Intelligent matching and supply chain collaboration-based whole-process procurement optimization method and system

Abstract

The application discloses a kind of based on intelligent matching and supply chain cooperation's whole-process procurement optimization method and system, it is related to supply chain management and information technology cross field.The method is first received multiple small agricultural product orders of purchaser, and based on the category of order, product control, address, time and preservation requirement, agricultural product procurement batch task and first section logistics instruction are generated to regional distribution center by intelligent matching algorithm, then execute instruction and collect first section logistics process data, and when agricultural product arrives distribution center, product control data is collected and bound, determine its distribution permission state, then end terminal set single and path optimization are carried out to the agricultural product of permission distribution, execute last section logistics instruction and collect process data, finally aggregate whole-link data for each order, generate unique digital quality record and provide to purchaser, whereby comprehensive procurement cost can be effectively reduced, loss is reduced, supply chain cooperation efficiency and trust degree are improved.

Description

Technical Field

[0001] This invention belongs to the interdisciplinary field of supply chain management and information technology, specifically involving a whole-process procurement optimization method and system based on intelligent matching and supply chain collaboration, which is particularly suitable for the efficient connection between the decentralized upstream production of agricultural products and the massive and scattered downstream demand. Background Technology

[0002] With the development of e-commerce and consumption upgrading, the procurement model for fresh agricultural products is shifting from traditional large-scale wholesale to high-frequency, small-batch digital procurement. Downstream buyers, such as those in the catering or retail sectors, often generate a large number of small, scattered order demands. However, existing technological solutions face the following pressing technical challenges when addressing these demands: (1) The cost and efficiency dilemma caused by small-scale procurement, namely, the small orders of downstream buyers are processed separately, which cannot form economically efficient procurement and transportation batches, resulting in high procurement and logistics costs per unit product, low vehicle loading rate, and serious waste of resources; although some existing order aggregation platforms have tried to aggregate demand, they mostly stay at the level of information display and transaction matching, and have failed to delve into the actual operation of the supply chain, and cannot fundamentally solve the problem of uneconomical scale caused by order fragmentation. (2) Data disconnection and lack of trust in quality control throughout the entire chain. Agricultural products, especially fresh produce, are perishable and non-standardized. Their quality is highly dependent on the temperature and humidity environment and operational timeliness during the logistics process. The various links in the existing system (such as supplier management, logistics tracking and warehouse quality control) often operate independently, forming "data silos". It is difficult for the purchaser to obtain a complete and reliable data flow from the source to the delivery, to verify whether the logistics environment meets the preservation requirements, and to accurately link the final quality control results with specific batches and orders, resulting in frequent quality control disputes and high industry trust costs. (3) The disconnect between static planning and dynamic execution means that traditional procurement and logistics planning is often based on static planning, which is difficult to adapt to the dynamic factors that change in real time in the fresh food supply chain. For example, the delivery addresses of orders are scattered and the time requirements are different, and simple carpooling is difficult to achieve the optimal route. Key information such as the supplier's performance, abnormal environmental data during logistics and quality control results after arrival cannot be fed back in real time and used to adjust subsequent scheduling decisions. This disconnect between planning and execution makes the overall supply chain slow to respond, weak to interference, and difficult to effectively control losses.

[0003] In summary, how to provide an integrated technical solution that can deeply integrate information technology and physical operations, intelligently aggregate scattered orders into economic batches, achieve transparent monitoring and data binding throughout the entire process, and make dynamic optimization decisions based on real-time execution data, so as to systematically solve the efficiency, cost, and trust issues in the scenario of small-scale procurement of agricultural products, is a topic that urgently needs to be studied by those skilled in the art. Summary of the Invention

[0004] The purpose of this invention is to provide a method, system, computer equipment, computer-readable storage product, and computer program product for optimizing the entire procurement process based on intelligent matching and supply chain collaboration, in order to solve the problems of cost and efficiency dilemmas caused by small-scale procurement, data disconnection and lack of trust in quality control throughout the entire chain, and / or the disconnect between static planning and dynamic execution in existing agricultural product procurement schemes.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: Firstly, a full-process procurement optimization method based on intelligent matching and supply chain collaboration is provided, including: Receive multiple small orders from multiple agricultural product purchasers, wherein the small order refers to an agricultural product order in which the amount of a single order is lower than a preset amount threshold or the quantity of goods in a single order is lower than a preset quantity threshold; Based on the product category, quality control conditions, delivery address, delivery time, and freshness requirements of agricultural products in the multiple small orders, an intelligent matching algorithm is used to generate at least one agricultural product procurement batch task that is used to fulfill the multiple small orders and points from the supplier to the regional distribution center, as well as at least one first-segment logistics instruction corresponding to the at least one agricultural product procurement batch task. The agricultural product procurement batch task refers to an independent procurement task that purchases the agricultural product from a single supplier in a procurement batch for a single category of agricultural products. For each task in the at least one agricultural product procurement batch task, the corresponding first-stage logistics instruction is executed, and then the first-stage logistics process data of the corresponding agricultural product from the corresponding supplier to the regional distribution center is collected during the instruction execution process, and the first-stage logistics process data is bound to the corresponding task. For each of the multiple small orders, when the agricultural product corresponding to a certain batch of agricultural product procurement task used to fulfill the corresponding order is delivered to the regional distribution center, the quality control data collected by the quality control terminal for the agricultural product is received, the quality control data is bound to the corresponding order, and the delivery permission status of the agricultural product is determined based on the quality control data and the quality control conditions in the corresponding order. For agricultural products with a delivery permit status, the system performs order aggregation and route optimization based on the delivery address and delivery time in the small order used to purchase the corresponding agricultural products, generates and executes the corresponding last-mile logistics instructions, and then collects the last-mile logistics process data of the corresponding agricultural products from the regional distribution center to the delivery address during the instruction execution process, and binds the last-mile logistics process data to the small order. For each order, a unique digital quality history is generated based on the first-stage logistics process data bound to the agricultural product procurement task used to fulfill the corresponding order, as well as the quality control data and last-stage logistics process data bound to the corresponding order. This digital quality history is then provided to the agricultural product procurement party that initiated the corresponding order.

[0006] Based on the above-mentioned invention, a new integrated agricultural product procurement solution that deeply integrates information technology and physical operations is provided. This solution first receives small-amount agricultural product orders from multiple buyers. Based on the order's category, quality control, address, time, and preservation requirements, an intelligent matching algorithm generates batch procurement tasks and initial logistics instructions for agricultural products directed to a regional distribution center. Then, the instructions are executed, and initial logistics process data is collected. Upon delivery of the agricultural products to the distribution center, quality control data is collected and bound to determine their delivery permit status. Next, for permitted delivery agricultural products, final order aggregation and route optimization are performed, final logistics instructions are executed, and process data is collected. Finally, all-link data for each order is aggregated to generate a unique digital quality history, which is provided to the buyer. This enables intelligent order aggregation and precise matching, visualized monitoring of the entire logistics process, and data binding of quality control information, ultimately forming traceable digital deliverables. This effectively reduces overall procurement costs, minimizes losses, and improves supply chain collaboration efficiency and trust, facilitating practical application and promotion.

[0007] In one possible design, based on the product category, quality control conditions, delivery address, delivery time, and agricultural product freshness requirements in the multiple small orders, an intelligent matching algorithm generates at least one agricultural product procurement batch task from the supplier to the regional distribution center to fulfill the multiple small orders, and at least one first-stage logistics instruction corresponding to the at least one agricultural product procurement batch task, including: Based on the product category, quality control conditions, delivery address, and delivery time of the multiple small orders, a clustering algorithm is used to aggregate the multiple small orders with the same product category and similar quality control conditions, delivery address, and delivery time to form at least one order set; For each order set in the at least one order set, based on the production capacity, historical performance rating and logistics cost of multiple agricultural product suppliers, and on the premise of meeting the freshness time limit requirements of agricultural products for all small orders in the corresponding set, at least one optimal supplier is determined from the multiple agricultural product suppliers for the corresponding set based on a multi-objective optimization algorithm with the objectives of minimizing the overall procurement cost, maximizing the supplier performance rating and / or minimizing the procurement loss rate. For each of the at least one optimal suppliers, a batch of agricultural product procurement tasks and the first logistics instruction are generated from the corresponding supplier to the regional distribution center, wherein the procurement demand of a set of orders is fulfilled by one or more of the batch of agricultural product procurement tasks.

