Intelligent distribution management optimization method and system for agricultural product supply chain

By optimizing transportation routes and packing strategies in the agricultural product supply chain through the Internet of Things and machine learning, and by assessing losses in real time and making closed-loop adjustments, the problems of freshness and cost control in agricultural product distribution have been solved, and efficient and flexible smart distribution management has been achieved.

CN121094671APending Publication Date: 2025-12-09QINGYUAN POLYTECHNIC
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202511304165.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-09-12
Publication Date
2025-12-09

AI Technical Summary

Technical Problem

The existing agricultural product distribution routes are limited and lack monitoring, resulting in reduced freshness of agricultural products, high transportation costs, and difficulty in effectively predicting and controlling losses.

Method used

By collecting multi-source heterogeneous data in real time through IoT devices, generating datasets based on dynamic desensitization and encryption rules, and combining machine learning algorithms to optimize transportation routes and packing strategies, real-time assessment of losses and closed-loop strategy adjustments are made to achieve intelligent delivery management.

Benefits of technology

It has enabled end-to-end intelligent distribution management of the agricultural product supply chain, reducing transportation costs and losses, improving distribution efficiency and customer satisfaction, and ensuring the freshness and quality of agricultural products.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121094671A_ABST
    Figure CN121094671A_ABST
Patent Text Reader

Abstract

The invention relates to an intelligent distribution management optimization method for an agricultural product supply chain, and the method comprises the steps: collecting multi-source heterogeneous data of production, processing and transportation ends in real time through the Internet of Things, and uploading the data to a cloud end after dynamic grading desensitization and encryption to form a complete traceability record; automatically generating an initial transportation and profit scheme according to a customer order, and dynamically optimizing a route and assembly by using a loss model driven by real-time environment / road condition data; when the actual loss exceeds the threshold value, the transportation strategy is adjusted immediately, integration of data safety, path optimization and loss closed-loop control is achieved, the freshness of agricultural product supply is improved, meanwhile, the transportation cost is reduced, and the performance efficiency and the customer experience are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the technical field of intelligent distribution, and in particular to a method and system for optimizing intelligent distribution management in the agricultural product supply chain. Background Technology

[0002] In recent years, my country's agricultural product trading channels have diversified, with a surge in agricultural product transactions through fresh food e-commerce, fresh food chains, and community-supported agriculture. Consequently, the demand for agricultural product logistics has also increased dramatically, and customers have higher and higher requirements for the quality of agricultural product delivery. Therefore, how to optimize agricultural product distribution management, reduce agricultural product losses, improve the quality of agricultural product delivery, and reduce agricultural product logistics costs are urgent problems to be solved in the current operation of the agricultural product supply chain.

[0003] The delivery process from supplier to customer is one of the major factors contributing to the high transportation costs of agricultural products and affecting the freshness of fresh agricultural products. Due to unreasonable vehicle transportation arrangements, the perishable nature of fresh agricultural products, and the difficulty of mixing batches of products with different freshness levels for delivery, there are huge losses in the process of delivering agricultural products from suppliers to retail stores.

[0004] Based on the existing technology, the following technical problems exist: the existing agricultural product distribution routes are limited, and the lack of monitoring of agricultural products results in delayed intervention in distribution, thereby reducing the freshness of agricultural product supply. Therefore, improvements are needed. Summary of the Invention

[0005] To improve the freshness of agricultural product supply, this application provides a method and system for optimizing intelligent distribution management in the agricultural product supply chain.

[0006] Firstly, the above-mentioned inventive objective of this application is achieved through the following technical solution:

[0007] A method for optimizing intelligent distribution management in an agricultural product supply chain, the method comprising the following steps:

[0008] The multi-source heterogeneous data from the production, processing and transportation ends is collected in real time by IoT devices, and the multi-source heterogeneous data is classified and graded based on pre-set dynamic desensitization rules to generate a multi-source heterogeneous dataset.

[0009] The data in the multi-source heterogeneous dataset is encrypted based on the preset encryption rules, and the processed data is uploaded to the cloud server to generate a traceability record.

[0010] Acquire customer order data and analyze the customer order data to generate corresponding initial transportation and delivery data and initial profit expectation data, wherein the initial transportation and delivery data includes initial agricultural product assembly demand data and initial transportation route data;

[0011] The initial transportation and distribution data is optimized based on the actual selected priority to generate optimized transportation and distribution data, wherein the optimized transportation and distribution data includes optimized agricultural product assembly demand data and optimized transportation route data;

[0012] The system acquires real-time environmental data and real-time road condition data corresponding to the optimized transportation route data, and calculates the actual transportation loss based on the real-time environmental data and real-time road condition data. The actual transportation loss includes agricultural product loss and transportation vehicle loss.

[0013] The actual transportation loss is compared with a preset loss threshold. If the loss exceeds the threshold, the transportation strategy is adjusted.

[0014] By adopting the above technical solutions, a complete chain of IoT real-time data collection, dynamic desensitization, cloud encryption, intelligent scheduling, real-time loss assessment, and closed-loop strategy correction is achieved, realizing end-to-end intelligent distribution management of the agricultural product supply chain. Existing route optimization only considers static distance or cost. This application incorporates real-time congestion index, environmental parameters, and agricultural product physiological models into dynamic loss calculation, achieving a three-dimensional balance of "timeliness-loss-cost". Traditional threshold alarms are post-event remedies. This application, through a closed-loop strategy of prediction-marking-simulation-correction, intervenes before cargo damage occurs, truly achieving "predictive scheduling". This flexibly and effectively controls distribution costs and adjusts allocation according to actual transportation and customer needs, improving customer satisfaction.

