Agricultural product supply chain transportation management method and system based on multi-source data fusion

Abstract

The invention discloses an agricultural product supply chain transportation management method and system based on multi-source data fusion, relates to the technical field of agricultural product logistics, and solves the problems that in the prior art, a long-term fixed mode is mostly adopted for supplier cooperation, sudden demand changes or regional supply shortage cannot be flexibly responded, and the transportation efficiency is high. The agricultural products cannot be supplied timely and effectively, and the accuracy and efficiency of the agricultural product supply chain transportation management method are low. Agricultural product demand prediction quantity is generated based on market data and environmental data; dynamically searching an optimal cooperative supplier based on the agricultural product demand prediction amount; the optimal transportation path is generated based on the cooperative suppliers, and when the agricultural product demand prediction amount is obtained, the supply condition and the comprehensive capability of the current cooperative suppliers are analyzed, so that whether the suppliers need to be adjusted is determined, the supply demand can be accurately and timely completed, the supply stability is improved, and the supply efficiency is improved. And the accuracy and efficiency of the supply chain transportation management method are improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of agricultural product logistics, and in particular relates to an agricultural product supply chain transportation management method and system based on multi-source data fusion. BACKGROUND

[0002] Agricultural product supply chain transportation management refers to a systematic management activity of planning, organizing, coordinating and controlling the transportation link in the entire flow process of agricultural products from producers to consumers. The core lies in ensuring the effective connection of agricultural products in time, space and quality, reducing logistics loss and improving circulation efficiency through scientific and efficient transportation management. This management process involves the selection of transportation modes and the optimization of transportation paths, and is an important link to ensure the safety of agricultural product circulation and stable market supply.

[0003] In the prior art, long-term fixed mode is often adopted for supplier cooperation when selecting suppliers, which cannot flexibly respond to sudden demand changes or regional supply shortages, resulting in that agricultural products cannot be supplied in time and effectively, and thus the accuracy and efficiency of the agricultural product supply chain transportation management method are low. Therefore, the agricultural product supply chain transportation management method still needs to be further improved. SUMMARY

[0004] The present application aims to at least solve one of the technical problems existing in the prior art. To this end, the present application provides an agricultural product supply chain transportation management method and system based on multi-source data fusion, which is used to solve the technical problem that long-term fixed mode is often adopted for supplier cooperation in the prior art, which cannot flexibly respond to sudden demand changes or regional supply shortages, resulting in that agricultural products cannot be supplied in time and effectively, and thus the accuracy and efficiency of the agricultural product supply chain transportation management method are low.

[0005] To achieve the above-mentioned purpose, the first aspect of the present application provides an agricultural product supply chain transportation management method based on multi-source data fusion, comprising:

[0006] obtaining market data and environment data;

[0007] generating an agricultural product demand prediction quantity based on the market data and the environment data; the agricultural product demand prediction quantity refers to a demand prediction quantity of agricultural products in a future period of time;

[0008] dynamically finding an optimal cooperative supplier based on the agricultural product demand prediction quantity;

[0009] generating an optimal transportation path based on the cooperative supplier; the transportation path refers to an optimal transportation path selected for each supplier.

[0010] The application analyzes the supply status and comprehensive ability of the current cooperative supplier on the basis of obtaining the agricultural product demand prediction quantity through the above steps, and then judges whether the supplier needs to be adjusted, which helps to accurately and timely meet the supply demand, effectively improves the stability of the supply, and enhances the accuracy and execution efficiency of the supply chain transportation management method.

[0011] Further, the agricultural product demand prediction quantity is generated based on market data and environment data, comprising:

[0012] The market data comprises sales data and social public opinion; and the environment data comprises temperature, rainfall and weather forecast;

[0013] The sales data, social public opinion, environment data and time characteristics corresponding to a plurality of historical time points in a sliding time window are extracted; the sliding time window refers to the size of the historical time range;

[0014] The sales data, social public opinion, environment data and time characteristics corresponding to the same historical time point are spliced into historical time point data;

[0015] The historical time point data corresponding to a plurality of historical time points are integrated into a demand time sequence according to the chronological order;

[0016] The demand time sequence is input into an agricultural product demand prediction model to obtain an agricultural product demand prediction quantity; the agricultural product demand prediction model is constructed by a machine learning model and is used for predicting the agricultural product demand in a future period of time.

[0017] Further, the agricultural product demand prediction model is constructed by a machine learning model, comprising:

[0018] A plurality of historical demand time sequences and their corresponding historical agricultural product demand quantities are obtained;

[0019] The plurality of historical demand time sequences and their corresponding historical demand quantities are divided into training data, validation data and test data; the training data, validation data and test data are preprocessed to obtain a training set, a validation set and a test set;

[0020] A machine learning model is selected as a base model;

[0021] The base model is trained by the training set, and the learning rate and hyperparameters are adjusted on the validation set to obtain a pre-trained model;

[0022] The pre-trained model is verified on the test set, and finally an agricultural product demand prediction model with the demand time sequence as the input and the agricultural product demand prediction quantity as the output is obtained.