[0008] In one possible design, based on the product category, quality control conditions, delivery address, and delivery time of the multiple small orders, a clustering algorithm is used to aggregate the multiple small orders with the same product category and similar quality control conditions, delivery address, and delivery time to form at least one order set, including: Based on the product categories in the multiple small orders, the multiple small orders are grouped by product category to obtain at least one order group; For each order group in the at least one order group, based on the quality control conditions, delivery address, and delivery time of all small orders in the corresponding group, a spatiotemporal clustering algorithm is used to aggregate all small orders in the corresponding group that have similar quality control conditions, delivery address, and delivery time to form at least one order set. The spatiotemporal clustering algorithm first maps the quality control conditions and the delivery address to coordinates in a multidimensional feature space, converts the delivery time into a time window, and then uses the proximity of the feature coordinates and the similarity of the time window as the core metrics for clustering.

[0009] In one possible design, the multi-objective optimization algorithm is implemented by constructing and solving a multi-objective optimization model, wherein the objective function of the multi-objective optimization model includes at least the following components (A) to (C): (A) Minimize the overall procurement cost, which is calculated by taking into account the procurement cost, the transportation cost from the supplier to the regional distribution center, and the loss cost estimated based on the agricultural product category and logistics time. (B) Maximize supplier performance score, which is calculated based on a weighted score of the supplier’s historical order on-time rate, quality control compliance rate and customer complaint rate; (C) Minimize the procurement loss rate, which is calculated based on the basic loss rate of agricultural product categories and the estimated additional loss rate due to the discrepancy between logistics environment data and shelf life requirements.

[0010] In one possible design, the method further includes: For any smallest SKU of inventory within the regional distribution center, historical sales data, seasonal factors, shelf life, historical loss rate data, and actual inbound / outbound rhythm data of the corresponding agricultural products are obtained. Based on the obtained data, a time-series forecast model is used to obtain the sales forecast value of the corresponding agricultural products in a specified future period. The historical loss rate data and the actual inbound / outbound rhythm data are extracted from the digital quality history, respectively. Based on the sales forecast, the preset service level target, and the procurement lead time, calculate the safety stock and reorder point for any minimum inventory unit SKU, and determine the optimal inventory level for any minimum inventory unit SKU based on the calculation results. The system monitors the current available inventory of any minimum inventory unit SKU in real time, and when the current available inventory is lower than the reorder point, it generates and executes a replenishment plan for any minimum inventory unit SKU, which includes the replenishment quantity and expected delivery time, based on the optimal inventory level and the sales forecast.

[0011] In one possible design, the method further includes: Periodically extract data reflecting actual performance from the generated digital quality history, including actual logistics timeliness, logistics environment compliance rate, quality control compliance rate, and goods loss rate. The extracted data reflecting actual performance is fed back into the decision model used for the intelligent matching algorithm, dynamically updating the cost model parameters, supplier performance score, and / or the calculation weights of the supplier performance score in the decision model.

[0012] Secondly, a full-process procurement optimization system based on intelligent matching and supply chain collaboration is provided, including a small order receiving unit, a task instruction generation unit, a first-stage logistics processing unit, a quality control and delivery determination unit, a last-stage logistics processing unit, and a quality history generation unit. The small order receiving unit is used to receive multiple small orders from multiple agricultural product purchasers, wherein the small order refers to an agricultural product order in which the amount of a single order is lower than a preset amount threshold or the quantity of goods in a single order is lower than a preset quantity threshold. The task instruction generation unit is communicatively connected to the small order receiving unit. It is used to generate, based on the category, quality control conditions, delivery address, delivery time, and agricultural product freshness time limit requirements in the multiple small orders, at least one agricultural product procurement batch task from the supplier to the regional distribution center and at least one first-segment logistics instruction corresponding to the at least one agricultural product procurement batch task through an intelligent matching algorithm. The agricultural product procurement batch task refers to an independent procurement task for purchasing agricultural products from a single supplier in a procurement batch for a single category of agricultural products. The first-stage logistics processing unit is communicatively connected to the task instruction generation unit. It is used to execute the corresponding first-stage logistics instruction for each task in the at least one agricultural product procurement batch task, and then collect the first-stage logistics process data of the corresponding agricultural product from the corresponding supplier to the regional distribution center during the instruction execution process, and bind the first-stage logistics process data to the corresponding task. The quality control and delivery determination unit is communicatively connected to the small order receiving unit and the task instruction generation unit, respectively. It is used to receive quality control data collected by the quality control terminal for each of the multiple small orders when the agricultural product corresponding to a certain batch of agricultural product procurement task used to fulfill the corresponding order is delivered to the regional distribution center, bind the quality control data with the corresponding order, and determine the delivery permission status of the agricultural product based on the quality control data and the quality control conditions in the corresponding order. The last-stage logistics processing unit is communicatively connected to the quality control and delivery determination unit. It is used to perform order aggregation and route optimization for agricultural products with a delivery permit status, based on the delivery address and delivery time in the small order used to purchase the corresponding agricultural products, generate and execute the corresponding last-stage logistics instructions, and then collect the last-stage logistics process data of the corresponding agricultural products from the regional distribution center to the delivery address during the instruction execution process, and bind the last-stage logistics process data to the small order. The quality history generation unit is communicatively connected to the first-stage logistics processing unit, the quality control and delivery determination unit, and the last-stage logistics processing unit. It is used to generate a unique digital quality history for each order based on the first-stage logistics process data bound to the agricultural product procurement task used to fulfill the corresponding order, as well as the quality control data and last-stage logistics process data bound to the corresponding order. The digital quality history is then provided to the agricultural product purchaser who initiates the corresponding order.

[0013] Thirdly, the present invention provides a computer device comprising a storage module, a processing module, and a transceiver module connected in sequence for communication, wherein the storage module is used to store a computer program, the transceiver module is used to send and receive messages, and the processing module is used to read the computer program and execute the end-to-end procurement optimization method as described in the first aspect or any possible design in the first aspect.

[0014] Fourthly, the present invention provides a computer-readable storage product storing instructions that, when executed on a computer, perform the end-to-end procurement optimization method as described in the first aspect or any possible design within the first aspect.

[0015] Fifthly, the present invention provides a computer program product, including a computer program or instructions, wherein when the computer program or instructions are executed by a computer, they implement the end-to-end procurement optimization method as described in the first aspect or any possible design in the first aspect.

