the ratio of the total decayed transaction quote to the decay coefficient, determine the average transaction quote.

[0126] Among them, the average transaction quote is the average value of the quotes considering time decay and volatility.

[0127] In the embodiments of the present disclosure, by the ratio of the total decayed transaction quote to the total decay coefficient, an average quote closer to the current market level is obtained.

[0128] The above technical solution automatically calibrates the volatility according to historical data, adapts to different market environments. Reduces the weight of outdated quotes and avoids interference from historical data in current decisions. At the same time, considering the volatility of quotes and time factors, makes the average quote more reasonable. Through the decay weight and adaptive volatility, clarifies the time value of the quote and the impact of market fluctuations.

[0129] In a possible implementation manner, in step S123, the determining the weights of the delivery-time sensitive edges from each of the source points to the intermediate nodes in the transaction weighted directed graph according to the committed delivery cycle of the matched target trading party for the electronic component, the corresponding historical overdue days of the matched target trading party, and the expected delivery cycle corresponding to the transaction matching request includes:

[0130] In step S1231, according to the corresponding historical overdue days of each of the matched target trading parties, determine the corresponding performance decay rate of each of the matched target trading parties;

[0131] Among them, the historical overdue days is the total number of days that the actual delivery cycle of the supplier exceeded the committed cycle in past transactions. The performance decay rate is an indicator used to measure the supplier's performance ability. The more overdue days, the higher the decay rate (indicating lower reliability).

[0132] In the embodiments of the present disclosure, the total overdue days of the supplier within the evaluation period (such as the past 12 months) are counted. Divide the overdue days by the total number of days in the evaluation period to obtain the performance decay rate.

[0133] In step S1232, according to the performance decay rate corresponding to each of the matched target trading parties and the cycle difference between the committed delivery cycle of each of the matched target trading parties for the electronic component and the desired delivery cycle corresponding to the transaction matching request, determine the lead-time sensitive edge weights from each of the source points to the intermediate node in the transaction weighted directed graph.

[0134] Among them, the committed lead time is the number of days promised by the supplier for delivery. The desired lead time is the number of days desired by the requester for delivery. The cycle difference is the difference between the committed delivery cycle and the desired cycle (which may be negative).

[0135] In the embodiments of the present disclosure, calculate the cycle difference: if the committed cycle > the desired cycle, the difference is positive; otherwise, it is negative. Combining with the performance decay rate, convert the difference into an edge weight.

[0136] The lead-time sensitive edge weight = |cycle difference| × (1 - performance decay rate), where the absolute value is used to ensure that the positive or negative of the difference does not affect the weight size, only focusing on the deviation amplitude. 1 - performance decay rate is used to reduce the impact of the decay rate (the higher the decay rate, the smaller the increase in weight).

[0137] The above technical solution converts the reliability of the supplier into a decay rate using the historical overdue days, avoiding subjective judgment. By multiplying the cycle difference and the decay rate, both the delivery deviation and the performance risk are punished at the same time. The performance decay rate makes the penalty for the delivery deviation of unreliable suppliers heavier. The edge weight directly reflects the superposition effect of the delivery deviation and the performance risk, facilitating the optimization of the subsequent matching algorithm.

[0138] In a possible implementation manner, in step S13, according to the capacities of the edges in each path from each source point to the sink point via the intermediate node in the transaction weighted directed graph, by finding the maximum tradable volume of the augmenting path, determine multiple matching paths for the matched target trading party to trade with the transaction requester via the intermediate node, including:

[0139] In step S131, according to each path from each source point to the sink point via the intermediate node in the transaction weighted directed graph, construct a residual network;

[0140] Among them, the residual network is an auxiliary network constructed based on the original weighted directed graph for calculating the maximum flow. The augmenting path is a feasible path from the source point to the sink point in the residual network, and the edge capacity of which has not been fully utilized.

[0141] In the embodiments of the present disclosure, reverse edges are added to each edge in the original graph, and the initial capacity is 0. The residual capacity is defined as the remaining capacity of the forward edge (initial capacity - allocated flow) or the reflux capacity of the reverse edge (allocated flow). The residual network is used to dynamically track the available capacity of the edges.

