Agricultural supply chain demand analysis and prediction system based on digital driving

Abstract

The invention relates to the technical field of prediction and analysis, and discloses an agricultural supply chain demand analysis and prediction system based on digital driving. Comprising a time period formulation module, a data acquisition module, a demand prediction module, a state analysis module and a supply chain adjustment module, and is used for summarizing demand analysis data into a demand combination set, predicting a supply demand value of a next demand time period of a target time period through a demand prediction model, analyzing a demand supply state at a future moment, and adjusting the supply chain. The demand supply state is dynamically adjusted; compared with the prior art, the agricultural product demand quantity can be predicted in advance from the whole cycle flow of the supply chain, limitation existing in single-dimension data collection and local supply chain data collection is avoided, and before the phenomenon of insufficient supply or excessive supply occurs, optimization and adjustment measures of the agricultural supply chain are formulated in advance, so that the agricultural product demand quantity can be predicted in advance. And thus, the agricultural supply chain can be always in a dynamic supply-demand balance state.

Description

Technical Field

[0001] The present invention relates to the field of forecasting and analysis technology, and more specifically, to a digital-driven agricultural supply chain demand analysis and forecasting system. Background Art

[0002] The digitally driven agricultural supply chain refers to the use of modern information technology to digitally manage and optimize each link of the agricultural supply chain in order to improve the efficiency of the supply chain, reduce costs, and enhance the quality and safety of agricultural products. When the demand for agricultural products in the agricultural supply chain increases or decreases, it can timely and accurately judge and predict changes in demand, thereby providing data support for the optimization and adjustment of the agricultural supply chain.

[0003] The patent application with reference publication number CN119026875A discloses a method and system for predicting agricultural supply chain demand based on digital drive, step S1: data collection, collecting historical agricultural supply chain data; step S2: data preprocessing; step S3: establishing an agricultural supply chain demand prediction model, slicing the original data, and performing serial operations based on feature representation; step S4: model parameter adjustment, dividing the parameter interval into multiple sub-intervals to design a unique initialization strategy; step S5: agricultural supply chain demand prediction based on digital drive, realizing agricultural supply chain demand prediction based on the established agricultural supply chain demand prediction model; When conducting demand analysis and forecasting, the existing agricultural supply chain collects data from a certain part or a single dimension of the agricultural supply chain as the basis for forecasting, and combines it with the forecasting model to predict the demand for agricultural products in the agricultural supply chain. For example, in the above-mentioned patent application, it collects crop growth data, climate data, environmental data, soil data and time data as input data of the forecasting model to represent the dynamic changes in the upstream of the agricultural supply chain. As a result, the data dimension involved in the subsequent demand analysis and forecasting is relatively single, and it is impossible to conduct full-cycle comprehensive data collection for the upstream, midstream and downstream of the agricultural supply chain. As a result, the results of the subsequent demand analysis and forecasting will be negatively affected by the data limitations, thereby reducing the accuracy of demand analysis and forecasting in the future.

[0004] In view of this, the present invention proposes a digital-driven agricultural supply chain demand analysis and forecasting system to solve the above problems. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solutions: a digital-driven agricultural supply chain demand analysis and forecasting system, applied to a demand forecasting platform, comprising: The time period formulation module collects the transportation scheduling parameters of the agricultural supply chain, introduces the analysis and prediction mechanism, and formulates the demand time periods with serial numbers. The analysis and prediction mechanism is that two adjacent demand time periods are in a continuous state on the timeline; The data collection module collects demand analysis data from the downstream and upstream supply chains during the demand period, and aggregates the demand analysis data into a demand combination set based on the sequence number; The demand forecasting module takes the demand period of the demand combination set as the target period and predicts the supply demand value of the next demand period of the target period through the demand forecasting model; The status analysis module collects comprehensive logistics data in the middle of the supply chain, calculates the effective storage value of the next demand period during the target period, and analyzes the demand and supply status of the agricultural supply chain at future times; The supply chain adjustment module determines the supply chain adjustment strategy corresponding to the demand and supply status, and dynamically adjusts the demand and supply status of the agricultural supply chain.

[0006] Furthermore, the agricultural supply chain includes the upstream supply chain, the midstream supply chain and the downstream supply chain, and the transportation scheduling parameters include the first transportation time, the warehousing scheduling time and the second transportation time.

[0007] Furthermore, the first transportation time is collected by: Find out the locations of the upstream and midstream of the supply chain through electronic maps, and record them as upstream locations and midstream locations respectively; Determine the vehicle type and transportation date of the transportation vehicle from the upstream location to the midstream location, input the vehicle type and transportation date into the navigation software, and query A first transportation routes from the upstream location to the midstream location; Query the tolls and travel times of A first transport routes, combine the A tolls and the A travel times to obtain A first transport suitability values, and record the travel time corresponding to the minimum of the first transport suitability values ​​as the first transport time.