[0016] The beneficial effects of the above scheme are: (1) This invention creatively provides a new integrated agricultural product procurement solution that can deeply integrate information technology and physical operation. First, it receives small agricultural product orders from multiple buyers. Based on the order's category, quality control, address, time, and preservation requirements, it generates agricultural product procurement batch tasks and first-stage logistics instructions pointing to the regional distribution center through an intelligent matching algorithm. Then, it executes the instructions and collects the first-stage logistics process data. When the agricultural products arrive at the distribution center, it collects and binds quality control data to determine their delivery permit status. Then, it performs end-of-line order aggregation and route optimization for the permitted agricultural products, executes the end-of-line logistics instructions and collects process data. Finally, it aggregates the full-link data for each order, generates a unique digital quality history and provides it to the buyer. This enables intelligent aggregation and accurate matching of orders, full-process visual monitoring of logistics, and data binding of quality control. Ultimately, it forms traceable digital deliverables, thereby effectively reducing overall procurement costs, reducing losses, and improving supply chain collaboration efficiency and trust. (2) It can realize full-process data-driven quality traceability and build a supply chain trust foundation. That is, by accurately binding the order, procurement task, and the data of the first and last logistics processes with the quality control data, a unique "digital quality history" is generated. This method establishes a transparent chain from farm to store. This makes the source of agricultural products, transportation environment and quality inspection results of each order verifiable, fundamentally solving the problem of trust loss caused by information opacity in the fresh food sector. (3) Through intelligent aggregation and optimization decision-making, the problem of uneconomical scale of fragmented procurement can be solved. Specifically, for small orders, the method adopts a two-stage intelligent matching algorithm of "clustering + multi-objective optimization": First, fragmented orders are aggregated into standardized "order sets" according to multiple dimensions such as category, quality control, time and space. Then, each set is treated as a whole for supplier selection. This transforms fragmented demand into economic procurement batches from the source, effectively reducing the procurement and logistics costs per unit product. Furthermore, by optimizing the historical performance rating of suppliers, the stability of supply is guaranteed. (4) It can realize fine-grained collaborative scheduling under multiple constraints, improve the certainty and efficiency of fulfillment. That is, when scheduling, the algorithm does not only consider cost or distance. For example, the spatiotemporal clustering algorithm considers quality control conditions (to ensure consistent quality), delivery address (to ensure geographical proximity) and delivery time (to ensure time compatibility) at the same time, ensuring that orders in the same batch are highly coordinated in terms of quality control, region and time. This lays the foundation for efficient last-mile delivery and greatly improves the feasibility and execution efficiency of the fulfillment plan. (5) It can accurately quantify and control the unique losses of agricultural products, and achieve real comprehensive cost optimization. The innovation of the multi-objective model lies in the design of a special optimization target for the characteristics of agricultural products. It incorporates the estimated loss cost into the total cost calculation and sets an independent target to minimize the procurement loss rate. The calculation of this target combines the basic loss rate of the category and the additional loss due to the non-compliance of the logistics environment. This makes the cost optimization model more in line with the actual fresh food business and can actively select the supply chain route with lower loss through the algorithm to achieve real cost reduction and efficiency improvement. (6) The traditional static, experience-driven inventory management model can be transformed into a dynamic, data-driven intelligent early warning and automatic replenishment model. That is, by integrating real business data such as loss and rhythm in the history, the sales forecast can be more in line with reality, thereby calculating the accurate dynamic inventory level. Through real-time monitoring and automatic triggering, the transformation from "manual discovery and handling when inventory is insufficient" to "automatic generation and execution of replenishment plan by the system before inventory reaches the critical point" has been realized, which significantly reduces the risk of stockouts and backlogs, reduces manual dependence, and improves the overall inventory turnover efficiency and supply chain response speed. (7) It can also achieve continuous learning and optimization based on real business feedback, so that intelligent matching decisions can continuously adapt to changes in actual operation, become more intelligent the more it is used, and facilitate practical application and promotion. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating the end-to-end procurement optimization method based on intelligent matching and supply chain collaboration provided in this application embodiment.

[0019] Figure 2 This is a schematic diagram of the structure of the end-to-end procurement optimization system based on intelligent matching and supply chain collaboration provided in the embodiments of this application.

[0020] Figure 3A schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in conjunction with the accompanying drawings and descriptions of the embodiments or the prior art. Obviously, the following description of the structure of the accompanying drawings is only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these embodiments without creative effort. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0022] It should be understood that although the terms "first" and "second", etc., may be used herein to describe various objects, these objects should not be limited by these terms. These terms are only used to distinguish one object from another. For example, the first object may be referred to as the second object, and similarly, the second object may be referred to as the first object, without departing from the scope of the exemplary embodiments of the invention.

[0023] It should be understood that the term "and / or" that may appear in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, or A and B exist simultaneously. Another example is A, B and / or C, which can mean that any one of A, B, and C or any combination thereof exists. The term " / and" that may appear in this document describes another relationship between related objects, indicating that two relationships can exist. For example, A / and B can mean: A exists alone or A and B exist simultaneously. In addition, the character " / " that may appear in this document generally indicates that the related objects before and after it are in an "or" relationship.

[0024] Example like Figure 1 As shown, the end-to-end procurement optimization method based on intelligent matching and supply chain collaboration provided in the first aspect of this embodiment can be executed, but is not limited to, by a management server with certain computing resources, such as by an agricultural product procurement management platform server. Figure 1 As shown, the whole-process procurement optimization method includes, but is not limited to, the following steps S1 to S6.

[0025] S1. Receive multiple small orders from multiple agricultural product purchasers, wherein the small order refers to an agricultural product order whose single order amount is lower than a preset amount threshold or whose single order quantity is lower than a preset quantity threshold.

[0026] In step S1, the specific methods for receiving small orders include, but are not limited to: receiving electronic purchase orders submitted by multiple downstream small and medium-sized purchasers (e.g., community fresh food supermarkets, small and medium-sized catering enterprises, or unit canteens) through platform clients (such as web pages or mobile applications); different small orders can come from different agricultural product purchasers or from the same agricultural product purchaser. The small orders need to contain structured purchase demand information, and specific fields include, but are not limited to: category (e.g., "Shandong Red Fuji apples"), quality control conditions (e.g., "sugar content ≥12°Brix, no mechanical damage"), delivery address (accurate to the store or warehouse address), delivery time (e.g., "9:00-12:00 the next day"), and freshness time limit requirements based on category characteristics (e.g., "from picking to delivery no more than 72 hours"). The procurement management platform can preset thresholds for monetary value (e.g., a single order amount less than 1,000 RMB) and quantity (e.g., a single order weight less than 50 kg) to automatically determine whether an order meets either condition. Orders meeting either condition are then marked as small-value orders and included in subsequent batch aggregation processes. Furthermore, all received orders are stored in the platform's database in real time, forming a set of orders awaiting processing.

[0027] S2. Based on the product category, quality control conditions, delivery address, delivery time, and agricultural product freshness time limit requirements in the multiple small orders, generate at least one agricultural product procurement batch task for fulfilling the multiple small orders and pointing from the supplier to the regional distribution center, and at least one first-segment logistics instruction corresponding to the at least one agricultural product procurement batch task, wherein the agricultural product procurement batch task refers to an independent procurement task for purchasing agricultural products from a single supplier in a procurement batch for a single category of agricultural products.

[0028] In step S2, the purpose of this step is to transform the scattered and heterogeneous procurement needs of the downstream into efficient and standardized supply tasks and logistics plans that can be executed by the upstream through computational intelligence. Specifically, based on the categories, quality control conditions, delivery addresses, delivery times, and freshness requirements of agricultural products in the multiple small orders, an intelligent matching algorithm is used to generate at least one agricultural product procurement batch task that is used to fulfill the multiple small orders and points from the supplier to the regional distribution center, as well as at least one first-stage logistics instruction corresponding to the at least one agricultural product procurement batch task, including but not limited to the following steps S21 to S23.

[0029] S21. Based on the product category, quality control conditions, delivery address, and delivery time of the multiple small orders, a clustering algorithm is used to aggregate the multiple small orders with the same product category and similar quality control conditions, delivery address, and delivery time to form at least one order set.

[0030] In step S21, the purpose of this step is to intelligently group a large number of scattered orders according to a collaboratively executable standard. Specifically, based on the category, quality control conditions, delivery address, and delivery time of the multiple small orders, a clustering algorithm is used to aggregate the multiple small orders with the same category and similar quality control conditions, delivery address, and delivery time to form at least one order set, including but not limited to the following steps S211 to S212.

[0031] S211. Based on the product categories in the plurality of small orders, group the plurality of small orders by product category to obtain at least one order group.

[0032] In step S211, all small orders are quickly grouped according to the core attribute of "category" (for example, all "tomato" orders and all "broccoli" orders are grouped into different groups) in order to significantly reduce the data heterogeneity within each cluster unit, and ensure that subsequent algorithms focus on optimizing quality control and spatiotemporal features under the same category dimension, thereby improving clustering efficiency and the business rationality of the results.