[0142] In step S132, according to breadth-first search or depth-first search, determine the residual capacity of all edges in the residual network;

[0143] Among them, breadth-first search (BFS) can traverse the nodes of the graph layer by layer to find the shortest path. Depth-first search (DFS) can traverse deeply along a single path and backtrack to find other paths. Residual Capacity is the remaining available capacity of the edge under the current allocation.

[0144] In the embodiments of the present disclosure, BFS or DFS is used to traverse the residual network to calculate the residual capacity of each edge. The residual capacity of the forward edge = initial capacity - allocated flow; the residual capacity of the reverse edge = allocated flow. If the residual capacity > 0, then the edge can participate in the augmenting path.

[0145] In step S133, according to the capacities of the edges in each path from each source point to the sink point via the intermediate node in the weighted directed graph of transactions, by finding the maximum tradable volume of the augmenting path, calculate the bottleneck capacity of a single path where a single matching target trading party supplies alone to meet the matching request volume, or the bottleneck capacity of a combined path where multiple matching target trading parties jointly supply to meet the matching request volume;

[0146] Among them, the bottleneck capacity is the edge with the smallest capacity in the augmenting path, which determines the maximum tradable volume of the path. A single path is a complete trading path that only involves one supplier. A combined path is a set of parallel or serial paths where multiple suppliers supply collaboratively.

[0147] In the embodiments of the present disclosure, for each augmenting path, traverse all its edges to find the minimum residual capacity as the bottleneck capacity. The bottleneck capacity of a single path directly limits the trading volume of that path. The bottleneck capacity of a combined path needs to comprehensively consider the capacity limitations of each sub-path (for example, take the minimum value for parallel paths and the sum of the minimum values of each segment for serial paths).

[0148] In step S134, according to the bottleneck capacity of each single path and / or each combined path, update the residual capacity of all edges in the residual network. When it is impossible to perform augmenting path search in the residual network, obtain multiple single paths and / or combined paths;

[0149] Among them, the residual capacity update adjusts the residual network according to the flow allocation of the augmenting path. The Max-Flow Min-Cut Theorem states that the maximum flow of a network is equal to the capacity of the minimum cut.

[0150] In the embodiments of the present disclosure, for each augmenting path, the flow is allocated according to the bottleneck capacity. Update the residual network: the residual capacity of the forward edge -= the bottleneck capacity. The residual capacity of the reverse edge += the bottleneck capacity. Repeat finding the augmenting path until there is no feasible path in the residual network (i.e., the maximum flow is achieved).

[0151] In step S135, according to the multiple single paths and / or the combined paths, multiple matching paths for the matching target trading party to trade with the trading request party via the intermediate node are determined.

[0152] In the embodiments of the present disclosure, all augmenting paths and their allocated flows are collected. According to the supplier combination of the paths and the flow allocation, a matching path list is generated. Ensure that the total allocated flow is equal to the trading request volume, and the flow of each path does not exceed its bottleneck capacity.

[0153] Through the above steps, the trading matching problem is transformed into a network maximum flow problem to ensure global optimality. The available capacity is updated in real time through the residual network, supporting multi-round path optimization and realizing dynamic capacity allocation. It allows flexible matching of single suppliers or combined suppliers, enhancing the robustness of transactions. Combining BFS / DFS and bottleneck capacity calculation ensures that a feasible solution can be found within polynomial time.

[0154] In a possible implementation manner, in step S11, in response to receiving a trading matching request, according to the electronic component name carried in the trading matching request, dynamically determining a matching target trading party for electronic component trading matching from the trading sellers selling the electronic components corresponding to the electronic component name includes:

[0155] In step S111, in response to receiving a trading matching request, according to the electronic component name carried in the trading matching request, querying the trading sellers registered with the electronic components corresponding to the electronic component name in the trading matching intermediary;

[0156] In the embodiments of the present disclosure, a database can be maintained to record all registered sold electronic components and their corresponding seller information. When a trading matching request is received, the system queries the database according to the electronic component name in the request to obtain a list of all sellers selling the component. The query results may include metadata such as the names of the sellers, historical trading records, inventory information, etc.

[0157] In step S112, determine the historical view count and view frequency of each of the transaction sellers at the transaction matching intermediary;

[0158] Among them, the historical view count is the total number of times the electronic components on the platform are viewed by potential buyers. The view frequency is the statistical value of the number of views within a preset time period (such as the daily average view volume).