[0015] In a preferred embodiment, this application can be further configured as follows: after acquiring customer order data and analyzing the customer order data to generate corresponding initial transportation and delivery data and initial profit expectation data, the following steps are included:

[0016] Associate historical customer order data with a pre-defined public timeline and construct a historical order dataset;

[0017] The pre-set order demand prediction model analyzes the historical order dataset based on a machine learning algorithm to generate predicted agricultural product demand data, which includes predicted agricultural product type data, agricultural product quantity data, and agricultural product destination data corresponding to the agricultural product;

[0018] Based on the predicted agricultural product demand data, agricultural product warehousing and allocation data are generated.

[0019] By adopting the above technical solutions, we can achieve "procurement based on demand and warehousing based on procurement", reduce emergency allocation and transportation costs, reduce the capital tied up in working capital, reduce the delivery distance of agricultural products when customer orders are received, improve delivery efficiency, and reduce the risk of agricultural product spoilage.

[0020] In a preferred embodiment, this application can be further configured as follows: after the step of acquiring real-time environmental data and real-time traffic data corresponding to the optimized transportation route data, and calculating the actual transportation loss based on the real-time environmental data and real-time traffic data, the application includes the following steps:

[0021] The historical actual transportation losses are associated with the corresponding agricultural product classification data, and an agricultural product transportation loss dataset is constructed.

[0022] The pre-set agricultural product transportation loss analysis model analyzes the agricultural product transportation loss dataset based on a machine self-learning algorithm to generate agricultural product transportation loss factors.

[0023] By adopting the above technical solutions, experience-based losses can be quantified into calculable factors, allowing for advance prediction of category-level losses. This provides a quantitative basis for decision-making regarding dynamic temperature control, adding ice, or line modification, effectively reducing overall cargo loss.

[0024] In a preferred example, this application can be further configured as follows: after the step of analyzing the agricultural product transportation loss dataset based on a machine learning algorithm using a pre-set agricultural product transportation loss analysis model to generate agricultural product transportation loss factors, the following steps are included:

[0025] Based on the predicted agricultural product demand data and the transportation loss factor, the scenario of shipping from each storage point is simulated to generate the predicted transportation loss corresponding to each storage and shipping point.

[0026] Compare the predicted transportation losses for each warehouse and shipping point to generate transportation loss difference data.

[0027] The agricultural product warehousing and allocation data are optimized based on the transportation loss difference data to adjust the warehousing management of agricultural products, so that different storage points can reduce transportation costs while meeting the warehousing needs of agricultural products.

[0028] In a preferred example, this application can be further configured as follows: after the step of analyzing the agricultural product transportation loss dataset based on a machine learning algorithm using a pre-set agricultural product transportation loss analysis model to generate agricultural product transportation loss factors, the application includes the following steps:

[0029] Acquire the caloric data of agricultural products during transportation and associate the caloric data with a pre-set public time axis to construct a transportation caloric change dataset;

[0030] Data on agricultural product packing during transportation is obtained, and the packing data is input into the transportation heat change dataset and associated with the corresponding heat data to construct a packing-heat dataset, wherein the packing data includes the classification, quantity and corresponding placement data of the packed agricultural products;

[0031] The pre-set packing heat change analysis model analyzes the packing-heat dataset based on a machine self-learning algorithm to generate heat change correlation coefficients for different agricultural product categories. These heat change correlation coefficients are used to predict heat changes during packing and placement for different agricultural product categories.

[0032] In a preferred example, this application can be further configured as follows: after the step of analyzing the packing-heat dataset based on a pre-set packing heat change analysis model using a machine self-learning algorithm to generate heat change correlation coefficients corresponding to different agricultural product categories, the following steps are included:

[0033] Obtain actual agricultural product packaging data, and calculate the first predicted heat change data based on the heat change correlation coefficient;

[0034] The real-time environmental data and real-time traffic data are acquired, and the first predicted heat change data is corrected to generate the second predicted heat change data.

[0035] A corresponding temperature control strategy for the container is generated based on the preset heat threshold for agricultural product preservation and the second predicted heat change data.

[0036] In a preferred embodiment, this application can be further configured such that the calculation formula for the first predicted heat change data is: ,in This is the first predicted data for heat change. The correlation coefficient for heat changes corresponding to the packaging location of agricultural products. This is the preset reference temperature change value;

[0037] The formula for calculating the second predicted heat change data is as follows: ,in, The first predicted heat change corresponding to different placement locations (derived from the correlation coefficient of placement location), T: real-time ambient temperature (°C), RH: real-time relative humidity (%), CI: real-time congestion index. , , These are the temperature environment correction coefficient, humidity environment correction coefficient, and congestion correction coefficient, respectively, all calculated from historical data regression. , , This is the preset baseline value.

[0038] Secondly, the above-mentioned inventive objective of this application is achieved through the following technical solutions:

[0039] A smart distribution management optimization device for agricultural product supply chain, the device includes: a multi-source heterogeneous dataset generation unit, used to collect multi-source heterogeneous data from the production end, processing end and transportation end in real time through Internet of Things devices, and classify and grade the multi-source heterogeneous data based on a pre-set dynamic desensitization rule to generate a multi-source heterogeneous dataset;

[0040] The source tracing record generation unit is used to encrypt the data in the multi-source heterogeneous dataset based on preset encryption rules, and upload the processed data to the cloud server to generate source tracing records;

[0041] The customer order data analysis unit is used to acquire customer order data and analyze the customer order data to generate corresponding initial transportation and delivery data and initial profit expectation data, wherein the initial transportation and delivery data includes initial agricultural product assembly demand data and initial transportation route data.