[0033] S212. For each order group in the at least one order group, based on the quality control conditions, delivery address, and delivery time of all small orders in the corresponding group, a spatiotemporal clustering algorithm is used to aggregate all small orders in the corresponding group that have similar quality control conditions, delivery address, and delivery time to form at least one order set. The spatiotemporal clustering algorithm first maps the quality control conditions and the delivery address to coordinates in a multidimensional feature space, converts the delivery time into a time window, and then uses the proximity of the feature coordinates and the similarity of the time window as the core measure for clustering.

[0034] In step S212, specifically for orders within the same category of orders, a spatiotemporal clustering algorithm is initiated. This algorithm transforms three key features of each order into computable model inputs: (1) Feature space mapping, which encodes unstructured quality control conditions (such as "organic certification" or "premium fruit") into numerical labels, and together with the geographical coordinates (latitude and longitude) of the delivery address, forms a multidimensional feature vector as the basis for spatial clustering; (2) Time window conversion, which converts ambiguous delivery time periods (such as "morning") into comparable standardized time window objects (such as timestamp range). Subsequently, the spatiotemporal clustering algorithm uses the spatial distance of feature vectors (representing quality control consistency and geographical proximity) and the degree of overlap of time windows as the core measurement standards, dynamically aggregating orders with similar features and compatible time windows into the final executable order set, thereby ensuring that orders within the same set are highly coordinated in terms of quality control standards, delivery areas, and time requirements. This is the key to achieving end-of-line order co-delivery and reducing fulfillment costs.

[0035] Therefore, based on the above steps S211 to S212, orders with the same product category and similar quality control requirements, delivery geographical areas, and expected delivery time windows can be aggregated into an order set. This organizes the originally disordered scattered demands into several "demand packages" with unified internal characteristics that can be processed in batches. This lays the data structure foundation for subsequent unified supplier sourcing and logistics planning for the set, and is a key pre-processing step for achieving economies of scale.

[0036] S22. For each order set in the at least one order set, based on the production capacity, historical performance rating and logistics cost of multiple agricultural product suppliers, and on the premise of meeting the freshness requirements of agricultural products for all small orders in the corresponding set, at least one optimal supplier is determined from the multiple agricultural product suppliers for the corresponding set based on a multi-objective optimization algorithm with the objectives of minimizing the overall procurement cost, maximizing the supplier performance rating and / or minimizing the procurement loss rate.

[0037] In step S22, the purpose of this step is not simply to pursue the lowest price, but to achieve a multi-dimensional balance decision of cost, quality and reliability under the rigid constraint of ensuring the freshness of agricultural products. In detail, the multi-objective optimization algorithm is implemented by constructing and solving a multi-objective optimization model, wherein the optimization objective function of the multi-objective optimization model includes at least the following components (A) to (C).

[0038] (A) Minimize the overall procurement cost, which is calculated by considering the procurement cost, transportation cost from the supplier to the regional distribution center, and estimated loss cost based on the agricultural product category and logistics time. This objective aims to achieve the optimal total cost, rather than the lowest single cost. Its calculation needs to comprehensively consider: procurement cost – calculated based on supplier quotations and procurement quantity (unit price × procurement quantity); transportation cost – calculated based on transportation distance, cargo weight / volume, and unit freight rate (freight rate × distance × cargo quantity) model; estimated loss cost (to reflect the characteristics of agricultural products, an estimated loss cost based on category and experience data is introduced), which is calculated as (procurement price + transportation cost) × estimated loss rate. The estimated loss rate can be estimated by looking up a preset loss rate-time comparison table or a simple linear model based on the inherent perishability of the agricultural product category and the planned logistics time.

[0039] (B) Maximize the supplier performance score, which is calculated based on a weighted score of the supplier's historical order on-time delivery rate, quality control compliance rate, and customer complaint rate. This objective aims to select the best long-term, reliable partners. The supplier performance score is a quantifiable comprehensive performance indicator, calculated by weighting its historical performance: Score = w1 × Historical Order On-Time Delivery Rate + w2 × Historical Quality Control Compliance Rate - w3 × Historical Customer Complaint Rate, where w1, w2, and w3 are preset positive weighting coefficients (e.g., w1=0.5, w2=0.3, w3=0.2). The historical order on-time delivery rate, historical quality control compliance rate, and historical customer complaint rate are statistically derived from all past transaction records of the supplier in the platform database. By maximizing this weighted score, the optimization algorithm can favor suppliers with more punctual delivery, more stable quality, and better service reputation.

[0040] (C) Minimize the procurement loss rate, which is calculated based on the basic loss rate of agricultural product categories and the estimated additional loss rate due to discrepancies between logistics environment data and shelf-life requirements. This objective aims to control loss risk at the source, and its calculation is broken down into two parts: (1) the basic loss rate, which is an inherent baseline value set according to the agricultural product category (such as leafy vegetables or berries); (2) the additional estimated loss rate, a dynamic adjustment item, that is, the system will evaluate the logistics route and duration assigned in the plan. If it is judged from historical data that the logistics environment data (such as average temperature) may not be able to fully meet the shelf-life requirements of the order, an additional estimated loss rate will be added according to the degree of difference. Finally, the procurement loss rate = basic loss rate + additional estimated loss rate; minimizing this objective can drive the optimization algorithm to prioritize suppliers and routes that can provide better shelf-life logistics conditions.

[0041] In step S22, the production capacity, historical performance rating and logistics cost of each agricultural product supplier can be collected in the following manner: (1) Supplier production capacity: The supplier first updates the available supply of each category of agricultural products regularly (e.g. daily) through its platform management backend. This data is stored in the platform supplier database in real time, so that the production capacity data can be obtained by accessing the database; (2) Historical performance rating: Based on the data of all completed orders of the supplier in the past, the supplier automatically calculates its historical order on-time rate (number of on-time delivered orders divided by the total number of orders), historical quality control compliance rate (number of quality inspection qualified orders divided by the total number of orders) and customer complaint rate (number of complaint orders divided by the total number of orders), and calculates the quantified historical performance rating according to the preset formula (e.g., rating = 0.5 × on-time rate + 0.3 × compliance rate - 0.2 × complaint rate), and updates it regularly; (3) Logistics cost: The integrated map and logistics service provider API is called to estimate the unit transportation cost based on the coordinates of the supplier's warehouse and regional distribution center, the weight and volume of the goods and the current market freight rate. To achieve efficient solution, a linear weighted sum method can be used to transform the aforementioned multiple objectives into a single comprehensive objective, specifically: Comprehensive objective value = α × (normalized value of comprehensive procurement cost) - β × (normalized value of supplier performance score) + γ × (normalized value of estimated procurement loss rate), where α, β, and γ are preset positive weighting coefficients (e.g., α = 0.6, β = 0.3, γ = 0.1), used to adjust the relative importance of cost, reliability, and loss in the decision-making process (maximizing the supplier performance score is transformed into a minimization problem by adding a negative sign); solving for the minimum value of this comprehensive objective function yields the Pareto optimal solution that balances the demands of multiple parties. Furthermore, in the optimization process, for example, but not limited to, using a genetic algorithm to solve the above model, the steps are briefly described as follows: (1) Encoding: Encode a possible supplier allocation scheme as a "chromosome", where the gene represents the supplier number assigned to each order set; (2) Initialization: Randomly generate a set of initial schemes to form an initial population; (3) Evaluation: Calculate the comprehensive target value of each scheme in the population as its fitness; (4) Selection, crossover, and mutation: Select excellent individuals based on fitness, generate new schemes through crossover operations, and introduce new features with a small probability of mutation to generate the next generation population; (5) Iteration: Repeat the evaluation and evolution steps until the convergence condition is met (such as reaching the preset number of iterations), and finally output the scheme with the highest fitness, which is the determined optimal supplier set. Through the aforementioned implementation method, it can be ensured that the algorithm can efficiently and automatically obtain high-quality decisions under complex constraints.