[0159] In the embodiments of the present disclosure, a browsing log of each seller's electronic components is recorded, including the timestamp and viewer information. The number of views is calculated by counting the total number of browsing records of the component in the log. The view frequency is obtained by dividing the number of views within a time window (such as the most recent 30 days) by the length of the time window. Frequency calculation needs to consider time decay (such as higher weight for recent views).

[0160] In step S113, according to the historical view count and view frequency of each of the transaction sellers at the transaction matching intermediary, dynamically determine the matching target transaction parties for electronic component transaction matching from the transaction sellers selling the electronic components corresponding to the electronic component name.

[0161] In the embodiments of the present disclosure, by combining the number of views and the frequency, construct an "attention" index (such as weighted sum or product) of the seller. The attention index reflects the market attractiveness and potential transaction willingness of the seller. Sort the sellers according to the attention, and select the sellers with higher rankings as the matching targets. A threshold mechanism (such as only selecting sellers with a frequency higher than a certain value) or hierarchical screening (such as filtering by component type, region) can be introduced.

[0162] Through the above steps, use the browsing behavior data to quantify the attractiveness of the seller, avoiding manual subjective judgment. Update the browsing index in real time to reflect the changes in market heat, improving the dynamic adaptability. Narrow the matching range to high-attention sellers, reducing the computational complexity of the subsequent matching algorithm and improving the matching efficiency.

[0163] The embodiments of the present disclosure also provide an AI-based intelligent dynamic transaction matching device for electronic components, which is applied to the transaction matching intermediary. Refer to Figure 2 as shown, the device includes:

[0164] A first determination module 210, configured to, in response to receiving a transaction matching request, dynamically determine the matching target transaction parties for electronic component transaction matching from the transaction sellers selling the electronic components corresponding to the electronic component name carried in the transaction matching request;

[0165] A second determination module 220, configured to construct a weighted directed graph of transactions with the transaction matching intermediary as the intermediate node, the matching target trading parties as the source points, and the transaction matching request as the sink point, according to the matching request volume carried in the transaction matching request and the sales volumes of the respective matching target trading parties, wherein the capacity of the edges in the weighted directed graph of transactions is determined according to the sales volume and the matching request volume;

[0166] A third determination module 230, configured to determine multiple matching paths for the matching target trading parties to trade with the transaction request party via the intermediate node by finding the maximum tradable volume of the augmenting path according to the capacities of the edges in each path from each source point to the sink point via the intermediate node in the weighted directed graph of transactions;

[0167] A fourth determination module 240, configured to determine an intelligent dynamic transaction matching result for the transaction matching request according to the maximum tradable volume in the multiple matching paths, and display the introduction information of the matching target trading parties corresponding to the selected intelligent dynamic transaction matching result in response to a selected browsing operation on the displayed intelligent dynamic transaction matching result.

[0168] In a possible implementation manner, the second determination module 220 is configured to:

[0169] dynamically determine the credibility coefficients of the respective matching target trading parties according to the historical performance rates of the respective matching target trading parties for the sales volume;

[0170] determine the price-sensitive edge weights from each source point to the intermediate node in the weighted directed graph of transactions according to the transaction quotes of the matching target trading parties for the electronic components and the expected price corresponding to the transaction matching request;

[0171] determine the delivery-time-sensitive edge weights from each source point to the intermediate node in the weighted directed graph of transactions according to the committed delivery cycles of the matching target trading parties for the electronic components, the historical overdue days corresponding to the matching target trading parties, and the expected delivery cycle corresponding to the transaction matching request;

[0172] determine the capacities of the edges from the source points corresponding to the respective matching target trading parties to the intermediate node in the weighted directed graph of transactions according to the credibility coefficients of the respective matching target trading parties, the price-sensitive edge weights, the delivery-time-sensitive edge weights, and the product of the smaller value between the sales volume and the matching request volume corresponding to the respective matching target trading parties;

[0173] Determine the capacity of the edge from the sink node to the intermediate node in the weighted transaction digraph according to the matching request volume carried in the transaction matching request, and construct the weighted transaction digraph according to the capacity of the edge from the source node corresponding to each matching target trading party in the weighted transaction digraph to the intermediate node and the capacity of the edge from the sink node to the intermediate node.