[0023] The application can more accurately predict the demand in a future period of time, realize equivalent matching of production and sales, and reduce the loss of unsold agricultural products by using a pre-trained agricultural product demand prediction model, combining current and historical market data and environmental data, and adding social public opinion data in the market data.

[0024] Further, the dynamic searching for the optimal cooperative supplier based on the agricultural product demand prediction quantity comprises:

[0025] obtaining supplier data; the supplier data comprises cooperative supplier IDs and corresponding cooperative supplier data, and alternative supplier IDs and corresponding alternative supplier data;

[0026] generating cooperative priority and alternative priority based on the supplier data;

[0027] when the cooperative priority is not 0, extracting the maximum supply quantity in the cooperative supplier data; the maximum supply quantity represents the maximum quantity that the current cooperative supplier can supply for the agricultural product demand prediction quantity;

[0028] judging whether the agricultural product demand prediction quantity is less than the maximum supply quantity;

[0029] Yes, keep the cooperative supplier ID unchanged;

[0030] No, calculating the agricultural product demand gap quantity by difference calculation on the agricultural product demand prediction quantity and the maximum supply quantity; sequentially selecting a plurality of alternative supplier IDs based on the high-low order of the alternative priority until the agricultural product demand gap quantity is satisfied; and taking the plurality of alternative supplier IDs and the cooperative supplier ID as cooperative suppliers;

[0031] when the cooperative priority is 0, suspending the transaction of the cooperative supplier ID;

[0032] sequentially selecting a plurality of alternative supplier IDs based on the high-low order of the alternative priority until the agricultural product demand prediction quantity is satisfied; and taking the plurality of alternative supplier IDs as cooperative suppliers.

[0033] Further, the generating of the cooperative priority and the alternative priority based on the supplier data comprises:

[0034] the supplier data comprises cooperative supplier data and alternative supplier data;

[0035] obtaining weight labels, agricultural product demand prediction increment, demand prediction fluctuation variance, and agricultural product value; the weight labels comprise cooperative labels and alternative labels;

[0036] splicing the agricultural product demand prediction increment, the demand prediction fluctuation variance, and the agricultural product value into weight analysis data;

[0037] inputting the weight label and the weight analysis data into a weight generation model to obtain a plurality of weight coefficients corresponding to the weight label; the weight generation model is constructed by an artificial intelligence model and is used for predicting an influence degree of the agricultural product on the supplier priority;

[0038] extracting a supply capacity score, a quality reliability score and a delivery agility score in the cooperative supplier data; the supply capacity score is determined by a ratio between a maximum supply amount of the cooperative supplier and a demand prediction amount of the agricultural product; the quality reliability score is determined by a ratio between a qualified batch and a total batch in the cooperative supplier data in the last N months; the delivery agility score is determined by a ratio between a delayed order number and a total order number in the cooperative supplier data in the last N months; wherein N is an integer;

[0039] when any one of the supply capacity score, the quality reliability score and the delivery agility score is less than a corresponding score threshold, setting the cooperative priority to 0;

[0040] otherwise, the supply capacity score, the quality reliability score and the delivery agility score of the cooperative supplier data are weight fused by a plurality of weight coefficients of the cooperative label to obtain the cooperative priority;

[0041] extracting a supply potential score, a quality assessment score, a delivery security score and a financial health score in the alternative supplier data; the supply potential score is comprehensively evaluated by a production line scale and a maximum theoretical production capacity of the alternative supplier; the quality assessment score is comprehensively scored by a quality inspection laboratory grade and an agricultural product defect rate in the alternative supplier data; the delivery security score is comprehensively evaluated by a transportation cooperation partner list and a severe weather distribution plan in the alternative supplier data; and the financial health score is comprehensively evaluated by an asset-liability ratio and a cash flow coverage rate in the alternative supplier data;

[0042] when any one of the supply potential score, the quality assessment score, the delivery security score and the financial health score is less than a corresponding score threshold, setting the alternative priority to 0;

[0043] otherwise, the supply potential score, the quality assessment score, the delivery security score and the financial health score of the alternative supplier data are weight fused by a plurality of weight coefficients of the alternative label to obtain the alternative priority.

[0044] Further, the weight generation model is constructed by an artificial intelligence model, comprising:

[0045] obtaining historical weight analysis data and historical weight coefficients corresponding to the weight label;

[0046] The historical weight analysis data corresponding to the weight label and the historical weight coefficients are divided into training data, verification data and test data; the training data, the verification data and the test data are preprocessed to obtain a training set, a verification set and a test set;

[0047] The artificial intelligence model is selected as the base model;

[0048] The base model is trained by the training set, and the learning rate and the hyperparameters are adjusted on the verification set to obtain a pre-trained model;

[0049] The pre-trained model is verified on the test set, and finally a weight generation model is obtained, which inputs the weight label and the weight analysis data and outputs the weight coefficients corresponding to the weight label.

[0050] The application predicts the demand prediction of agricultural products in the future period of time, analyzes it to obtain a plurality of demand prediction change parameters, and then obtains the weight coefficients corresponding to the index scores of different types of suppliers in the calculation of priority through the pre-trained weight generation model, so that the priority of the supplier can be dynamically adjusted according to the real-time prediction result when the priority of the supplier is analyzed, realizing the real-time linkage of demand prediction and supplier evaluation, improving the accuracy of supplier priority calculation, and providing accurate data support for subsequent dynamic selection of suppliers.