[0042] S23. For each of the at least one optimal suppliers, generate a batch task for the agricultural product procurement and the first logistics instruction from the corresponding supplier to the regional distribution center, wherein the procurement demand of a set of orders is fulfilled by one or more batch tasks for the agricultural product procurement.

[0043] In step S23, a specific, executable physical operation instruction package can be generated for each supplier. The details include, but are not limited to: (1) generating a procurement batch task, that is, creating a unique task number, clearly specifying the supplier, the single agricultural product category to be procured (such as "Hebei production area A-grade Crown Pear"), and the precise procurement quantity (the total demand of the order set under the responsibility of the supplier); (2) generating the first logistics instruction, that is, associating the above task, generating an instruction containing the pickup address (supplier warehouse), destination (regional distribution center), latest loading time, suggested carrier, and route (this instruction will be issued to the corresponding supplier and logistics service provider). At the same time, let me briefly explain the "one-to-many" fulfillment relationship between the order set (which can be understood as a procurement batch) and the task: the procurement demand of an order set may be allocated by the system to multiple optimal suppliers to fulfill due to insufficient capacity of a single supplier or for risk diversification considerations. Therefore, multiple parallel procurement batch tasks and logistics instructions will be generated to ensure that the demand is fully met. In addition, all generated tasks and instructions are precisely associated with the source order set within the platform.

[0044] S3. For each task in the at least one agricultural product procurement batch task, execute the corresponding first-stage logistics instruction, and then collect the first-stage logistics process data of the corresponding agricultural product from the corresponding supplier to the regional distribution center during the instruction execution process, and bind the first-stage logistics process data to the corresponding task.

[0045] In step S3, the aim is to transform the digital instructions generated at the planning layer (i.e., the first logistics instruction) into physical transportation operations and achieve full-process transparent monitoring. Specifically, the generated initial logistics instruction (which includes information such as pick-up address, destination distribution center, license plate number, and planned time window) can be synchronized to the corresponding carrier's dispatch system and the driver's mobile terminal via an interface. This allows the supplier to prepare goods according to the instruction requirements, and the carrier to arrange vehicles to perform trunk transportation tasks from the supplier to the designated regional distribution center. During transportation, IoT devices installed on the transport vehicles can automatically and continuously collect the initial logistics process data (which mainly includes: spatiotemporal trajectory data, i.e., geographical location, driving speed, and timestamp sequence collected by the vehicle's GPS device; environmental indicator data, i.e., time-series records of temperature and relative humidity inside the cargo compartment collected by temperature and humidity sensors; and key node event data, i.e., operation timestamps and operator information recorded at loading and unloading points through QR code / RFID scanning), and transmit it back to the platform in real time or near real time via the mobile network. After receiving the aforementioned streaming data, the platform server automatically associates all process data and stores it in the database record of the agricultural product procurement batch task corresponding to the number, based on the unique transportation task number carried in the data stream (this number is created when the instruction is generated). This enables the full-dimensional process data generated by each physical transportation operation to be uniquely and accurately bound to its planned tasks in the digital world, providing a complete data chain foundation for subsequent traceability, analysis, and history generation.

[0046] S4. For each of the multiple small orders, when the agricultural product corresponding to a certain agricultural product procurement batch task used to fulfill the corresponding order is delivered to the regional distribution center, the quality control data collected by the quality control terminal for the agricultural product is received, and the quality control data is bound to the corresponding order. Based on the quality control data and the quality control conditions in the corresponding order, the delivery permission status of the agricultural product is determined.

[0047] In step S4, this step is performed at the receiving stage of the regional distribution center. It aims to conduct objective quality inspection of the delivered agricultural products and make an automated decision on whether to allow them to enter the next distribution stage. Specifically, when the transport vehicle carrying goods with a specific batch of agricultural products arrives at the platform of the regional distribution center, the staff uses a quality control terminal (such as a dedicated PDA or a tablet computer with a specific application installed) to inspect the goods and obtain quality control data. The quality control data can be collected and uploaded to the platform in real time through at least one of the following methods: (1) Image recognition, that is, taking multi-angle photos of agricultural products, and automatically analyzing their color, size and surface defects by an integrated AI model, and outputting the specification grade and defect detection results; (2) Sensor detection, that is, using a handheld saccharimeter or a rapid pesticide residue detector to conduct sampling inspection, and the data is automatically synchronized to the quality control terminal via Bluetooth; (3) Manual entry, that is, checking or filling in the observation results (such as packaging integrity or abnormal odor) on the terminal form. Upon receiving the quality control data, the platform server automatically and uniquely binds it to the original purchase order associated with the batch barcode scanned by the quality control terminal. Immediately after binding, it performs automated judgment: comparing the collected quality control data (e.g., "sweetness value: 13.5°Brix") with the quality control conditions agreed upon in the corresponding order (e.g., "sweetness ≥ 12°Brix"). If all key indicators meet or exceed the order requirements, the status of the batch of agricultural products is automatically marked as "permitted" (i.e., delivery allowed). If any key indicator fails to meet the requirements, it is marked as "rejected," triggering an exception handling process (e.g., return, exchange, or downgrade). This step transforms subjective quality control experience into objective data-driven decisions, ensuring that only compliant products can proceed to subsequent stages and precisely linking quality responsibility with specific orders.

[0048] S5. For agricultural products with a delivery permit status of "permitted", perform order aggregation and route optimization based on the delivery address and delivery time in the small order used to purchase the corresponding agricultural products, generate and execute the corresponding last-mile logistics instructions, and then collect the last-mile logistics process data of the corresponding agricultural products from the regional distribution center to the delivery address during the instruction execution process, and bind the last-mile logistics process data with the small order.

[0049] In step S5, the "last mile" intelligent delivery of quality-controlled agricultural products from the regional distribution center to the final customer's address is carried out. Specifically, the original orders associated with all agricultural products in the "permitted" status are dynamically aggregated according to their delivery address and delivery time. For example, all orders that need to be delivered within a 3-kilometer radius of "Zhongguancun Street, Haidian District, Beijing" between 9:00 and 12:00 the next day are grouped into one delivery batch. Subsequently, based on the geographic information system, a vehicle route planning algorithm considering time window constraints is used to calculate the optimal delivery route, vehicle allocation, and parking sequence for this batch, with the goal of minimizing the total mileage, maximizing the vehicle loading rate, and ensuring that the delivery time of each order is met. After optimization, detailed last-mile logistics instructions are automatically generated based on the optimization results. These instructions include, but are not limited to: delivery personnel information, a vehicle cargo manifest (corresponding to specific orders), a planned delivery route sequence, the estimated arrival time at each stop, and the corresponding order information. These instructions are issued via the delivery personnel's app, allowing them to load the goods at the distribution center according to the manifest and then execute the delivery task following the app's navigation instructions. Finally, during delivery, the delivery personnel's app, in conjunction with the vehicle's IoT (Internet of Things) devices, automatically collects last-mile logistics data. This data includes, but is not limited to: real-time location tracking based on the phone's GPS (Global Positioning System), the signing time and geostamp at each customer point via the app scanning the order QR code, and, if necessary, temperature and humidity data of the final transportation environment recorded by portable devices. After all data is uploaded in real time, the platform automatically and accurately binds the entire last-mile trajectory and related event data to the specific small-value order using the unique order code included in the scanning action, completing the data loop for the delivery process.

[0050] S6. For each order, based on the first segment logistics process data bound to the agricultural product procurement task used to fulfill the corresponding order, as well as the quality control data and last segment logistics process data bound to the corresponding order, a unique digital quality history is generated, and the digital quality history is provided to the agricultural product procurement party that initiates the corresponding order.