[0174] In a possible implementation manner, the second determination module 220 is configured to:

[0175] Determine the transaction offer diameter according to the first offer difference between the highest transaction offer and the lowest transaction offer in the transaction offers of the matching target trading parties for the electronic component;

[0176] Determine the average transaction offer according to the transaction offers of each matching target trading party for the electronic component and the time offset of each transaction offer from the current time;

[0177] Determine the transaction offer offset corresponding to each matching target trading party according to the absolute value of the second offer difference between the transaction offer of each matching target trading party for the electronic component and the average transaction offer;

[0178] Determine the price-sensitive edge weights from each source node to the intermediate node in the weighted transaction digraph according to the product of the ratio of the transaction offer offset corresponding to each matching target trading party to the transaction offer diameter and the weight decay coefficient corresponding to the preset deviation threshold range in which the ratio is located, where each preset deviation threshold range is correspondingly set with the weight decay coefficient.

[0179] In a possible implementation manner, the second determination module 220 is configured to:

[0180] Determine the adaptive volatility according to the preset volatility and the mean and standard deviation of the transaction offers corresponding to each matching target trading party within the historical preset duration from the current time;

[0181] Determine the decay weight corresponding to each matching target trading party according to the time offset of the transaction offer corresponding to each matching target trading party from the current time and the adaptive volatility;

[0182] Determine the decay transaction offer corresponding to each matching target trading party according to the product of the decay weight corresponding to each matching target trading party and the transaction offer of the corresponding electronic component;

[0183] Sum up the decay transaction quotes corresponding to each of the matched target trading parties to determine the total decay transaction quote, and sum up the decay weights corresponding to each of the matched target trading parties to determine the decay coefficient;

[0184] Determine the average transaction quote according to the ratio of the total decay transaction quote to the decay coefficient.

[0185] In a possible implementation manner, the second determination module 220 is configured to:

[0186] Determine the performance decay rate corresponding to each of the matched target trading parties according to the historical overdue days corresponding to each of the matched target trading parties;

[0187] Determine the delivery-time sensitive edge weights from each source point to the intermediate node in the transaction weighted directed graph according to the performance decay rate corresponding to each of the matched target trading parties and the cycle difference between the committed delivery cycle of each of the matched target trading parties for the electronic component and the expected delivery cycle corresponding to the transaction matching request.

[0188] In a possible implementation manner, the third determination module 230 is configured to:

[0189] Construct a residual network according to each path from each source point through the intermediate node to the sink point in the transaction weighted directed graph;

[0190] Determine the residual capacity of all edges in the residual network according to breadth-first search or depth-first search;

[0191] According to the capacity of the edges in each path from each source point through the intermediate node to the sink point in the transaction weighted directed graph, calculate the bottleneck capacity of a single path where a single matched target trading party supplies alone to meet the matching request quantity, or the bottleneck capacity of a combined path where multiple matched target trading parties jointly supply to meet the matching request quantity by finding the maximum tradable quantity of the augmenting path;

[0192] Update the residual capacity of all edges in the residual network according to the bottleneck capacity of each single path and / or each combined path, and obtain multiple single paths and / or combined paths when it is impossible to execute the augmenting path search in the residual network;

[0193] Determine multiple matching paths for the matched target trading party to trade with the transaction requester through the intermediate node according to the multiple single paths and / or combined paths.

[0194] In a possible implementation manner, the first determination module 210 is configured to:

[0195] In response to receiving a transaction matching request, query, according to the name of the electronic component carried in the transaction matching request, a transaction seller of the electronic component corresponding to the name of the electronic component registered with the transaction matching intermediary;

[0196] Determine the historical view count and view frequency of each of the transaction sellers at the transaction matching intermediary;

[0197] According to the historical view count and view frequency of each of the transaction sellers at the transaction matching intermediary, dynamically determine a matching target transaction party for performing electronic component transaction matching from the transaction sellers selling the electronic component corresponding to the name of the electronic component.

[0198] An embodiment of the present disclosure further provides an electronic device, including:

[0199] A processor;

[0200] A memory for storing executable instructions executable by the processor;

[0201] Wherein, the processor is configured to execute the executable instructions stored in the memory to implement the method according to any one of the foregoing embodiments.

[0202] An embodiment of the present disclosure further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method according to any one of the foregoing embodiments are implemented.