[0042] An optimized transportation and distribution data generation unit is used to optimize the initial transportation and distribution data based on the actually selected priority to generate optimized transportation and distribution data, wherein the optimized transportation and distribution data includes optimized agricultural product assembly demand data and optimized transportation route data;

[0043] The actual transportation loss calculation unit is used to acquire real-time environmental data and real-time road condition data corresponding to the optimized transportation route data, and calculate the actual transportation loss based on the real-time environmental data and real-time road condition data. The actual transportation loss includes agricultural product loss and transportation vehicle loss.

[0044] The transportation strategy adjustment unit is used to compare the actual transportation loss with a preset loss threshold, and adjust the transportation strategy if the loss exceeds the threshold.

[0045] Thirdly, the above-mentioned objectives of this application are achieved through the following technical solutions:

[0046] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described intelligent distribution management optimization method for an agricultural product supply chain.

[0047] Fourthly, the above-mentioned objectives of this application are achieved through the following technical solutions:

[0048] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described intelligent distribution management optimization method for an agricultural product supply chain.

[0049] In summary, this application includes at least one of the following beneficial technical effects:

[0050] 1. The complete chain of IoT real-time data collection, dynamic desensitization, cloud encryption, intelligent scheduling, real-time loss assessment, and closed-loop strategy correction enables end-to-end intelligent distribution management of the agricultural product supply chain. Existing route optimization only considers static distance or cost. This application incorporates real-time congestion index, environmental parameters, and agricultural product physiological models into dynamic loss calculation to achieve a three-dimensional balance of "timeliness-loss-cost". Traditional threshold alarms are post-event remedies. This application uses a closed-loop strategy of prediction-marking-simulation-correction to intervene before cargo loss occurs, truly achieving "predictive scheduling". It flexibly and effectively controls distribution costs and adjusts allocation according to actual transportation and customer needs to improve customer satisfaction.

[0051] 2. Achieve "demand-based procurement and procurement-based warehousing", reduce emergency allocation and transportation costs, reduce the capital tied up in working capital, reduce the delivery distance of agricultural products when customer orders are received, improve delivery efficiency, and reduce the risk of agricultural product spoilage;

[0052] 3. Quantify experience-based losses into calculable factors to predict category-level losses in advance, providing a quantitative basis for dynamic temperature adjustment, adding ice, or line modification, effectively reducing overall cargo loss. Attached Figure Description

[0053] Figure 1 This is a flowchart of an intelligent distribution management optimization method for an agricultural product supply chain according to one embodiment of this application;

[0054] Figure 2 This is a schematic diagram of a smart distribution management optimization device for an agricultural product supply chain according to one embodiment of this application;

[0055] Figure 3 This is a schematic diagram of an electronic device according to an embodiment of this application.

[0056] Icon labels:

[0057] 1. Multi-source heterogeneous dataset generation unit; 2. Traceability record generation unit; 3. Customer order data analysis unit; 4. Optimized transportation and delivery data generation unit; 5. Actual transportation loss calculation unit; 6. Transportation strategy adjustment unit. Detailed Implementation

[0058] The present application will be further described in detail below with reference to the accompanying drawings.

[0059] In one embodiment, such as Figure 1 As shown, this application discloses a smart distribution management optimization method for agricultural product supply chains, which specifically includes the following steps:

[0060] S10: Collect multi-source heterogeneous data from the production, processing and transportation ends in real time through IoT devices, and classify and grade the multi-source heterogeneous data based on the preset dynamic desensitization rules to generate a multi-source heterogeneous dataset.

[0061] Specifically, the production end uses soil moisture sensors (to monitor humidity and pH value) and UAV multispectral imaging (to capture leaf disease index); the processing end uses intelligent sorting line weight sensors (to record sorting loss rate) and machine vision systems (to identify fruit damage); and the transportation end uses refrigerated truck temperature and humidity sensors and triaxial accelerometers (to monitor the intensity of sudden braking / bumping). In the dynamic desensitization rules of this application embodiment, farmers only open the yield forecast value (e.g., "expected harvest tomorrow 800kg±5%)", processing enterprises open the inventory status (e.g., "cold storage remaining volume 30%)", and logistics enterprises open the vehicle location (latitude and longitude drift ±100 meters). In other embodiments, dynamic desensitization can be performed by "setting crop batch ID as sensitive level, geographic location as internal level, and driver identity as public level", forming a multi-source heterogeneous dataset with hierarchical labels.

[0062] S20: Encrypt the data in the multi-source heterogeneous dataset based on the preset encryption rules, and upload the processed data to the cloud server to generate a traceability record;

[0063] Specifically, in this embodiment, AES-256 is used to encrypt the "sensitive" fields, SM4 is used to encrypt the "internal" fields, and only hash verification is performed on the "public" fields. After encryption, the data is uploaded to the cloud server via the MQTT-TLS channel, and a triple of <batch ID, hash, timestamp> is written to the blockchain sidechain to generate an immutable traceability record.

[0064] S30: Obtain customer order data and analyze the customer order data to generate corresponding initial transportation and delivery data and initial profit expectation data;

[0065] The initial transportation and distribution data includes initial agricultural product assembly demand data and initial transportation route data;

[0066] Specifically, a customer places an order via their app for "500 boxes of organic broccoli + 200 boxes of bell peppers, requiring delivery to a market in Qingyuan within 48 hours." After parsing the order, the system generates initial transportation and delivery data: Assembly requirements = "300 boxes of broccoli from Base 1 + 200 boxes of broccoli from Base 2 + 200 boxes of bell peppers from Base 3", Initial route = "Base 1 - Base 2 - Base 3 - Qingyuan warehouse - a market in Qingyuan", Initial profit expectation = Order amount - Estimated cost.

[0067] Automatically translate order language into executable logistics tasks, reducing manual scheduling time and lowering the mismatch rate.