[0051] In step S6, this step is the value delivery stage, which aims to aggregate and encapsulate the discrete data generated in all previous steps and precisely bound to the order into a complete and tamper-proof "digital product" for delivery to the purchaser. Specifically, when an order completes its final delivery and is signed for, the history generation program is automatically triggered. Based on the unique identifier of the order, the program associates and extracts three core data from the database: (1) supply and trunk logistics data, that is, the first logistics process data associated with the agricultural product procurement batch task that fulfills the order (such as supplier information, departure time, and the temperature and humidity curve of the trunk transportation); (2) quality verification data, that is, the quality control data directly bound to the order (such as test reports or photos of quality inspection results); (3) final delivery data, that is, the final logistics process data bound to the order (such as departure time from the distribution center, delivery trajectory, signing time, and location snapshot). Subsequently, these multi-source, heterogeneous but logically related data are serialized chronologically and encapsulated into a digital quality history file with a specific data structure (such as JSON format) (this history has a globally unique identifier). The generated history file is stored on the platform server and simultaneously provided to the agricultural product purchaser who initiated the order through two main methods: (a) API interface push, i.e., the history data is pushed to the purchaser's internal management system in real time through the standard API interface integrated into the purchaser's system; (b) platform visualization query, i.e., a visual history viewing interface is provided on the purchaser's platform order details page, which intuitively displays key data and events throughout the entire process in the form of a timeline or chart. In addition, the digital quality history constitutes a complete and credible digital evidence chain for this procurement activity, which can be used by the purchaser for purposes such as quality traceability, financial reconciliation, and supplier performance evaluation.

[0052] Therefore, based on the end-to-end procurement optimization method described in steps S1 to S6 above, a new integrated end-to-end agricultural product procurement solution is provided that can deeply integrate information technology and physical operations. This solution first receives small-amount agricultural product orders from multiple buyers. Based on the order's category, quality control, address, time, and preservation requirements, an intelligent matching algorithm generates batch procurement tasks and initial logistics instructions for agricultural products directed to a regional distribution center. Then, the instructions are executed, and initial logistics process data is collected. Upon delivery of the agricultural products to the distribution center, quality control data is collected and bound to determine their delivery permit status. Next, for permitted delivery agricultural products, final order aggregation and route optimization are performed, final logistics instructions are executed, and process data is collected. Finally, end-to-end data is aggregated for each order, generating a unique digital quality history and providing it to the buyer. This enables intelligent order aggregation and precise matching, full-process visual monitoring of logistics, and data-driven binding of quality control, ultimately forming traceable digital deliverables. This effectively reduces overall procurement costs, minimizes losses, and improves supply chain collaboration efficiency and trust, facilitating practical application and promotion.

[0053] Based on the technical solution of the first aspect mentioned above, this embodiment also provides a possible design for how to drive dynamic prediction and automatic replenishment of inventory based on end-to-end data, that is, the method also includes, but is not limited to, the following steps S71 to S73.

[0054] S71. For any smallest stock keeping unit (SKU) within the regional distribution center, acquire historical sales data, seasonal factors, shelf life, historical loss rate data, and actual inbound / outbound rhythm data for the corresponding agricultural product category. Based on the acquired data, obtain the sales forecast value for the corresponding agricultural product category in a specified future period through a time series prediction model. The historical loss rate data and the actual inbound / outbound rhythm data are extracted from the digital quality history, respectively.

[0055] In step S71, this step aims to provide accurate demand forecasts for inventory decisions. Specifically, for each SKU (e.g., "300g boxed Shandong cherry tomatoes") in the distribution center, modeling data is obtained from multiple data sources: (1) the internal business system obtains its historical sales data, seasonal factors (e.g., holiday markers), and shelf life; (2) the average historical loss rate related to the SKU is extracted from the historical digital quality history database, and inbound and outbound rhythm data such as the average daily actual outbound volume are obtained based on the timestamp analysis in the history. Subsequently, these cleaned time-series data are input into the time-series forecasting model (e.g., using an ARIMA model that considers seasonal decomposition or an LSTM neural network model) to output the daily sales forecast value of the SKU in the next replenishment cycle (e.g., the next 7 days).

[0056] S72. Based on the sales forecast, the preset service level target, and the procurement lead time, calculate the safety stock and reorder point for any minimum inventory unit SKU, and determine the optimal inventory level for any minimum inventory unit SKU based on the calculation results.

[0057] In step S72, the predicted value is transformed into an executable inventory strategy. Specifically, based on the sales forecast, combined with a preset service level target (e.g., requiring 99% demand availability) and procurement lead time (days from order placement to goods receipt), a quantitative calculation is performed using an inventory model. Specifically, the safety stock is typically calculated based on the forecast error (e.g., standard deviation) within the lead time and the service level factor (e.g., Z-score). The reorder point is equal to the sum of the predicted sales within the procurement lead time plus the safety stock. The optimal inventory level (i.e., the target maximum inventory) is typically set as the reorder point plus an Economic Order Quantity (EOQ) or the predicted sales within a fixed coverage period. The final calculation result provides dynamic inventory control parameters for the SKU.

[0058] S73. Monitor the current available inventory of any minimum inventory unit SKU in real time, and when the current available inventory is lower than the reorder point, generate a replenishment plan for any minimum inventory unit SKU, including the replenishment quantity and expected delivery time, based on the optimal inventory level and the sales forecast, and execute it.

[0059] In step S73, this step achieves an automated closed-loop inventory management system. Specifically, it monitors the current available inventory (i.e., physical inventory minus reserved but not yet picked-up quantity) of each SKU in real time. When the available inventory of a certain SKU is detected to have dropped to its reorder point, the replenishment plan generation process is automatically triggered: the suggested replenishment quantity is calculated based on the difference between the optimal inventory level and the current inventory, combined with sales forecasts for fine-tuning; simultaneously, the expected delivery time is calculated backward based on the supplier's procurement lead time; subsequently, a structured replenishment instruction is automatically generated and sent directly to the procurement execution system or the task list of relevant personnel through an interface, driving the procurement task to be initiated, thereby achieving automatic early warning of insufficient inventory risk and rapid replenishment response.

[0060] Based on the aforementioned possible design, the traditional static, experience-driven inventory management model can be transformed into a dynamic, data-driven intelligent early warning and automatic replenishment model. This involves integrating real business data such as losses and pace from the inventory history to make sales forecasts more realistic, thereby calculating accurate dynamic inventory levels. Through real-time monitoring and automatic triggering, the system shifts from "manually discovering and handling inventory shortages" to "automatically generating and executing replenishment plans before inventory reaches a critical point." This significantly reduces the risk of stockouts and overstocking, reduces reliance on manual labor, and improves overall inventory turnover efficiency and supply chain responsiveness.

[0061] Based on the technical solution of the first aspect mentioned above, this embodiment also provides a possible design two for how to realize the data intelligent closed loop of "evaluation-feedback-optimization", that is, the method also includes, but is not limited to, the following steps S81 to S82.

[0062] S81. Periodically extract data reflecting actual performance from the generated digital quality history, including actual logistics timeliness, logistics environment compliance rate, quality control compliance rate, and goods loss rate.

[0063] In step S81, this step aims to quantify historical performance results into analyzable performance indicators. Specifically, the generated and stored digital quality history database will be scanned periodically (e.g., weekly or monthly), and structured data used to evaluate the overall performance over the past period will be extracted in batches through preset statistical queries. The specific indicators extracted and calculated include, but are not limited to: (1) actual logistics timeliness, i.e., calculating the actual average time from supplier delivery to final receipt for each order and comparing it with the planned time; (2) logistics environment compliance rate, i.e., the proportion of time points during transportation when the temperature and humidity data recorded by sensors meet the preset freshness range for the agricultural product category; (3) quality control compliance rate, i.e., the proportion of batches that are judged as "permitted" among the batches inspected at the distribution center; (4) goods loss rate, i.e., the ratio of actual goods loss to goods loss based on the shipment volume and final receipt volume in the history; etc.

[0064] S82. The extracted data reflecting the actual performance of the contract is fed back into the decision model used for the intelligent matching algorithm, and the cost model parameters, supplier performance score and / or the calculation weight of the supplier performance score in the decision model are dynamically updated.