[0203] Figure 3 The AI-based intelligent dynamic transaction matching device 100 for electronic components shown can be configured as a server, display a user interface through a terminal website, and then receive a transaction matching request or introduction information on the user interface. The device 100 includes: a processor 1001 and a memory 1003. Wherein, the processor 1001 and the memory 1003 are connected, such as connected through a bus 1002. Optionally, the AI-based intelligent dynamic transaction matching device 100 for electronic components may further include a communication component 1004, and the communication component 1004 can be used for data interaction between the device 100 and other devices, such as sending and / or receiving data, etc. It should be noted that in actual scheduling, the communication component 1004 is not limited to one, and the structure of the AI-based intelligent dynamic transaction matching device 100 for electronic components does not constitute a limitation to the embodiments of the present application.

[0204] The processor 1001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute various exemplary logical blocks, modules, and circuits described in connection with the disclosure of this application. The processor 1001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0205] The bus 1002 may include a path for transmitting information between the above components. The bus 1002 may be a PCI (Peripheral Component Interconnect) bus, an EISA (Extended Industry Standard Architecture) bus, or the like. The bus 1002 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 only a thick line is shown herein, but it does not mean that there is only one bus or one type of bus.

[0206] The memory 1003 may be a ROM (Read Only Memory) or other type of static storage device that can store static information and instructions, a RAM (Random Access Memory) or other type of dynamic storage device that can store information and instructions, or it may also be an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory), or other optical disc storage, optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media, other magnetic storage devices, or any other medium that can be used to carry or store program code and can be read by a computer, which is not limited herein.

[0207] The memory 1003 is used to store the program code for implementing the embodiments of the present disclosure, and is controlled by the processor 1001 for execution. The processor 1001 is configured to execute the program code stored in the memory 1003 to implement the steps shown in the foregoing embodiments of the AI-based intelligent dynamic trading matching method for electronic components.

[0208] The embodiments of the present disclosure further provide a computer-readable storage medium, on which program code is stored. When the program code is executed by a processor, the steps and corresponding content of the foregoing embodiments of the AI-based intelligent dynamic trading matching method for electronic components can be implemented.

[0209] The preferred embodiments of the present disclosure have been described in detail above in conjunction with the accompanying drawings. However, the present disclosure is not limited to the specific details in the above embodiments. Within the technical concept of the present disclosure, these embodiments can be subject to various changes, modifications, substitutions, and variations, and all of these changes, modifications, substitutions, and variations fall within the protection scope of the present disclosure.

[0210] In addition, it should be noted that, in the various specific technical features described in the above specific embodiments, without conflict, they can be combined in any appropriate manner, and the same should be regarded as the content disclosed by the present disclosure. To avoid unnecessary repetition, the present disclosure does not separately describe various possible combination manners. The technical scope of this application is not limited to the content in the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. An AI-based intelligent dynamic transaction matching method for electronic components, characterized in that: The method is applied to a transaction matching intermediary, and the method comprises: In response to receiving a transaction matching request, dynamically determining a matching target transaction party for electronic component transaction matching from transaction sellers that sell electronic components corresponding to the electronic component names according to the electronic component names carried in the transaction matching request; According to the matching request quantity carried in the transaction matching request and the sales quantity of each matching target transaction party, a transaction weighted directed graph is constructed with the transaction matching intermediary as the intermediate node, the matching target transaction party as the source point, and the transaction matching request as the sink point, including: dynamically determining the credibility coefficient of each matching target transaction party according to the historical fulfillment rate of each matching target transaction party for the sales quantity; determining the price-sensitive edge weights from each source point to the intermediate node in the transaction weighted directed graph according to the transaction quotation of the matching target transaction party for the electronic component and the expected price corresponding to the transaction matching request; determining the price-sensitive edge weights from each source point to the intermediate node in the transaction weighted directed graph according to the promised delivery cycle of the matching target transaction party for the electronic component, the historical overdue days corresponding to the matching target transaction party, and the expected delivery time corresponding to the transaction matching request. According to the delivery cycle, the weight of the delivery-sensitive edge from each source point to the intermediate node in the transaction weighted directed graph is determined; according to the product of the smaller value of the sales volume and the matching request volume corresponding to each matching target transaction party, the capacity of the edge from the source point to the intermediate node corresponding to each matching target transaction party in the transaction weighted directed graph is determined; according to the matching request volume carried in the transaction matching request, the capacity of the edge from the sink point to the intermediate node in the transaction weighted directed graph is determined, and according to the capacity of the edge from the source point to the intermediate node corresponding to each matching target transaction party in the transaction weighted directed graph and the capacity of the edge from the sink point to the intermediate node, the transaction weighted directed graph is constructed; According to the capacity of the edges in each path from each source point in the transaction weighted directed graph through the intermediate node to the sink point, by finding the maximum tradable amount of the augmented path, determining multiple matching paths for the matching target transaction party to trade with the transaction request party through the intermediate node; Based on the maximum tradable quantity in the multiple matching paths, an intelligent dynamic transaction matching result for the transaction matching request is determined, and in response to a selection and browsing operation on the displayed intelligent dynamic transaction matching result, introduction information on electronic components of the matching target transaction party corresponding to the selected intelligent dynamic transaction matching result is displayed.