[0008] Furthermore, the second transportation time is collected by: Find out the location of the downstream of the supply chain through electronic maps and record it as the downstream location; Determine the vehicle type and transportation date of the transportation vehicle from the midstream location to the downstream location, input the vehicle type and transportation date into the navigation software, and query B second transportation routes from the midstream location to the downstream location; The tolls and travel times of B second transport routes are queried, and after combining the B tolls and the B travel times, B second transport suitability values ​​are obtained, and the travel time corresponding to the minimum value of the second transport suitability value is recorded as the second transport time.

[0009] Furthermore, the method for formulating the demand period is as follows: The supply chain cycle is obtained by adding the first transportation time, the warehouse scheduling time, and the second transportation time; Mark C intervals of equal length in the supply chain cycle, and record the period between two adjacent intervals as the demand period, and obtain D demand periods; The D demand periods are numbered in ascending order from front to back along the timeline, so as to formulate D demand periods with serial numbers.

[0010] Furthermore, the demand analysis data includes time period attributes, total output value, unit price excess value and time period transaction rate; The method for collecting the excess unit price is as follows: Calculate the minimum and maximum durations for selling a unit of agricultural products in the upstream supply chain, add the minimum and maximum durations together and calculate the average to obtain the mean duration; Using a duration mean as the division standard, E interval distribution sub-intervals are divided in D demand periods; Query all the sales unit prices in the E sub-intervals of the upstream supply chain one by one, remove the maximum and minimum sales unit prices, and sort the remaining sales unit prices in ascending order from small to large to generate a unit price queue; If the number of sales unit prices in the unit price queue is an even number, add the two sales unit prices in the middle of the unit price queue and calculate the average to obtain the median sales unit price; If the number of sales unit prices in the unit price queue is an odd number, the sales unit price in the middle of the unit price queue is recorded as the median sales unit price; The first unit price, the median unit price, and the last unit price in the unit price queues of the E sub-intervals are added together and averaged to obtain the E unit price means; The E unit price means of D demand periods are added up one by one and averaged, and the average is subtracted from the standard unit price to obtain D unit price excess values.

[0011] Furthermore, the method for collecting the transaction rate of a period is as follows: At the first moment of D demand periods, count the quantity of all agricultural products to be sold in the downstream of the supply chain to obtain D total supply values; At the last moment of D demand periods, count the quantity of all agricultural products to be sold in the downstream of the supply chain and obtain D supply surplus values; Set the sales loss rate of agricultural products, and multiply the sales loss rate by D total supply values ​​to obtain D time period loss values; Subtract the D supply surplus values ​​from the D time period loss values, and compare the difference with the D total supply values ​​to obtain the D time period transaction rate.

[0012] Furthermore, D time period attributes, D total output values, D unit price excess values, and D time period transaction rates are correspondingly summarized to generate D demand combination sets; The training method of the demand forecasting model is: Collect multiple sets of demand combinations and supply demand values ​​in advance, convert peak hours into 1, flat hours into 2, and off-peak hours into 3; Convert multiple sets of demand combinations into multiple feature vectors using a sliding window method, convert supply demand values ​​into labels corresponding to the demand combination sets according to the sliding step length, one feature vector corresponds to one label, and form a set of training data, presetting the prediction time step length, sliding step length, and sliding window length; The feature vector is used as the input of the machine learning model, the last demand period in the sliding window length is taken as the target period, the supply demand value of the demand period after the time step of the target period is predicted as the output, the subsequent supply demand value of each training set is used as the prediction target, and the minimized sum of prediction errors is used as the training target. The machine learning model is trained to train a demand forecasting model.

[0013] Furthermore, the comprehensive logistics data includes the inbound quantity value, outbound quantity value, storage loss value and current balance value; The calculation method of effective storage value is: Add the current surplus value of the next demand period of the target period to the incoming quantity value, and then subtract them from the outgoing quantity value and the storage loss value to obtain the effective storage value of the next demand period of the target period; The demand and supply status includes the supply and demand imbalance status, the supply and demand balance status and the supply and demand excess status. The analysis methods for the supply and demand imbalance status, the supply and demand balance status and the supply and demand excess status are as follows: The effective storage value of the next demand period of the target period in the midstream of the supply chain is subtracted from the predicted supply demand value and the preset storage red line value to obtain the supply-demand difference; When the difference between supply and demand is greater than 0, the demand-supply status is recorded as excess supply and demand status; When the difference between supply and demand is equal to 0, the demand-supply state is recorded as the supply-demand balance state; When the supply-demand difference is less than 0, the demand-supply status is recorded as a supply-demand imbalance state.