[0065] In step S82, this step is to transform the performance data into the optimization basis of the algorithm model. Specifically, the aggregated performance data extracted in step S81 is used as a feedback signal to be input into the underlying decision model on which the intelligent matching algorithm depends, triggering the self-updating of the model parameters: (1) Update the cost model parameters, for example, take the actual average loss rate of the current period as the new prediction benchmark, adjust the loss cost calculation coefficient in the cost model, so that the future cost prediction is more in line with reality; (2) Update the supplier performance score, that is, take the current period's quality control compliance rate and environmental compliance rate as new historical performance data, iteratively update the long-term performance score of the corresponding supplier, so that the score can reflect its latest performance in real time; (3) Update the score calculation weight, that is, automatically analyze the correlation between each performance indicator (such as timeliness, loss) and overall satisfaction or cost, dynamically adjust the weight coefficients of each sub-indicator (such as on-time rate or compliance rate) in the supplier performance score formula, so that the score model is more scientific.

[0066] Based on the above-mentioned second possible design, continuous learning and optimization based on real business feedback can be achieved, enabling intelligent matching decisions to continuously adapt to changes in actual operating conditions and become more intelligent with use.

[0067] like Figure 2 As shown, the second aspect of this embodiment provides a virtual system for implementing the full-process procurement optimization method described in the first aspect or any possible design in the first aspect, including a small order receiving unit, a task instruction generation unit, a first-stage logistics processing unit, a quality control and delivery determination unit, a last-stage logistics processing unit, and a quality history generation unit. The small order receiving unit is used to receive multiple small orders from multiple agricultural product purchasers, wherein the small order refers to an agricultural product order in which the amount of a single order is lower than a preset amount threshold or the quantity of goods in a single order is lower than a preset quantity threshold. The task instruction generation unit is communicatively connected to the small order receiving unit. It is used to generate, based on the category, quality control conditions, delivery address, delivery time, and agricultural product freshness time limit requirements in the multiple small orders, at least one agricultural product procurement batch task from the supplier to the regional distribution center and at least one first-segment logistics instruction corresponding to the at least one agricultural product procurement batch task through an intelligent matching algorithm. The agricultural product procurement batch task refers to an independent procurement task for purchasing agricultural products from a single supplier in a procurement batch for a single category of agricultural products. The first-stage logistics processing unit is communicatively connected to the task instruction generation unit. It is used to execute the corresponding first-stage logistics instruction for each task in the at least one agricultural product procurement batch task, and then collect the first-stage logistics process data of the corresponding agricultural product from the corresponding supplier to the regional distribution center during the instruction execution process, and bind the first-stage logistics process data to the corresponding task. The quality control and delivery determination unit is communicatively connected to the small order receiving unit and the task instruction generation unit, respectively. It is used to receive quality control data collected by the quality control terminal for each of the multiple small orders when the agricultural product corresponding to a certain batch of agricultural product procurement task used to fulfill the corresponding order is delivered to the regional distribution center, bind the quality control data with the corresponding order, and determine the delivery permission status of the agricultural product based on the quality control data and the quality control conditions in the corresponding order. The last-stage logistics processing unit is communicatively connected to the quality control and delivery determination unit. It is used to perform order aggregation and route optimization for agricultural products with a delivery permit status, based on the delivery address and delivery time in the small order used to purchase the corresponding agricultural products, generate and execute the corresponding last-stage logistics instructions, and then collect the last-stage logistics process data of the corresponding agricultural products from the regional distribution center to the delivery address during the instruction execution process, and bind the last-stage logistics process data to the small order. The quality history generation unit is communicatively connected to the first-stage logistics processing unit, the quality control and delivery determination unit, and the last-stage logistics processing unit. It is used to generate a unique digital quality history for each order based on the first-stage logistics process data bound to the agricultural product procurement task used to fulfill the corresponding order, as well as the quality control data and last-stage logistics process data bound to the corresponding order. The digital quality history is then provided to the agricultural product purchaser who initiates the corresponding order.

[0068] The working process, working details and technical effects of the aforementioned system provided in the second aspect of this embodiment can be found in the first aspect or any possible design of the full-process procurement optimization method described in the first aspect, and will not be repeated here.

[0069] like Figure 3 As shown, the third aspect of this embodiment provides a computer device for executing the end-to-end procurement optimization method as described in the first aspect or any possible design within the first aspect. The device includes a storage module, a processing module, and a transceiver module connected in sequence. The storage module stores a computer program, the transceiver module sends and receives messages, and the processing module reads the computer program and executes the end-to-end procurement optimization method as described in the first aspect or any possible design within the first aspect. Specifically, the storage module may include, but is not limited to, random-access memory (RAM), read-only memory (ROM), flash memory, first-in-first-out (FIFO) memory, and / or first-in-last-out (FILO) memory, etc.; the processing module may, but is not limited to, use a microprocessor of the STM32F105 series. Furthermore, the computer device may also include, but is not limited to, a power supply module, a display screen, and other necessary components.

[0070] The working process, working details and technical effects of the aforementioned computer equipment provided in the third aspect of this embodiment can be found in the first aspect or any possible design in the first aspect of the whole process procurement optimization method, and will not be repeated here.

[0071] This fourth aspect of the embodiment provides a computer-readable storage product that stores instructions comprising the end-to-end procurement optimization method as described in the first aspect or any possible design within the first aspect. Specifically, the computer-readable storage product stores instructions that, when executed on a computer, perform the end-to-end procurement optimization method as described in the first aspect or any possible design within the first aspect. The computer-readable storage product refers to a data storage medium, which may include, but is not limited to, computer-readable storage media such as floppy disks, optical disks, hard disks, flash memory, USB flash drives, and / or Memory Sticks. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0072] The working process, working details and technical effects of the aforementioned computer-readable storage product provided in the fourth aspect of this embodiment can be found in the full-process procurement optimization method described in the first aspect or any possible design in the first aspect, and will not be repeated here.

[0073] This fifth aspect of the embodiment provides a computer program product, including a computer program or instructions, which, when executed by a computer, implement the end-to-end procurement optimization method as described in the first aspect or any possible design within the first aspect. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0074] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for optimizing the entire procurement process based on intelligent matching and supply chain collaboration, characterized in that, include: Receive multiple small orders from multiple agricultural product purchasers, wherein the small order refers to an agricultural product order in which the amount of a single order is lower than a preset amount threshold or the quantity of goods in a single order is lower than a preset quantity threshold; Based on the product category, quality control conditions, delivery address, delivery time, and freshness requirements of agricultural products in the multiple small orders, an intelligent matching algorithm is used to generate at least one agricultural product procurement batch task that is used to fulfill the multiple small orders and points from the supplier to the regional distribution center, as well as at least one first-segment logistics instruction corresponding to the at least one agricultural product procurement batch task. The agricultural product procurement batch task refers to an independent procurement task that purchases the agricultural product from a single supplier in a procurement batch for a single category of agricultural products. For each task in the at least one agricultural product procurement batch task, the corresponding first-stage logistics instruction is executed, and then the first-stage logistics process data of the corresponding agricultural product from the corresponding supplier to the regional distribution center is collected during the instruction execution process, and the first-stage logistics process data is bound to the corresponding task. For each of the multiple small orders, when the agricultural product corresponding to a certain batch of agricultural product procurement task used to fulfill the corresponding order is delivered to the regional distribution center, the quality control data collected by the quality control terminal for the agricultural product is received, the quality control data is bound to the corresponding order, and the delivery permission status of the agricultural product is determined based on the quality control data and the quality control conditions in the corresponding order. For agricultural products with a delivery permit status, the system performs order aggregation and route optimization based on the delivery address and delivery time in the small order used to purchase the corresponding agricultural products, generates and executes the corresponding last-mile logistics instructions, and then collects the last-mile logistics process data of the corresponding agricultural products from the regional distribution center to the delivery address during the instruction execution process, and binds the last-mile logistics process data to the small order. For each order, a unique digital quality history is generated based on the first-stage logistics process data bound to the agricultural product procurement task used to fulfill the corresponding order, as well as the quality control data and last-stage logistics process data bound to the corresponding order. This digital quality history is then provided to the agricultural product procurement party that initiated the corresponding order.