[0051] Further, the optimal transportation path based on the cooperative supplier includes:

[0052] Extracting a plurality of path data corresponding to the cooperative supplier;

[0053] Generating a plurality of target sub-functions MZF based on the path data;

[0054] Constructing a path selection target function; the path selection target function satisfies the following formula:

[0055] Wherein, i represents a plurality of path numbers corresponding to the cooperative supplier, HS represents the total number of target sub-functions, and n represents the number of target sub-functions; MQ i,n The target weight coefficient corresponding to the nth target sub-function in the ith path is represented as MQ

[0056] The transportation path corresponding to the cooperative supplier is obtained by solving the path selection target function.

[0057] Further, the plurality of target sub-functions MZF are generated based on the path data, including:

[0058] The target sub-function includes a transportation timeliness sub-function, a comprehensive cost sub-function and a risk disaster sub-function;

[0059] extracting path distance LJ, node quantity JS, path average vehicle speed LPV, accident occurrence rate SFL, path flatness LPD and geological disaster probability DZL in the path data;

[0060] constructing a transportation timeliness sub-function YSZF by formula i ; wherein i represents a plurality of path numbers corresponding to the cooperative supplier; j represents a node number, ZXTS i,j represents the loading and unloading stay time required at the jth node in the ith path;

[0061] constructing a comprehensive cost sub-function ZCZF by formula i ; wherein i represents a plurality of path numbers corresponding to the cooperative supplier; GF i represents the high-speed fee required in the ith path; NH represents the unit energy consumption; NDJ represents the unit price of energy consumption; a represents the vehicle time depreciation rate, and the specific value is set according to experience;

[0062] constructing a risk disaster sub-function FZZF by formula i ; wherein i represents a plurality of path numbers corresponding to the cooperative supplier; QFD i represents the meteorological risk grade in the ith path, which is obtained by weather forecast in the environmental data; β1 represents an accident risk sensitive coefficient, β1 ∈ (0, 1); β2 represents a disaster sensitive coefficient, β2 > 0.

[0063] Further, the target weight coefficient is obtained by the following method, comprising:

[0064] obtaining the agricultural product ID and the agricultural product type;

[0065] The agricultural product type includes high-value fresh food, high-value storage-resistant products, low-value fresh food and low-value storage-resistant products.

[0066] Based on the agricultural product type, a plurality of target weight coefficients corresponding to the agricultural product ID are obtained from a type target weight table; the type target weight table is set by experts according to the influence of the agricultural product type on transportation.

[0067] Another aspect of the present application provides a multi-source data fusion based agricultural product supply chain transportation management system, comprising: a data acquisition module and a data analysis module; the data acquisition module and the data analysis module are connected;

[0068] The data acquisition module: obtaining market data and environmental data through a data acquisition device;

[0069] ​​​The data analysis module comprises a demand prediction unit, a supply adjustment unit and a path planning unit;

[0070] The demand prediction unit generates a demand prediction quantity of agricultural products based on market data and environmental data;

[0071] The supply adjustment unit dynamically searches for an optimal cooperative supplier based on the demand prediction quantity of agricultural products;

[0072] The path planning unit generates an optimal transportation path based on the cooperative supplier.

[0073] Compared with the prior art, the application has the following beneficial effects:

[0074] 1. The application generates a demand prediction quantity of agricultural products based on market data and environmental data, dynamically searches for an optimal cooperative supplier based on the demand prediction quantity of agricultural products, and generates an optimal transportation path based on the cooperative supplier. When the demand prediction quantity of agricultural products is obtained, the supply situation and comprehensive capability of the current cooperative supplier are analyzed to determine whether the supplier needs to be adjusted, so that the supply demand can be accurately and timely completed, and the stability of the supply and the accuracy and efficiency of the supply chain transportation management method are improved.

[0075] 2. The application analyzes the supply priority of the current cooperative supplier and the alternative supplier, dynamically adjusts the selection of the supplier under different supply priority states, so that the supplier can meet the demand in the supply chain, and according to the achievement of the demand prediction quantity of agricultural products, the appropriate supplier is selected in order of the priority of the supplier, which reduces the risk of chain breakage caused by single supplier dependence, and further improves the accuracy and efficiency of the supply chain transportation management method.