[0014] Furthermore, the supply chain adjustment strategy includes a supply increase strategy, a supply maintenance strategy, and a supply reduction strategy. The method for determining the supply increase strategy, the supply maintenance strategy, and the supply reduction strategy is as follows: When the demand-supply state of the agricultural supply chain in the next demand period of the target period is in a supply-demand imbalance state, a strategy to increase supply is determined; When the demand-supply state of the agricultural supply chain in the next demand period of the target period is in a supply-demand balance state, a strategy for maintaining supply is determined; When the demand-supply status of the agricultural supply chain in the next demand period during the target period is in a state of excess supply and demand, a strategy to reduce supply is determined.

[0015] The technical effects and advantages of the digital-driven agricultural supply chain demand analysis and forecasting system of the present invention are as follows: (1): By collecting demand analysis data during the demand period, aggregating the demand analysis data into a demand combination set, and predicting the supply demand value of the next demand period of the target period through the demand forecasting model, it can not only comprehensively collect multi-dimensional data that affects the final demand for agricultural products in the upstream, midstream and downstream of the supply chain, but also avoid the limitations of single-dimensional data collection and local supply chain data collection. At the same time, combined with the prediction method of the demand forecasting model, it can make early and accurate analysis and prediction of the demand for agricultural products in the downstream of the supply chain, thereby effectively avoiding the inefficiency and delay of traditional real-time data collection and analysis and prediction, and improving the effectiveness of demand analysis and prediction.

[0016] (2): By calculating the effective storage value to analyze the demand and supply status at future times, and determining the supply chain adjustment strategy, the demand and supply status of the agricultural supply chain can be dynamically adjusted. Based on the results of advance demand analysis and prediction, the balance between supply and demand of agricultural products in the agricultural supply chain can be accurately analyzed and judged. Before the phenomenon of insufficient or excessive supply occurs, optimization and adjustment measures for the agricultural supply chain can be formulated in advance to prevent the agricultural supply chain from having an imbalance in supply and demand in the future, thereby ensuring that the agricultural supply chain can always be in a dynamic supply and demand balance state. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A module schematic diagram of a digitally driven agricultural supply chain demand analysis and forecasting system provided in Embodiment 1 of the present invention; Figure 2 A flow chart of a method for analyzing and predicting agricultural supply chain demand based on digital drive provided in the second embodiment of the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] Example 1: Please refer to Figure 1 As shown, the agricultural supply chain demand analysis and forecasting system based on digital drive described in this embodiment is applied to the demand forecasting platform, including: The time period setting module collects the transportation scheduling parameters of the agricultural supply chain and introduces an analysis and prediction mechanism to develop demand time periods with serial numbers; The agricultural supply chain refers to a complete chain formed from the procurement of agricultural production materials to the production, storage, transportation and sales of agricultural products. It can summarize the three aspects of upstream production, midstream storage and downstream sales of agricultural products in all dimensions. Therefore, the agricultural supply chain includes the upstream supply chain, the midstream supply chain and the downstream supply chain. Specifically, the upstream of the supply chain corresponds to the planting and harvesting of agricultural products in the agricultural supply chain, the midstream of the supply chain corresponds to the storage and processing of agricultural products in the agricultural supply chain, and the downstream of the supply chain corresponds to the sales and demand of agricultural products in the agricultural supply chain.

[0020] Transportation scheduling parameters refer to the diversified parameters that can affect the overall operation time of agricultural product transportation and scheduling in the agricultural supply chain, thereby representing the changes in the transportation and scheduling of agricultural products in the agricultural supply chain, and providing basic data support for the development and progress of advanced productivity in the agricultural supply chain; The transport scheduling parameters include the first transport duration, the storage scheduling duration and the second transport duration; The first transportation time refers to the transportation time required for agricultural products to be transported from the upstream to the midstream of the supply chain, which can be used to represent the time consumed by agricultural products from the upstream to the midstream of the supply chain; The first transportation time is collected as follows: Find out the locations of the upstream and midstream of the supply chain through electronic maps, and record them as upstream locations and midstream locations respectively; Determine the vehicle type and transportation date of the transportation vehicle from the upstream location to the midstream location, and input the vehicle type and transportation date into the navigation software to query the A first transportation routes from the upstream location to the midstream location; through the automatic route planning method of the navigation software, the required transportation route can be automatically and accurately queried and simulated to ensure the accuracy of the transportation route; Query the tolls and travel times of A first transport routes, combine the A tolls and the A travel times to obtain A first transport suitability values, and record the travel time corresponding to the minimum of the first transport suitability values ​​as the first transport time; when the first transport suitability value is smaller, it means that the toll and travel time are smaller at this time, the cost and time spent on the transport route during the transport process will be less, and the cost performance of the transport route will be higher; The expression of the first transport suitability value is: ; In the formula, For the The first transport suitability value of the first transport route, =1,2,...,A, For the The toll of the first transport route, For the The duration of the first transport route, , All of them are proportional coefficients greater than 0.