2. The end-to-end procurement optimization method according to claim 1, characterized in that, Based on the product categories, quality control conditions, delivery addresses, delivery times, and freshness requirements of agricultural products in the multiple small orders, an intelligent matching algorithm generates at least one agricultural product procurement batch task that fulfills the multiple small orders and points from the supplier to the regional distribution center, along with at least one first-stage logistics instruction corresponding to each of the at least one agricultural product procurement batch task, including: Based on the product category, quality control conditions, delivery address, and delivery time of the multiple small orders, a clustering algorithm is used to aggregate the multiple small orders with the same product category and similar quality control conditions, delivery address, and delivery time to form at least one order set; For each order set in the at least one order set, based on the production capacity, historical performance rating and logistics cost of multiple agricultural product suppliers, and on the premise of meeting the freshness time limit requirements of agricultural products for all small orders in the corresponding set, at least one optimal supplier is determined from the multiple agricultural product suppliers for the corresponding set based on a multi-objective optimization algorithm with the objectives of minimizing the overall procurement cost, maximizing the supplier performance rating and / or minimizing the procurement loss rate. For each of the at least one optimal suppliers, a batch of agricultural product procurement tasks and the first logistics instruction are generated from the corresponding supplier to the regional distribution center, wherein the procurement demand of a set of orders is fulfilled by one or more of the batch of agricultural product procurement tasks.

3. The end-to-end procurement optimization method according to claim 2, characterized in that, Based on the product category, quality control conditions, delivery address, and delivery time of the multiple small orders, a clustering algorithm is used to aggregate the multiple small orders with the same product category and similar quality control conditions, delivery address, and delivery time to form at least one order set, including: Based on the product categories in the multiple small orders, the multiple small orders are grouped by product category to obtain at least one order group; For each order group in the at least one order group, based on the quality control conditions, delivery address, and delivery time of all small orders in the corresponding group, a spatiotemporal clustering algorithm is used to aggregate all small orders in the corresponding group that have similar quality control conditions, delivery address, and delivery time to form at least one order set. The spatiotemporal clustering algorithm first maps the quality control conditions and the delivery address to coordinates in a multidimensional feature space, converts the delivery time into a time window, and then uses the proximity of the feature coordinates and the similarity of the time window as the core metrics for clustering.

4. The end-to-end procurement optimization method according to claim 2, characterized in that, The multi-objective optimization algorithm is implemented by constructing and solving a multi-objective optimization model, wherein the objective function of the multi-objective optimization model includes at least the following components (A) to (C): (A) Minimize the overall procurement cost, which is calculated by taking into account the procurement cost, the transportation cost from the supplier to the regional distribution center, and the loss cost estimated based on the agricultural product category and logistics time. (B) Maximize supplier performance score, which is calculated based on a weighted score of the supplier’s historical order on-time rate, quality control compliance rate and customer complaint rate; (C) Minimize the procurement loss rate, which is calculated based on the basic loss rate of agricultural product categories and the estimated additional loss rate due to the discrepancy between logistics environment data and shelf life requirements.

5. The end-to-end procurement optimization method according to claim 1, characterized in that, The method further includes: For any smallest SKU of inventory within the regional distribution center, historical sales data, seasonal factors, shelf life, historical loss rate data, and actual inbound / outbound rhythm data of the corresponding agricultural products are obtained. Based on the obtained data, a time-series forecast model is used to obtain the sales forecast value of the corresponding agricultural products in a specified future period. The historical loss rate data and the actual inbound / outbound rhythm data are extracted from the digital quality history, respectively. Based on the sales forecast, the preset service level target, and the procurement lead time, calculate the safety stock and reorder point for any minimum inventory unit SKU, and determine the optimal inventory level for any minimum inventory unit SKU based on the calculation results. The system monitors the current available inventory of any minimum inventory unit SKU in real time, and when the current available inventory is lower than the reorder point, it generates and executes a replenishment plan for any minimum inventory unit SKU, which includes the replenishment quantity and expected delivery time, based on the optimal inventory level and the sales forecast.

6. The end-to-end procurement optimization method according to claim 1, characterized in that, The method further includes: Periodically extract data reflecting actual performance from the generated digital quality history, including actual logistics timeliness, logistics environment compliance rate, quality control compliance rate, and goods loss rate. The extracted data reflecting actual performance is fed back into the decision model used for the intelligent matching algorithm, dynamically updating the cost model parameters, supplier performance score, and / or the calculation weights of the supplier performance score in the decision model.

7. A full-process procurement optimization system based on intelligent matching and supply chain collaboration, characterized in that, It includes a small order receiving unit, a task instruction generation unit, a first-stage logistics processing unit, a quality control and delivery confirmation unit, a last-stage logistics processing unit, and a quality history generation unit; The small order receiving unit is used to receive multiple small orders from multiple agricultural product purchasers, wherein the small order refers to an agricultural product order in which the amount of a single order is lower than a preset amount threshold or the quantity of goods in a single order is lower than a preset quantity threshold. The task instruction generation unit is communicatively connected to the small order receiving unit. It is used to generate, based on the category, quality control conditions, delivery address, delivery time, and agricultural product freshness time limit requirements in the multiple small orders, at least one agricultural product procurement batch task from the supplier to the regional distribution center and at least one first-segment logistics instruction corresponding to the at least one agricultural product procurement batch task through an intelligent matching algorithm. The agricultural product procurement batch task refers to an independent procurement task for purchasing agricultural products from a single supplier in a procurement batch for a single category of agricultural products. The first-stage logistics processing unit is communicatively connected to the task instruction generation unit. It is used to execute the corresponding first-stage logistics instruction for each task in the at least one agricultural product procurement batch task, and then collect the first-stage logistics process data of the corresponding agricultural product from the corresponding supplier to the regional distribution center during the instruction execution process, and bind the first-stage logistics process data to the corresponding task. The quality control and delivery determination unit is communicatively connected to the small order receiving unit and the task instruction generation unit, respectively. It is used to receive quality control data collected by the quality control terminal for each of the multiple small orders when the agricultural product corresponding to a certain batch of agricultural product procurement task used to fulfill the corresponding order is delivered to the regional distribution center, bind the quality control data with the corresponding order, and determine the delivery permission status of the agricultural product based on the quality control data and the quality control conditions in the corresponding order. The last-stage logistics processing unit is communicatively connected to the quality control and delivery determination unit. It is used to perform order aggregation and route optimization for agricultural products with a delivery permit status, based on the delivery address and delivery time in the small order used to purchase the corresponding agricultural products, generate and execute the corresponding last-stage logistics instructions, and then collect the last-stage logistics process data of the corresponding agricultural products from the regional distribution center to the delivery address during the instruction execution process, and bind the last-stage logistics process data to the small order. The quality history generation unit is communicatively connected to the first-stage logistics processing unit, the quality control and delivery determination unit, and the last-stage logistics processing unit. It is used to generate a unique digital quality history for each order based on the first-stage logistics process data bound to the agricultural product procurement task used to fulfill the corresponding order, as well as the quality control data and last-stage logistics process data bound to the corresponding order. The digital quality history is then provided to the agricultural product purchaser who initiates the corresponding order.

8. A computer device, characterized in that, It includes a storage module, a processing module, and a transceiver module that are sequentially connected in communication. The storage module is used to store computer programs, the transceiver module is used to send and receive messages, and the processing module is used to read the computer programs and execute the end-to-end procurement optimization method as described in any one of claims 1 to 6.

9. A computer-readable storage product, characterized in that... The computer-readable storage product stores instructions that, when executed on a computer, perform the end-to-end procurement optimization method as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or the instructions are executed by the computer, they implement the end-to-end procurement optimization method as described in any one of claims 1 to 6.

Images