[0068] S40: Optimize the initial transportation and delivery data based on the actual selected priority to generate optimized transportation and delivery data;

[0069] The optimized transportation and distribution data includes optimized agricultural product assembly demand data and optimized transportation route data. Specifically, when the system detects a "typhoon warning" causing highway closure, the priority is switched to "timeliness priority." The algorithm is replanned: the assembly demand is adjusted to "500 boxes of broccoli from Base 1 + 200 boxes of bell peppers from Base 3", and the route is changed to "Base 1 - Qingyuan warehouse (air transport) - a market in Qingyuan", shortening the expected transportation time from 48 hours to 36 hours.

[0070] This step ensures timely delivery even in extreme weather scenarios, reducing customer complaints and breach of contract claims.

[0071] S50: Obtain real-time environmental data and real-time road condition data corresponding to the optimized transportation route data, and calculate the actual transportation loss based on the real-time environmental data and real-time road condition data. The actual transportation loss includes agricultural product loss and transportation vehicle loss.

[0072] Specifically, a vehicle carrying 700 boxes of vegetables passed through the high-temperature zone of South China. The system obtained real-time data on the road surface temperature (38°C), humidity (85%), and congestion index (2.1). Based on a broccoli respiration heat model, the algorithm calculated that the vegetable spoilage rate would be approximately 7% after 24 hours, with an additional diesel consumption of about 15 liters.

[0073] It should be noted that the Traffic Congestion Index (TCI) is a composite indicator that quantifies road traffic efficiency and is used to assess the degree of obstruction on transportation routes in real time. In the context of the agricultural product supply chain, it directly affects vehicle speed, idling time, and the risk of cargo damage. This embodiment adopts the following three-level calculation method:

[0074] Data acquisition layer: The vehicle GPS terminal uploads latitude and longitude and instantaneous speed every minute; at the same time, it calls the map API to obtain the real-time vehicle speed and historical free-flow vehicle speed (V0) of the road segment.

[0075] Exponential calculation layer: TCI = (V0 – V t ) / V0 × 100, where V t : This represents the average speed of vehicles on the current road segment.

[0076] TCI∈[0,100], the larger the TCI value, the more severe the congestion.

[0077] When TCI ≥ 2.1, the average vehicle speed is 46% lower than the free-flow speed, which is considered "moderate congestion".

[0078] Application mapping layer: The TCI value is directly substituted into the agricultural product respiratory heat model: Additional loss rate (%) = 0.8×TCI +0.3×temperature (°C) – 0.2×humidity (%), thereby quantifying the incremental loss of goods caused by congestion and providing a threshold trigger basis for S60 strategy adjustment.

[0079] This step quantifies environmental and road conditions into calculable loss indicators, providing data for subsequent threshold comparisons and strategy adjustments.

[0080] S60: Compare the actual transportation loss with a preset loss threshold; if the loss exceeds the loss threshold, adjust the transportation strategy.

[0081] Specifically, for example, when monitoring shows that the vegetable loss rate is 7%, which is greater than the set threshold of 5%, the system triggers a strategy adjustment: pre-cooling of the refrigerated truck compartment is initiated; a "Huizhou transit cold storage" is added as a buffer en route; and a combination of insulation blankets and ice packs is used for packaging. After the adjustment, the vegetable loss rate drops to 3% within 24 hours, while the overall cost increases by 2%.

[0082] This step achieves automatic closed-loop control when losses exceed the threshold, ensuring the quality of agricultural products while controlling additional costs and enhancing supply chain resilience.

[0083] In summary, through steps S10-S60—real-time IoT data collection, dynamic desensitization, cloud encryption, intelligent scheduling, real-time loss assessment, and closed-loop strategy correction—end-to-end intelligent delivery management of the agricultural product supply chain is achieved. Existing route optimization only considers static distance or cost. This application incorporates real-time congestion index, environmental parameters, and agricultural product physiological models into dynamic loss calculation, achieving a three-dimensional balance of "timeliness-loss-cost." Traditional threshold alarms are reactive measures. This application, through a closed-loop strategy of prediction-marking-simulation-correction, intervenes before cargo damage occurs, truly achieving "predictive scheduling." This flexibly and effectively controls delivery costs and adjusts allocations based on actual transportation and customer needs, improving customer satisfaction.

[0084] After step S30: acquiring customer order data and analyzing the customer order data to generate corresponding initial transportation and delivery data and initial profit expectation data, the following steps are included:

[0085] S301: Associate historical customer order data with a pre-defined common timeline and construct a historical order dataset;

[0086] S302: The pre-set order demand prediction model analyzes the historical order dataset based on a machine learning algorithm to generate predicted agricultural product demand data, which includes predicted agricultural product type data, agricultural product quantity data, and agricultural product destination data corresponding to the agricultural product destination data.

[0087] S303: Generate agricultural product warehousing and allocation data based on the predicted agricultural product demand data.

[0088] For steps S301-S303, this application embodiment continues to use the scenario of "broccoli South China - Qingyuan market 48 h cold chain order" as an example. The system traces back all customer orders in the South China region over the past 12 months, aligns them according to the "common timeline", and generates a structured CSV file containing "order date, agricultural product type, quantity, destination, and actual arrival time" to form a historical order dataset.

[0089] An LSTM neural network was used to train the historical order dataset. The model input was "week, holiday marker, temperature, and historical sales". The output was a prediction for the next 7 days: 500 boxes of broccoli and 200 boxes of bell peppers were needed. The destinations were concentrated in a market in Qingyuan (80%) and a wholesale agricultural market in Shenzhen (20%).

[0090] Based on the forecast results, the system issued an inbound instruction of "500 boxes of broccoli - Qingyuan warehouse" to Base 1 and "200 boxes of bell peppers - Shenzhen warehouse" to Base 3 72 hours in advance, and updated the WMS warehouse space reservation simultaneously.