2. The method according to claim 1, characterized in that The step of determining the price-sensitive edge weights from each source point to the intermediate node in the transaction weighted directed graph according to the transaction quotation of the matching target transaction party for the electronic component and the expected price of the transaction matching request includes: Determining a transaction quotation diameter according to a first quotation difference between a highest transaction quotation and a lowest transaction quotation among transaction quotations of the matching target transaction party for the electronic component; Determine an average transaction quotation according to the transaction quotation of each matching target transaction party for the electronic component and the time offset between each transaction quotation and the current time; Determining a transaction quotation offset corresponding to each of the matching target transaction parties according to an absolute value of a second quotation difference between a transaction quotation of each of the matching target transaction parties for the electronic component and the average transaction quotation; The price-sensitive edge weights from each source point to the intermediate node in the transaction weighted directed graph are determined based on the ratio of the transaction quotation offset corresponding to each of the matched target transaction parties to the transaction quotation diameter, and the product of the weight attenuation coefficient corresponding to the preset deviation threshold range in which the ratio is located, wherein each of the preset deviation threshold ranges is provided with a one-to-one corresponding weight attenuation coefficient.

3. The method according to claim 2, characterized in that The determining an average transaction quotation according to the transaction quotation of each matching target transaction party for the electronic component and the time offset between each transaction quotation and the current time includes: Determine the adaptive volatility according to the preset volatility and the mean and standard deviation of the transaction quotations corresponding to each of the matching target transaction parties within a preset historical time period from the current time; Determine the attenuation weight corresponding to each of the matching target trading parties according to the time offset between the transaction quote corresponding to each of the matching target trading parties and the current time and the adaptive volatility; Determine the attenuated transaction quotation corresponding to each of the matching target transaction parties according to the product of the attenuation weight corresponding to each of the matching target transaction parties and the corresponding transaction quotation of the electronic component; According to the attenuated transaction quotations corresponding to the matched target transaction parties, a sum of attenuated transaction quotations is determined, and according to the attenuation weights corresponding to the matched target transaction parties, an attenuation coefficient is determined by summing up; The average transaction quotation is determined according to the ratio of the sum of the attenuated transaction quotations to the attenuation coefficient.

4. The method according to claim 1, characterized in that The step of determining the delivery-time-sensitive edge weights from each source point to the intermediate node in the transaction weighted directed graph according to the promised delivery cycle of the matching target transaction party for the electronic component, the number of historical overdue days corresponding to the matching target transaction party, and the expected delivery cycle corresponding to the transaction matching request includes: Determine the fulfillment decay rate corresponding to each of the matching target transaction parties according to the number of historical overdue days corresponding to each of the matching target transaction parties; According to the fulfillment decay rate corresponding to each of the matching target transaction parties and the cycle difference between the promised delivery cycle of the electronic components by each of the matching target transaction parties and the expected delivery cycle corresponding to the transaction matching request, the delivery-sensitive edge weights from each source point to the intermediate node in the transaction weighted directed graph are determined.