[0076] 3. After determining the agricultural product supplier, the application designs a path selection target function based on the path data between each supplier and the destination, analyzes several dimensions of the supply chain demand to construct the corresponding target sub-function, and adaptively selects the target weight coefficient corresponding to each target sub-function according to the type of the transported agricultural products, so that the path selection target function can better meet the actual demand. Solving the path selection target function can obtain the optimal transportation path of the agricultural product supplier when transporting agricultural products, and the accuracy and efficiency of the agricultural product supply chain transportation management method are improved. BRIEF DESCRIPTION OF DRAWINGS

[0077] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the drawings needed in the embodiments or the prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0078] Figure 1 A flow chart of the agricultural product supply chain transportation management method based on multi-source data fusion of the present application;

[0079] Figure 2 A schematic diagram of the agricultural product supply chain transportation management system based on multi-source data fusion of the present application. DETAILED DESCRIPTION

[0080] The technical solutions of the present application will be described in detail below in conjunction with embodiments. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0081] Please refer to Figure 1 The first aspect embodiment of the present application provides an agricultural product supply chain transportation management method based on multi-source data fusion, comprising:

[0082] Obtaining market data and environmental data;

[0083] Generating an agricultural product demand prediction based on the market data and the environmental data; the agricultural product demand prediction refers to a demand prediction of the agricultural product in a future period of time;

[0084] Dynamically finding an optimal cooperative supplier based on the agricultural product demand prediction;

[0085] Generating an optimal transportation path based on the cooperative supplier; the transportation path refers to selecting an optimal transportation path for each supplier.

[0086] In the embodiment, generating an agricultural product demand prediction based on market data and environmental data comprises:

[0087] The market data includes sales data and social public opinion; the environmental data includes temperature, rainfall and weather forecast; the weather forecast is used to monitor the weather state in a future period of time, which has a great influence on the demand prediction of the agricultural product in the future period of time; the social public opinion also has a certain influence on the demand prediction, for example, the topic popularity of the social platform “seasonal agricultural product”, which can predict the consumption trend in advance when predicting the demand, so as to improve the accuracy of the demand prediction;

[0088] Extracting sales data, social public opinion, environmental data and time features corresponding to a plurality of historical time points in a sliding time window; the sliding time window refers to the size of the historical time range, and the specific value is set according to experience; in the embodiment, the sliding time window is set to 30 days; the time features include working days, holidays and other time features;

[0089] The sales data, social public opinion, environmental data and time characteristics corresponding to the same historical time point are spliced into historical time point data;

[0090] The historical time point data corresponding to a plurality of historical time points are integrated into a demand time sequence according to time sequence;

[0091] The demand time sequence is input into an agricultural product demand prediction model to obtain an agricultural product demand prediction value; the agricultural product demand prediction model is constructed by a machine learning model and is used to predict the agricultural product demand in a future period of time.

[0092] The agricultural product demand prediction model in the embodiment is constructed by a machine learning model, including:

[0093] A plurality of historical demand time sequences and corresponding historical agricultural product demand values are obtained;

[0094] The plurality of historical demand time sequences and corresponding historical demand values are divided into training data, verification data and test data; the training data, verification data and test data are preprocessed to obtain a training set, a verification set and a test set; the ratio between the training set, the test set and the verification set is 7:2:1;

[0095] A machine learning model is selected as a basic model; the machine learning model includes an LSTM model and the like;

[0096] The basic model is trained by the training set, and the learning rate and hyperparameters are adjusted on the verification set to obtain a pre-trained model;

[0097] The pre-trained model is verified on the test set, and finally an agricultural product demand prediction model with the demand time sequence as input and the agricultural product demand prediction value as output is obtained.

[0098] The embodiment is based on a pre-trained agricultural product demand prediction model, combines current and historical market data and environmental data, and introduces new influencing factors such as social public opinion in the market data, so that when predicting the agricultural product demand, the market demand change trend in a future period of time can be more accurately grasped, not only the effective matching of agricultural product production and sales is realized, but also the problem of unsalable and loss caused by supply and demand imbalance is significantly reduced, and the overall operation efficiency and economic benefit of the supply chain are improved.

[0099] The optimal cooperative supplier is dynamically found based on the agricultural product demand prediction value in the embodiment, including:

[0100] Supplier data is obtained; the supplier data includes cooperative supplier IDs and corresponding cooperative supplier data, and alternative supplier IDs and corresponding alternative supplier data;

[0101] generating a cooperation priority and an alternative priority based on the supplier data; the cooperation priority refers to a result of quantifying selection of a current cooperation supplier; the alternative priority refers to a result of quantifying selection of an alternative supplier;

[0102] when the cooperation priority is not 0, extracting a maximum supplyable amount in the cooperation supplier data; the maximum supplyable amount represents a maximum amount of supplyable by the current cooperation supplier to the agricultural product demand prediction amount;

[0103] judging whether the agricultural product demand prediction amount is less than the maximum supplyable amount;

[0104] yes, keeping the cooperation supplier ID unchanged;

[0105] no, calculating an agricultural product demand gap amount by difference between the agricultural product demand prediction amount and the maximum supplyable amount; selecting a plurality of alternative supplier IDs in a high-low order based on the alternative priority until the agricultural product demand gap amount is satisfied; taking the plurality of alternative supplier IDs and the cooperation supplier ID as cooperation suppliers; when a priority of the current cooperation supplier satisfies a cooperation requirement, the agricultural product demand gap amount that cannot be provided by the cooperation supplier ID is provided by the alternative supplier ID, so that the supply of the agricultural product demand prediction amount by the supplier reaches the requirement;

[0106] when the cooperation priority is 0, suspending transaction of the cooperation supplier ID;

[0107] selecting a plurality of alternative supplier IDs in a high-low order based on the alternative priority until the agricultural product demand prediction amount is satisfied; taking the plurality of alternative supplier IDs as cooperation suppliers.