[0091] Through steps S301-S303, "demand-based procurement and procurement-based warehousing" are achieved, reducing emergency allocation and transportation costs, lowering the capital tied up in working capital, reducing the delivery distance of agricultural products when customer orders are received, improving delivery efficiency, and reducing the risk of agricultural product spoilage.

[0092] After step S50: acquiring real-time environmental data and real-time traffic data corresponding to the optimized transportation route data, and calculating the actual transportation loss based on the real-time environmental data and real-time traffic data, the following steps are included:

[0093] S501: Associate the historical actual transportation losses with the corresponding agricultural product classification data, and construct an agricultural product transportation loss dataset;

[0094] Specifically, the system archives all transport batches from the past 12 months into two dimensions: "agricultural product category + actual transport loss". For example: broccoli: average loss rate 5.8%, mainly occurring during the high-temperature period of summer in South China; bell peppers: average loss rate 3.2%, concentrated during the high-humidity period of the plum rain season; tomatoes: average loss rate 7.1%, strongly correlated with a congestion index greater than 2.5.

[0095] In this embodiment, the data format is CSV: {category, date, route, carriage temperature, humidity, congestion index, loss rate}, forming a trainable agricultural product transportation loss dataset.

[0096] S502: The pre-set agricultural product transportation loss analysis model analyzes the agricultural product transportation loss dataset based on a machine self-learning algorithm to generate agricultural product transportation loss factors;

[0097] Specifically, the random forest algorithm is used to train the dataset. The input features of the agricultural product transportation loss analysis model include product category, temperature, humidity, congestion index, and transportation time; the output is the "transportation loss factor β". T = Temperature, H = Humidity, CI = Congestion Index The weight values ​​corresponding to temperature. The weighted values ​​corresponding to humidity. In this embodiment of the application, the weight value corresponding to the congestion index is... , and The results were calculated using an analysis model for agricultural product transportation losses.

[0098] For steps S501-S502, experience-based losses are quantified into calculable factors to predict category-level losses in advance, providing a quantitative basis for dynamic temperature adjustment, adding ice, or line modification, effectively reducing overall cargo loss.

[0099] Following step S502: The pre-set agricultural product transportation loss analysis model analyzes the agricultural product transportation loss dataset based on a machine learning algorithm to generate the agricultural product transportation loss factor, the following steps are included:

[0100] S5021: Based on the predicted agricultural product demand data and the transportation loss factor, simulate the shipment situation from each storage point to generate the predicted transportation loss corresponding to each storage and shipment point.

[0101] Specifically, the system takes "500 boxes of broccoli and 200 boxes of bell peppers destined for a market in Qingyuan in the next 7 days" as input, calls the loss factor β=0.058·T+0.032·H+0.021·CI, and simulates the predicted transportation loss for shipments from the Dongguan warehouse and the Huizhou warehouse respectively. Taking the average transportation distance of 180km, T=34 ℃, H=82%, CI=1.8 in the Dongguan warehouse as an example, the predicted loss rate is 4.7%. Taking the average transportation distance of 120km, T=33 ℃, H=80%, CI=1.5 in the Huizhou warehouse as an example, the predicted loss rate is 4.1%. The system outputs the predicted transportation loss matrix for the two warehouses.

[0102] S5022: Compare the predicted transportation losses for each warehouse shipping point to generate transportation loss difference data;

[0103] Specifically, by comparing the results of the two warehouses, the following difference data was obtained: {"Dongguan warehouse - Huizhou warehouse": loss difference = 0.6%, single box loss difference = 0.3 kg, corresponding cost difference = 0.18 yuan / box}. Using quantitative differences to replace experience-based judgments makes warehousing decisions transparent, comparable, and traceable.

[0104] S5023: Optimize the agricultural product warehousing and allocation data based on the transportation loss difference data to adjust the warehousing management of agricultural products, so that different storage points can reduce transportation costs while meeting the warehousing needs of agricultural products.

[0105] Specifically, based on the discrepancy data, the system adjusted the original statement "500 boxes of broccoli were all stored in the Dongguan warehouse" to:

[0106] 320 boxes of broccoli were stored in the Huizhou warehouse (low loss), and 180 boxes of broccoli were stored in the Dongguan warehouse (to replenish capacity), which reduced the overall predicted loss rate from 4.7% to 4.2%, saving about 90 yuan per batch in transportation costs.

[0107] While meeting warehousing requirements, data-driven fine-tuning achieves a double reduction in "cargo damage and freight costs," thereby improving the overall economic efficiency of the supply chain.

[0108] Following step S502: The pre-set agricultural product transportation loss analysis model analyzes the agricultural product transportation loss dataset based on a machine learning algorithm to generate the agricultural product transportation loss factor, the following steps are also included:

[0109] S5024: Acquire the calorific data of agricultural products during transportation, and associate the calorific data with a preset common time axis to construct a transportation calorific change dataset;

[0110] Specifically, three sets of NTC thermistors are installed at the top, middle, and bottom of the truck compartment, collecting temperature data every 30 seconds, while simultaneously recording the ambient temperature and solar radiation intensity outside the compartment. Aligned along a common time axis, this creates a time-temperature curve for each batch of vegetables, forming a dataset of transport heat changes. The temperature is further refined into a three-dimensional spatial curve, providing high-resolution thermal field data for subsequent analysis and avoiding errors caused by traditional single-point temperature measurements.