5. The method according to claim 1, characterized in that The determining of multiple matching paths for the matching target transaction party to trade with the transaction request party via the intermediate node by finding the maximum tradable amount of the augmented path according to the capacity of the edges in each path from each source point via the intermediate node to the sink point in the transaction weighted directed graph includes: Constructing a residual network according to each path from each source point in the transaction weighted directed graph through the intermediate node to the sink point; Determining the residual capacity of all edges in the residual network according to a breadth-first search or a depth-first search; According to the capacity of the edges in each path from each source point through the intermediate node to the sink point in the transaction weighted directed graph, by finding the maximum tradable amount of the augmented path, the bottleneck capacity of a single path that meets the matching request quantity supplied by a matching target transaction party alone, or the bottleneck capacity of a combined path that meets the matching request quantity supplied by multiple matching target transaction parties together, is calculated; According to the bottleneck capacity of each of the single paths and / or each of the combined paths, the residual capacity of all edges in the residual network is updated, and a plurality of the single paths and / or the combined paths cannot be obtained in the residual network when performing augmented path finding; According to the multiple single paths and / or the combined paths, multiple matching paths for the matching target transaction party to trade with the transaction request party via the intermediate node are determined.

6. The method according to any one of claims 1 to 5, characterized in that In response to receiving the transaction matching request, dynamically determining a matching target transaction party for electronic component transaction matching from transaction sellers that sell electronic components corresponding to the electronic component names according to the electronic component names carried in the transaction matching request, including: In response to receiving the transaction matching request, querying, according to the electronic component name carried in the transaction matching request, a transaction seller of the electronic component corresponding to the electronic component name registered with the transaction matching intermediary; Determine the number of times and frequency of historical browsing of each of the transaction sellers in the transaction matching intermediary; According to the number of times and frequency of historical browsing of each transaction seller at the transaction matching intermediary, a matching target transaction party for electronic component transaction matching is dynamically determined from transaction sellers selling electronic components corresponding to the electronic component names.

7. An AI-based intelligent dynamic transaction matching device for electronic components, characterized in that: Applied to a transaction matching intermediary, the device comprises: A first determination module is configured to, in response to receiving a transaction matching request, dynamically determine a matching target transaction party for electronic component transaction matching from transaction sellers that sell electronic components corresponding to the electronic component name according to the electronic component name carried in the transaction matching request; The second determination module is configured to construct a transaction weighted directed graph based on the matching request quantity carried in the transaction matching request and the sales quantity of each of the matching target transaction parties, with the transaction matching intermediary as the intermediate node, the matching target transaction party as the source point, and the transaction matching request as the sink point, including: dynamically determining the credibility coefficient of each of the matching target transaction parties based on the historical fulfillment rate of the sales quantity of each of the matching target transaction parties; determining the price-sensitive edge weights from each of the source points to the intermediate nodes in the transaction weighted directed graph based on the transaction quotation of the matching target transaction party for the electronic components and the expected price corresponding to the transaction matching request; determining the price-sensitive edge weights from each of the source points to the intermediate nodes based on the promised delivery cycle of the matching target transaction party for the electronic components, the historical overdue days corresponding to the matching target transaction party, and the expected price corresponding to the transaction matching request According to the expected delivery cycle, the weight of the delivery-sensitive edge from each source point to the intermediate node in the transaction weighted directed graph is determined; according to the product of the smaller value of the sales volume and the matching request volume corresponding to each matching target transaction party, the capacity of the edge from the source point to the intermediate node corresponding to each matching target transaction party in the transaction weighted directed graph is determined; according to the matching request volume carried in the transaction matching request, the capacity of the edge from the sink point to the intermediate node in the transaction weighted directed graph is determined, and according to the capacity of the edge from the source point to the intermediate node corresponding to each matching target transaction party in the transaction weighted directed graph and the capacity of the edge from the sink point to the intermediate node, the transaction weighted directed graph is constructed; A third determination module is configured to determine, according to the capacity of the edges in each path from each source point in the transaction weighted directed graph through the intermediate node to the sink point, a plurality of matching paths for the matching target transaction party to trade with the transaction request party through the intermediate node by finding the maximum tradable amount of the augmented path; The fourth determination module is configured to determine the intelligent dynamic transaction matching result for the transaction matching request based on the maximum tradable quantity in the multiple matching paths, and in response to a selection and browsing operation on the displayed intelligent dynamic transaction matching result, display the introduction information of the electronic components of the matching target transaction party corresponding to the selected intelligent dynamic transaction matching result.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method described in any one of claims 1 to 6 are implemented.

9. An electronic device, characterized in that: include: a memory having a computer program stored thereon; A processor, configured to execute the computer program in the memory to implement the steps of the method according to any one of claims 1 to 6.

Images