[0108] The embodiment realizes dynamic adjustment of supplier selection in different priority states by analyzing supply priorities of the current cooperation supplier and the alternative supplier, and ensures that the actual demand in the supply chain can be better satisfied; at the same time, according to the achievement of the agricultural product demand prediction, suitable suppliers are screened in a high-to-low order of the supplier priority, the risk of chain breakage caused by excessive dependence on a single supplier is effectively reduced, the accuracy and operation efficiency of the supply chain transportation management are further improved, and the ability of the system to respond to market fluctuations is enhanced.

[0109] The cooperation priority and the alternative priority based on the supplier data in the embodiment include:

[0110] the supplier data includes cooperation supplier data and alternative supplier data;

[0111] obtaining a weight label, an agricultural product demand prediction increase, a demand prediction fluctuation variance and an agricultural product value; the weight label includes a cooperation label and an alternative label;

[0112] The demand prediction increase, the demand prediction fluctuation variance, and the agricultural product value are spliced into weight analysis data;

[0113] The weight label and the weight analysis data are input into a weight generation model to obtain a plurality of weight coefficients corresponding to the weight label; the weight generation model is constructed by an artificial intelligence model, and is used to predict an influence degree of the agricultural product on the supplier priority;

[0114] The supply capacity score, the quality reliability score, and the delivery agility score in the cooperative supplier data are extracted; the supply capacity score is determined by a ratio between the maximum supply amount of the cooperative supplier and the agricultural product demand prediction amount; the quality reliability score is determined by a ratio between a qualified batch in the last N months and a total batch in the cooperative supplier data; the delivery agility score is determined by a ratio between a delayed order number in the last N months and a total order number in the cooperative supplier data; wherein N is an integer, and a specific value is set according to experience, and in the embodiment, N is set to 3;

[0115] When any one of the supply capacity score, the quality reliability score, and the delivery agility score is less than a corresponding score threshold, the cooperative priority is set to 0; the score threshold is set according to experience, and in the embodiment, the supply capacity score, the quality reliability score, and the delivery agility score have a value range of 0 to 100, and the score threshold corresponding to the supply capacity score, the quality reliability score, and the delivery agility score is all set to 60;

[0116] Otherwise, the supply capacity score, the quality reliability score, and the delivery agility score of the cooperative supplier data are fused by a plurality of weight coefficients of the cooperative label to obtain the cooperative priority;

[0117] The supply potential score, the quality evaluation score, the delivery guarantee score, and the financial health score in the alternative supplier data are extracted; the supply potential score is comprehensively evaluated by a production line scale and a maximum theoretical production capacity of the alternative supplier; the quality evaluation score is comprehensively scored by a quality inspection laboratory grade and an agricultural product defect rate in the alternative supplier data; the delivery guarantee score is comprehensively evaluated by a transportation cooperation partner list and a severe weather distribution plan in the alternative supplier data; and the financial health score is comprehensively evaluated by an asset-liability ratio and a cash flow coverage ratio in the alternative supplier data;

[0118] When any one of the supply potential score, the quality assessment score, the delivery guarantee score and the financial health score is less than the corresponding score threshold, the alternative priority is set to 0; the score threshold is set according to experience; the supply potential score, the quality assessment score, the delivery guarantee score and the financial health score in the embodiment have a value range of 0 to 100, and the corresponding score threshold of the supply potential score, the quality assessment score, the delivery guarantee score and the financial health score is set to 60 points, that is, any one score is less than 60 points, and then the corresponding priority of the supplier is directly set to 0. By default, the supplier is unqualified;

[0119] Otherwise, the supply potential score, the quality assessment score, the delivery guarantee score and the financial health score of the alternative supplier data are fused by weight through a plurality of weight coefficients of the alternative label to obtain the alternative priority.

[0120] The weight generation model in the embodiment is constructed by an artificial intelligence model, including:

[0121] The historical weight analysis data and the historical plurality of weight coefficients corresponding to the weight label are obtained.

[0122] The historical weight analysis data and the historical plurality of weight coefficients corresponding to the weight label are divided into training data, verification data and test data; the training data, the verification data and the test data are preprocessed to obtain a training set, a verification set and a test set; the ratio between the training set, the test set and the verification set is 7:2:1;

[0123] An artificial intelligence model is selected as a base model; the artificial intelligence model includes a BP model and the like;

[0124] The base model is trained through the training set, and the learning rate and the hyperparameter are adjusted on the verification set to obtain a pre-trained model;

[0125] The pre-trained model is verified on the test set, and finally a weight generation model with the input of the weight label and the weight analysis data and the output of the plurality of weight coefficients corresponding to the weight label is obtained.

[0126] The embodiment predicts the demand for agricultural products in the future period of time, and deeply analyzes the demand prediction change parameters; then, the pre-trained weight generation model is used to calculate the weight coefficients corresponding to the index scores in the priority evaluation of different types of suppliers, so that the supplier priority analysis can dynamically adjust the related weight coefficients according to the real-time prediction results, realize the real-time linkage between demand prediction and supplier evaluation, and significantly improve the accuracy and adaptability of supplier priority evaluation, providing scientific and reliable data support for the dynamic selection of subsequent suppliers.