[0111] S5025: Obtain agricultural product packing data during transportation, and input the packing data into the transportation heat change dataset and associate it with the corresponding heat data to construct a packing-heat dataset, wherein the packing data includes the classification, quantity and corresponding placement data of the packed agricultural products;

[0112] Specifically, the barcodes are scanned to record the type, weight (18 kg), box number, and placement location (front left to middle layer of the truck compartment) of each box of broccoli. UWB positioning tags are used to track the box coordinates, and after being correlated with heat change data, a four-dimensional dataset of "box number-type-location-temperature" is generated. This "packing layout" is converted into computable spatial coordinates, enabling heat analysis to be accurate to the individual box level, providing a data foundation for differentiated packing strategies.

[0113] S5026: The pre-set packing heat change analysis model analyzes the packing-heat dataset based on a machine self-learning algorithm to generate heat change correlation coefficients for different agricultural product categories. The heat change correlation coefficients are used to predict the heat change during packing and placement for different agricultural product categories.

[0114] Specifically, a gradient boosting tree model was trained on a four-dimensional dataset of "location-category-temperature" to obtain the correlation coefficient of broccoli heat changes during a certain period in summer: (Car roof:) = 0.85, Middle layer: = 0.62, bottom layer: = 0.41, indicating that the top chamber heats up the fastest, while the bottom chamber heats up the slowest.

[0115] Quantifying "placement location" into a predictable heat coefficient can be directly used to optimize packing and placement, which is expected to further reduce the damage rate of the same product category.

[0116] For steps S5024-S5026, some technologies typically only place 1-2 temperature probes in the center of the carriage to generate "single-point-curve" or "average temperature" reports. This fails to distinguish the micro-environmental differences between different placement locations, leading to the following drawbacks:

[0117] Large error: The temperature rise of the top chamber within the same batch may be 6–8 ℃ higher than that of the bottom chamber, and single-point data may mask the risk of localized overheating;

[0118] The packing is crude: relying on experience to "put heavy goods at the bottom and light goods at the top" or uniformly leaving ventilation gaps, lacking quantitative basis, resulting in a high damage rate, and sacrificing the goods at the top for transportation.

[0119] Remedial measures: After the damage was discovered, the loading was adjusted, but irreversible quality deterioration had already occurred.

[0120] This application achieves the following overall beneficial effects through the three-step closed-loop process of S5024-S5026:

[0121] Improved spatial resolution: Three-dimensional multi-point high-frequency sampling + UWB positioning reduces temperature error and enables visualization of the "box-level" thermal field;

[0122] Data-driven packing strategy: The correlation coefficient of heat change transforms "placement location" into a predictable β value, guiding "high heat-sensitive categories to be placed at the bottom and heat-resistant categories to be placed at the top," reducing the damage rate under the same conditions;

[0123] Pre-decision closed-loop: The model can output the β value before packing, which is linked to the loss factor, and provide the optimal stacking scheme in advance to avoid the additional costs of post-remediation.

[0124] General expansion: The correlation coefficient can be quickly transferred to other categories such as bell peppers and tomatoes without recalibrating the sensor, significantly reducing the cost of large-scale deployment.

[0125] After step S5026: The pre-set packing heat change analysis model analyzes the packing-heat dataset based on a machine self-learning algorithm to generate heat change correlation coefficients for different agricultural product categories, the following steps are included:

[0126] S50261: Obtain actual agricultural product packaging data, and calculate the first predicted heat change data based on the heat change correlation coefficient;

[0127] Specifically, the calculation formula for the first predicted heat change data in this application is as follows: ,in This is the first predicted data for heat change. The correlation coefficient for heat changes corresponding to the packaging location of agricultural products. This is the preset reference temperature change value.

[0128] S50262: Obtain the real-time environmental data and real-time traffic data, and correct the first predicted heat change data to generate the second predicted heat change data;

[0129] The formula for calculating the second predicted heat change data is as follows: ,in, The first predicted heat change corresponding to different placement locations (derived from the correlation coefficient of placement location), T: real-time ambient temperature (°C), RH: real-time relative humidity (%), CI: real-time congestion index. , , These are the temperature environment correction coefficient, humidity environment correction coefficient, and congestion correction coefficient, respectively, all calculated from historical data regression. , , This is the preset baseline value.

[0130] S50263: Generate a corresponding temperature control strategy inside the box based on the preset agricultural product preservation heat threshold and the second predicted heat change data;

[0131] For example, the heat threshold for broccoli preservation is set to "temperature rise inside the box not greater than 2.5 ℃".

[0132] When the temperature of the agricultural products placed on top rises to 3.96 ℃, which is greater than the preset 2.5 ℃, the internal temperature control strategy is triggered: the refrigeration unit is set to -2 ℃; a phase change ice plate is added to the top; and a top ventilation duct is reserved.

[0133] When the temperature of the agricultural products placed on the middle layer rises to 2.89 ℃, which is greater than 2.5 ℃, the internal temperature control strategy is triggered: the refrigeration unit setting is lowered to 0 ℃, the system outputs a temperature control command and synchronizes it to the vehicle controller.

[0134] It achieves precise temperature control at the "carriage-level" and reduces the risk of spoilage during the transportation of agricultural products, while also effectively reducing energy consumption for cooling, thus balancing quality and cost.

[0135] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0136] In one embodiment, a smart distribution management optimization device for agricultural product supply chains is provided, which corresponds one-to-one with the smart distribution management optimization method for agricultural product supply chains described in the above embodiments. For example... Figure 2 As shown, the intelligent distribution management optimization device for the agricultural product supply chain includes a multi-source heterogeneous dataset generation unit 1, which is used to collect multi-source heterogeneous data from the production end, processing end, and transportation end in real time through IoT devices, and classify and grade the multi-source heterogeneous data based on pre-set dynamic desensitization rules to generate a multi-source heterogeneous dataset.

[0137] The traceability record generation unit 2 is used to encrypt the data in the multi-source heterogeneous dataset based on the preset encryption rules, and upload the processed data to the cloud server to generate traceability records;

[0138] Customer order data analysis unit 3 is used to acquire customer order data and analyze the customer order data to generate corresponding initial transportation and delivery data and initial profit expectation data, wherein the initial transportation and delivery data includes initial agricultural product assembly demand data and initial transportation route data.

[0139] The optimized transportation and distribution data generation unit 4 is used to optimize the initial transportation and distribution data based on the actual selected priority to generate optimized transportation and distribution data, wherein the optimized transportation and distribution data includes optimized agricultural product assembly demand data and optimized transportation route data;

[0140] The actual transportation loss calculation unit 5 is used to acquire real-time environmental data and real-time road condition data corresponding to the optimized transportation route data, and calculate the actual transportation loss based on the real-time environmental data and real-time road condition data. The actual transportation loss includes agricultural product loss and transportation vehicle loss.

[0141] The transportation strategy adjustment unit 6 is used to compare the actual transportation loss with a preset loss threshold, and adjust the transportation strategy if the loss exceeds the threshold.

[0142] Specific limitations regarding the intelligent distribution management optimization device for agricultural product supply chains can be found in the limitations of the intelligent distribution management optimization method for agricultural product supply chains described above, and will not be repeated here. Each module in the aforementioned intelligent distribution management optimization device for agricultural product supply chains can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in an electronic device, or stored in the memory of an electronic device as software, so that the processor can call and execute the corresponding operations of each module.

[0143] In one embodiment, an electronic device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 3 As shown, this electronic device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores the database. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a smart distribution management optimization method for the agricultural product supply chain.

[0144] In one embodiment, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps:

[0145] The multi-source heterogeneous data from the production, processing and transportation ends is collected in real time by IoT devices, and the multi-source heterogeneous data is classified and graded based on pre-set dynamic desensitization rules to generate a multi-source heterogeneous dataset.

[0146] The data in the multi-source heterogeneous dataset is encrypted based on the preset encryption rules, and the processed data is uploaded to the cloud server to generate a traceability record.

[0147] Acquire customer order data and analyze the customer order data to generate corresponding initial transportation and delivery data and initial profit expectation data, wherein the initial transportation and delivery data includes initial agricultural product assembly demand data and initial transportation route data;

[0148] The initial transportation and distribution data is optimized based on the actual selected priority to generate optimized transportation and distribution data, wherein the optimized transportation and distribution data includes optimized agricultural product assembly demand data and optimized transportation route data;

[0149] The system acquires real-time environmental data and real-time road condition data corresponding to the optimized transportation route data, and calculates the actual transportation loss based on the real-time environmental data and real-time road condition data. The actual transportation loss includes agricultural product loss and transportation vehicle loss.

[0150] The actual transportation loss is compared with a preset loss threshold. If the loss exceeds the threshold, the transportation strategy is adjusted.

[0151] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0152] The multi-source heterogeneous data from the production, processing and transportation ends is collected in real time by IoT devices, and the multi-source heterogeneous data is classified and graded based on pre-set dynamic desensitization rules to generate a multi-source heterogeneous dataset.

[0153] The data in the multi-source heterogeneous dataset is encrypted based on the preset encryption rules, and the processed data is uploaded to the cloud server to generate a traceability record.

[0154] Acquire customer order data and analyze the customer order data to generate corresponding initial transportation and delivery data and initial profit expectation data, wherein the initial transportation and delivery data includes initial agricultural product assembly demand data and initial transportation route data;

[0155] The initial transportation and distribution data is optimized based on the actual selected priority to generate optimized transportation and distribution data, wherein the optimized transportation and distribution data includes optimized agricultural product assembly demand data and optimized transportation route data;

[0156] The system acquires real-time environmental data and real-time road condition data corresponding to the optimized transportation route data, and calculates the actual transportation loss based on the real-time environmental data and real-time road condition data. The actual transportation loss includes agricultural product loss and transportation vehicle loss.

[0157] The actual transportation loss is compared with a preset loss threshold. If the loss exceeds the threshold, the transportation strategy is adjusted.

[0158] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0159] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0160] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for optimizing intelligent distribution management in the agricultural product supply chain, characterized in that, The method includes the following steps: real-time collection of multi-source heterogeneous data from the production end, processing end, and transportation end through IoT devices; classification and grading of the multi-source heterogeneous data based on pre-set dynamic desensitization rules to generate a multi-source heterogeneous dataset. The data in the multi-source heterogeneous dataset is encrypted based on the preset encryption rules, and the processed data is uploaded to the cloud server to generate a traceability record. Acquire customer order data and analyze the customer order data to generate corresponding initial transportation and delivery data and initial profit expectation data, wherein the initial transportation and delivery data includes initial agricultural product assembly demand data and initial transportation route data; The initial transportation and distribution data is optimized based on the actual selected priority to generate optimized transportation and distribution data, wherein the optimized transportation and distribution data includes optimized agricultural product assembly demand data and optimized transportation route data; The system acquires real-time environmental data and real-time road condition data corresponding to the optimized transportation route data, and calculates the actual transportation loss based on the real-time environmental data and real-time road condition data. The actual transportation loss includes agricultural product loss and transportation vehicle loss. The actual transportation loss is compared with a preset loss threshold. If the loss exceeds the threshold, the transportation strategy is adjusted.