[0127] In this embodiment, generating the optimal transportation route based on cooperative suppliers includes:

[0128] Extract several path data corresponding to the cooperating suppliers;

[0129] Generate several objective sub-functions MZF based on path data;

[0130] Construct a path selection objective function; the path selection objective function satisfies the following formula:

[0131] Where i represents the path numbers corresponding to the cooperating suppliers, HS represents the total number of target sub-functions (in this embodiment, HS is set to 3), and n represents the number of the target sub-functions; MQ i,n The target weight coefficient is represented by the nth objective sub-function in the i-th path; min() represents the minimum value operation. In this embodiment, the path selection objective function is to select the path that makes the target weight coefficient the nth objective sub-function in the i-th path. The minimum value of i is obtained; once i is obtained, the optimal transportation route for the cooperating supplier in transporting agricultural products can be obtained.

[0132] The transportation routes corresponding to the cooperating suppliers are obtained by solving the objective function of route selection;

[0133] In this embodiment, for each cooperating supplier, an optimal transportation route will be selected based on the type of agricultural product being transported, ensuring that the agricultural product can reach the predetermined location in the best possible condition.

[0134] In this embodiment, several target sub-functions (MZFs) are generated based on path data, including:

[0135] The objective sub-functions include the transportation timeliness sub-function, the comprehensive cost sub-function, and the risk and disaster sub-function;

[0136] The path data is extracted as follows: path distance (LJ), number of nodes (JS), average vehicle speed (LPV), accident rate (SFL), path smoothness (LPD), and geological disaster probability (DZL). In this embodiment, the accident rate can be obtained from the traffic management platform, and the path smoothness can be obtained from the road network database. In this embodiment, the path smoothness can be divided into 5 levels, with level 1 representing the worst path smoothness and level 5 representing the best path smoothness. The geological disaster probability is obtained from the geological monitoring system, and its value ranges from 0 to 1. The higher the value, the greater the probability of a geological disaster occurring.

[0137] Through formula Construct the transportation timeliness sub-function YSZF i Where i represents several path numbers corresponding to the cooperating suppliers; j represents the node number, ZXTS i,jrepresents the loading and unloading stay time required at the jth node in the ith path;

[0138] The comprehensive cost sub-function ZCZF is constructed by the formula i ; wherein, i represents a plurality of path numbers corresponding to the cooperative supplier; GF i represents the high-speed cost required in the ith path; NH represents the unit energy consumption; NDJ represents the unit price of energy consumption; in the embodiment, NH includes oil consumption and electricity consumption, and the corresponding unit price of energy consumption NDJ includes the unit price of fuel and the unit price of electricity; α represents the vehicle time depreciation rate, and the specific value is set according to experience, and in the embodiment, α is set to 40 yuan / hour, which includes the driver cost;

[0139] The risk disaster sub-function FZZF is constructed by the formula i ; wherein, i represents a plurality of path numbers corresponding to the cooperative supplier; QFD i represents the meteorological risk level in the ith path, which is obtained by weather forecast in the environmental data, and in the embodiment, the meteorological risk level is in the range of 0 to 1; β1 represents the accident risk sensitive coefficient, β1 ∈ (0, 1); the specific value is set according to experience, and in the embodiment, β1 is set to 0.5; β2 represents the disaster sensitive coefficient, β2 > 0; the specific value is set according to experience, and in the embodiment, β2 is set to 2.

[0140] The target weight coefficient in the embodiment is obtained by the following method, including:

[0141] Obtaining the agricultural product ID and the agricultural product type;

[0142] The agricultural product type includes high-value fresh food, high-value storage product, low-value fresh food and low-value storage product;

[0143] Based on the agricultural product type, a plurality of target weight coefficients corresponding to the agricultural product ID are obtained from the type target weight table; the type target weight table is set by experts according to the influence of the agricultural product type on transportation;

[0144] The type target weight table in the embodiment is as follows:

[0145]

[0146] ​​The embodiment determines the agricultural product suppliers, combines path data between each supplier and the consignee, and constructs a path selection target function; through comprehensive analysis of multiple dimensions of the supply chain demand, corresponding target sub-functions are further designed, and according to the type of the transported agricultural products, the target weight coefficients corresponding to each sub-function are adaptively adjusted, so that the path selection target function is more in line with the actual transportation demand, and through solving the target function, the optimal transportation path can be efficiently obtained, thereby improving the accuracy and overall operation efficiency of the agricultural product supply chain transportation management in path planning.

[0147] Referring to Figure 2 Another aspect of the embodiment of the present application provides an agricultural product supply chain transportation management system based on multi-source data fusion, comprising a data acquisition module and a data analysis module;

[0148] The data acquisition module: market data and environmental data are obtained through a data acquisition device; the data acquisition device comprises a plurality of sensors, such as temperature sensors and the like;

[0149] The data analysis module comprises a demand prediction unit, a supply adjustment unit and a path planning unit;

[0150] The demand prediction unit: agricultural product demand prediction is generated based on market data and environmental data;

[0151] The supply adjustment unit: the optimal cooperative supplier is dynamically found based on the agricultural product demand prediction;

[0152] The path planning unit: the optimal transportation path is generated based on the cooperative supplier.

[0153] Some data in the above formula is calculated by removing the dimension and taking the numerical value, the formula is obtained by software simulation of a large amount of collected data to be closest to the real situation; the preset parameters and the preset threshold in the formula are set by the person skilled in the art according to the actual situation or obtained by a large amount of data simulation.

[0154] The working principle of the present application is: acquiring market data and environment data; generating agricultural product demand prediction based on market data and environment data; dynamically finding the optimal cooperative supplier based on the agricultural product demand prediction; and generating the optimal transportation path based on the cooperative supplier. When the agricultural product demand prediction is obtained, the supply situation and comprehensive ability of the current cooperative supplier are analyzed to determine whether the supplier needs to be adjusted, so that the supply demand can be accurately and timely completed, the supply stability is improved, and the accuracy and efficiency of the supply chain transportation management method are improved. The existing technology is long-term fixed mode for supplier cooperation, which cannot flexibly respond to sudden demand changes or regional supply shortages, leading to the problem that agricultural products cannot be supplied in time and effectively, and thus the accuracy and efficiency of the agricultural product supply chain transportation management method are low.

[0155] The above embodiments are only used to illustrate the technical method of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical method of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical method of the present application.

Claims

1. A method for agricultural product supply chain transportation management based on multi-source data fusion, characterized in that: include: Obtain market and environmental data; Demand forecasts for agricultural products are generated based on market and environmental data. The agricultural product demand forecast refers to the forecast of agricultural product demand over a future period of time. Dynamically identify the optimal cooperative supplier based on agricultural product demand forecasts. The optimal transportation route is generated based on the cooperating suppliers; the transportation route refers to selecting the optimal transportation route for each supplier.

2. The agricultural product supply chain transportation management method based on multi-source data fusion according to claim 1, characterized in that, The generation of agricultural product demand forecasts based on market and environmental data includes: The market data includes sales data and public opinion; the environmental data includes temperature, rainfall, and weather forecasts. Extract sales data, public opinion, environmental data, and time features corresponding to several historical time points within a sliding time window; the sliding time window refers to the size of the historical time range. By combining sales data, public opinion data, environmental data, and time characteristics corresponding to the same historical point in time, historical point-in-time data is formed. The historical time data corresponding to several historical time points are integrated into the required time sequence according to the chronological order. The agricultural product demand forecast model is obtained by inputting the demand time series into the agricultural product demand forecast model. The agricultural product demand forecast model is constructed by machine learning model and is used to predict the demand for agricultural products in the future.

3. The agricultural product supply chain transportation management method based on multi-source data fusion according to claim 2, characterized in that, The agricultural product demand forecasting model is constructed using a machine learning model, including: Obtain several historical demand time series and their corresponding historical agricultural product demand quantities; Several historical demand time series sequences and their corresponding historical demand quantities are divided into training data, validation data, and test data; data preprocessing is performed on the training data, validation data, and test data to obtain the training set, validation set, and test set. Choose a machine learning model as the base model; The base model is trained on the training set, and the learning rate and hyperparameters are adjusted on the validation set to obtain the pre-trained model. By validating the pre-trained model on the test set, we finally obtained an agricultural product demand forecasting model that takes demand time series as input and outputs the predicted quantity of agricultural product demand.

4. The agricultural product supply chain transportation management method based on multi-source data fusion according to claim 1, characterized in that, The method of dynamically finding the optimal cooperative supplier based on agricultural product demand forecasts includes: Obtain supplier data; the supplier data includes cooperative supplier IDs and their corresponding cooperative supplier data, as well as alternative supplier IDs and their corresponding alternative supplier data; Generate cooperation priorities and alternative priorities based on supplier data; When the cooperation priority is not 0, the maximum available quantity is extracted from the cooperative supplier data; the maximum available quantity is the maximum amount that the current cooperative supplier can supply to meet the predicted demand for agricultural products. Determine whether the projected demand for agricultural products is less than the maximum available quantity. Yes, keep the supplier ID unchanged; No, the agricultural product demand gap is calculated by subtracting the forecast demand from the maximum available supply; several candidate supplier IDs are selected sequentially based on their priority until the agricultural product demand gap is met; the candidate supplier IDs and the cooperating supplier IDs are then used as cooperating suppliers. When the cooperation priority is 0, the transaction of the cooperating supplier ID is suspended; Based on the order of priority of the candidates, several candidate supplier IDs are selected in sequence until the predicted demand for agricultural products is met; these candidate supplier IDs are then used as cooperative suppliers.