2. The intelligent distribution management optimization method for agricultural product supply chain according to claim 1, characterized in that, After acquiring customer order data and analyzing it to generate corresponding initial transportation and delivery data and initial profit expectation data, the process includes the following steps: Associate historical customer order data with a pre-defined public timeline and construct a historical order dataset; The pre-set order demand prediction model analyzes the historical order dataset based on a machine learning algorithm to generate predicted agricultural product demand data, which includes predicted agricultural product type data, agricultural product quantity data, and agricultural product destination data corresponding to the agricultural product; Based on the predicted agricultural product demand data, agricultural product warehousing and allocation data are generated.

3. The intelligent distribution management optimization method for agricultural product supply chain according to claim 2, characterized in that, After obtaining the real-time environmental data and real-time traffic data corresponding to the optimized transportation route data, and calculating the actual transportation loss based on the real-time environmental data and real-time traffic data, the following steps are included: The historical actual transportation losses are associated with the corresponding agricultural product classification data, and an agricultural product transportation loss dataset is constructed. The pre-set agricultural product transportation loss analysis model analyzes the agricultural product transportation loss dataset based on a machine self-learning algorithm to generate agricultural product transportation loss factors.

4. The intelligent distribution management optimization method for agricultural product supply chain according to claim 3, characterized in that, After the step of analyzing the agricultural product transportation loss dataset based on a machine learning algorithm using a pre-set agricultural product transportation loss analysis model to generate the agricultural product transportation loss factor, the following steps are included: Based on the predicted agricultural product demand data and the transportation loss factor, the scenario of shipping from each storage point is simulated to generate the predicted transportation loss corresponding to each storage and shipping point. Compare the predicted transportation losses for each warehouse and shipping point to generate transportation loss difference data. The agricultural product warehousing and allocation data are optimized based on the transportation loss difference data to adjust the warehousing management of agricultural products, so that different storage points can reduce transportation costs while meeting the warehousing needs of agricultural products.

5. The intelligent distribution management optimization method for agricultural product supply chain according to claim 4, characterized in that, After the step of analyzing the agricultural product transportation loss dataset based on a pre-set agricultural product transportation loss analysis model using a machine self-learning algorithm to generate the agricultural product transportation loss factor, the following steps are included: Acquire the caloric data of agricultural products during transportation and associate the caloric data with a pre-set public time axis to construct a transportation caloric change dataset; Data on agricultural product packing during transportation is obtained, and the packing data is input into the transportation heat change dataset and associated with the corresponding heat data to construct a packing-heat dataset, wherein the packing data includes the classification, quantity and corresponding placement data of the packed agricultural products; The pre-set packing heat change analysis model analyzes the packing-heat dataset based on a machine learning algorithm to generate heat change correlation coefficients for different agricultural product categories. These heat change correlation coefficients are used to predict heat changes during packing and placement for different agricultural product categories.

6. The intelligent distribution management optimization method for agricultural product supply chain according to claim 5, characterized in that, After the pre-set packing heat change analysis model analyzes the packing-heat dataset based on a machine self-learning algorithm to generate the heat change correlation coefficients corresponding to different agricultural product categories, the following steps are included: Obtain actual agricultural product packaging data, and calculate the first predicted heat change data based on the heat change correlation coefficient; The real-time environmental data and real-time traffic data are acquired, and the first predicted heat change data is corrected to generate the second predicted heat change data. A corresponding temperature control strategy for the container is generated based on the preset heat threshold for agricultural product preservation and the second predicted heat change data.

7. The intelligent distribution management optimization method for agricultural product supply chain according to claim 6, characterized in that, The formula for calculating the first predicted heat change data is: ,in This is the first predicted data for heat change. The correlation coefficient for heat changes corresponding to the packaging location of agricultural products. This is the preset reference temperature change value; The formula for calculating the second predicted heat change data is as follows: ,in, The first predicted heat change corresponds to different placement locations, where T is the real-time ambient temperature, RH is the real-time relative humidity, and CI is the real-time congestion index. , , These are the temperature environment correction coefficient, humidity environment correction coefficient, and congestion correction coefficient, respectively, all calculated from historical data regression. , , This is the preset baseline value.

8. A smart distribution management optimization device for an agricultural product supply chain, applied to the smart distribution management optimization method for an agricultural product supply chain as described in any one of claims 1-7, characterized in that, The device includes: The multi-source heterogeneous dataset generation unit (1) is used to collect multi-source heterogeneous data from the production end, processing end and transportation end in real time through IoT devices, and classify and grade the multi-source heterogeneous data based on the preset dynamic desensitization rules to generate a multi-source heterogeneous dataset. The traceability record generation unit (2) is used to encrypt the data in the multi-source heterogeneous dataset based on the preset encryption rules, and upload the processed data to the cloud server to generate traceability records; The customer order data analysis unit (3) is used to acquire customer order data and analyze the customer order data to generate corresponding initial transportation and delivery data and initial profit expectation data, wherein the initial transportation and delivery data includes initial agricultural product assembly demand data and initial transportation route data. The optimized transportation and distribution data generation unit (4) is used to optimize the initial transportation and distribution data based on the actual selected priority to generate optimized transportation and distribution data, wherein the optimized transportation and distribution data includes optimized agricultural product assembly demand data and optimized transportation route data; The actual transportation loss calculation unit (5) is used to obtain real-time environmental data and real-time road condition data corresponding to the optimized transportation route data, and calculate the actual transportation loss based on the real-time environmental data and real-time road condition data. The actual transportation loss includes agricultural product loss and transportation vehicle loss. The transportation strategy adjustment unit (6) is used to compare the actual transportation loss with a preset loss threshold, and adjust the transportation strategy if the loss threshold is exceeded.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the intelligent distribution management optimization method for agricultural product supply chain as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the intelligent distribution management optimization method for agricultural product supply chain as described in any one of claims 1 to 7.

Citation Information

Cited By

  • Vehicle-mounted agricultural product distribution process monitoring system and method

    CN121544145A