5. The agricultural product supply chain transportation management method based on multi-source data fusion according to claim 4, characterized in that, The generation of cooperation priorities and alternative priorities based on supplier data includes: The supplier data includes data on cooperating suppliers and data on alternative suppliers; Obtain weighted labels, the projected increase in agricultural product demand, the variance of demand forecast volatility, and the value of agricultural products; the weighted labels include cooperation labels and alternative labels. The forecast growth rate of agricultural product demand, the variance of demand forecast volatility, and the value of agricultural products are combined into weighted analysis data. The weight labels and weight analysis data are input into the weight generation model to obtain several weight coefficients corresponding to the weight labels; the weight generation model is constructed through an artificial intelligence model to predict the degree of influence of agricultural products on supplier priority. Extract supply capacity score, quality reliability score, and delivery agility score from the data of cooperative suppliers; When any one of the following scores—supply capacity score, quality reliability score, and delivery agility score—is less than its corresponding score threshold, the cooperation priority will be set to 0. Otherwise, the cooperation priority is obtained by weighting and fusing the supply capacity score, quality reliability score, and delivery agility score of the cooperative supplier data through several weighting coefficients of the cooperation label; Extract supply potential score, quality assessment score, delivery assurance score, and financial health score from the candidate supplier data; When any one of the following scores—supply potential score, quality assessment score, delivery assurance score, and financial health score—is less than its corresponding score threshold, the priority of the alternative will be set to 0. Otherwise, the supply potential score, quality assessment score, delivery assurance score, and financial health score of the candidate supplier data are weighted and fused using several weighting coefficients of the candidate labels to obtain the candidate priority.

6. The agricultural product supply chain transportation management method based on multi-source data fusion according to claim 5, characterized in that, The weight generation model is constructed using an artificial intelligence model, including: Obtain historical weight analysis data and several historical weight coefficients corresponding to the weight labels; The historical weight analysis data and several historical weight coefficients corresponding to the weight labels are divided into training data, validation data, and test data; the training data, validation data, and test data are preprocessed to obtain the training set, validation set, and test set. Choose an artificial intelligence model as the base model; The base model is trained on the training set, and the learning rate and hyperparameters are adjusted on the validation set to obtain the pre-trained model. By validating the pre-trained model on the test set, a weight generation model is finally obtained, which takes weight labels and weight analysis data as input and outputs several weight coefficients corresponding to the weight labels as output.

7. The agricultural product supply chain transportation management method based on multi-source data fusion according to claim 1, characterized in that, The process of generating the optimal transportation route based on cooperative suppliers includes: Extract several path data corresponding to the cooperating suppliers; Generate several objective sub-functions MZF based on path data; Construct a path selection objective function; the path selection objective function satisfies the following formula: Where i represents the path numbers corresponding to the cooperating suppliers, HS represents the total number of objective sub-functions, and n represents the number of the objective sub-functions; MQ i,n This is represented as the target weight coefficient corresponding to the nth target sub-function in the i-th path; The transportation routes corresponding to the cooperating suppliers are obtained by solving the objective function of route selection.

8. The agricultural product supply chain transportation management method based on multi-source data fusion according to claim 7, characterized in that, The method for generating several target sub-functions (MZF) based on path data includes: The objective sub-function includes a transportation timeliness sub-function, a comprehensive cost sub-function, and a risk and disaster sub-function; Extract the path distance (LJ), number of nodes (JS), average vehicle speed (LPV), accident rate (SFL), path smoothness (LPD), and probability of geological disasters (DZL) from the path data. Through formula Construct the transportation timeliness sub-function YSZF i Where i represents several path numbers corresponding to the cooperating suppliers; j represents the node number, ZXTS i,j This represents the loading / unloading dwell time required at the j-th node in the i-th path; Through formula Construct the comprehensive cost subfunction ZCZF i Where i represents several path numbers corresponding to the cooperating suppliers; GF i Let represent the highway cost required for the i-th path; NH represent the unit energy consumption; NDJ represent the unit price of energy consumption; and α represent the vehicle time depreciation rate. Through formula Construct the risk disaster sub-function FZZF i Where i represents several path numbers corresponding to the cooperating suppliers; QFD i β1 represents the meteorological risk level in the i-th path; β2 represents the accident risk sensitivity coefficient, β1∈(0,1); β2 represents the disaster sensitivity coefficient, β2>0.

9. The agricultural product supply chain transportation management method based on multi-source data fusion according to claim 7, characterized in that, The target weight coefficients are obtained through the following methods: Retrieve agricultural product ID and agricultural product type; The agricultural product types include high-value fresh produce, high-value storable produce, low-value fresh produce, and low-value storable produce. Based on the agricultural product type, several target weight coefficients corresponding to the agricultural product ID are obtained from the type target weight table; the type target weight table is set by experts according to the impact of agricultural product type on transportation.

10. An agricultural product supply chain transportation management system based on multi-source data fusion, applied to the agricultural product supply chain transportation management method based on multi-source data fusion as described in any one of claims 1-9, characterized in that, include: Interconnected data acquisition and data analysis modules; The data acquisition module acquires market data and environmental data through data acquisition equipment. The data analysis module includes a demand forecasting unit, a supply adjustment unit, and a route planning unit. The demand forecasting unit generates a forecast of agricultural product demand based on market data and environmental data. The supply adjustment unit dynamically identifies the optimal cooperative supplier based on the forecast of agricultural product demand. The route planning unit generates the optimal transportation route based on the cooperative suppliers.

Images