[0021] Warehouse scheduling time refers to the time between recording the entry of agricultural products into the warehouse and recording their exit from the warehouse in the middle of the supply chain, which can be used to represent the time consumed between the storage and scheduling of agricultural products in the middle of the supply chain; the warehouse scheduling time is obtained by collecting multiple storage times between entry and exit from the warehouse management system, adding the maximum storage time and the minimum storage time and then taking the average.

[0022] The second transportation time refers to the transportation time required for agricultural products to be transported from the midstream of the supply chain to the downstream of the supply chain, which can be used to represent the time consumed by agricultural products from the midstream of the supply chain to the downstream of the supply chain; The second transportation time is collected by: Find out the location of the downstream of the supply chain through electronic maps and record it as the downstream location; Determine the vehicle type and transportation date of the transportation vehicle from the midstream location to the downstream location, input the vehicle type and transportation date into the navigation software, and query B second transportation routes from the midstream location to the downstream location; Query the tolls and travel times of B second transport routes, combine the B tolls and the B travel times to obtain B second transport suitability values, and record the travel time corresponding to the minimum value of the second transport suitability value as the second transport time; The expression of the second transport suitability value is: ; In the formula, For the The second transport suitability value of the second transport route, =1,2,...,B, For the The toll of the second transport line, For the The duration of the second transport route, , All of them are proportional coefficients greater than 0.

[0023] Once the transport scheduling parameters are obtained, the time span of the entire agricultural supply chain can be determined based on the transport scheduling parameters. At the same time, an analysis and prediction mechanism needs to be introduced to finely divide the time span from the upstream to the downstream of the supply chain, so that the time range with a larger span can be divided into demand periods with a smaller time span. At this time, the demand period can divide the overall larger time range into multiple, continuous and smaller time ranges.

[0024] The analysis and prediction mechanism is used to improve the consistency and continuity of demand periods, ensuring that the demand periods can accurately and sufficiently segment the agricultural supply chain. Therefore, the analysis and prediction mechanism is that two adjacent demand periods are continuous on the timeline.

[0025] The method for formulating the demand period is as follows: The supply chain cycle is obtained by adding the first transportation time, the warehouse scheduling time, and the second transportation time; Mark C intervals of equal length in the supply chain cycle, and record the period between two adjacent intervals as the demand period, and obtain D demand periods; through the demand periods of equal length and continuous distribution, the supply chain cycle can be fully and completely segmented to ensure that the demand period can accurately and finely represent the entire cycle of the agricultural supply chain; The D demand periods are numbered in ascending order from front to back along the timeline, so as to formulate D demand periods with serial numbers.

[0026] It should be noted that the serial numbering can not only distinguish each demand period to avoid confusion among multiple demand periods, but also arrange each demand period in sequence in a timeline manner to facilitate subsequent identification and use of the demand period.

[0027] The data collection module collects demand analysis data from the downstream and upstream supply chains during the demand period, and summarizes the demand analysis data into a demand combination set based on the sequence number. The comprehensive demand data includes time period attributes, total output value, unit price excess value and time period transaction rate; Demand analysis data refers to the comprehensive data in the upstream and downstream supply chains that can affect the final demand analysis and forecast results of agricultural products, and serves as the data basis for predicting agricultural supply chain demand. Therefore, the collection objects of demand analysis data are the upstream and downstream supply chains; Demand analysis data includes time period attributes, total output value, unit price excess value and time period transaction rate; The time period attribute refers to the specific date type corresponding to each demand period in the downstream supply chain, which can represent the sales market trends under different date types. Specifically, the time period attributes include peak period, off-peak period and valley period. The peak period refers to the sales date corresponding to the demand period on holidays of more than 3 days. At this time, the demand for agricultural products in the downstream supply chain increases. The off-peak period refers to the sales date corresponding to the demand period on weekends. At this time, the demand for agricultural products in the downstream supply chain is medium. The valley period refers to the sales date corresponding to the demand period on weekdays. At this time, the demand for agricultural products in the downstream supply chain is small. The time period attribute is obtained by querying the specific date of each demand period through the calendar.

[0028] The total output value refers to the total production of agricultural products corresponding to each demand period in the upstream of the supply chain, which can be used to represent the output of agricultural products in different demand periods. The larger the total output value, the larger the total production of agricultural products corresponding to each demand period in the upstream of the supply chain. At this time, the demand for agricultural products in the downstream of the supply chain during the demand period is greater; the total output value is obtained by counting the total production of agricultural products in the upstream of the supply chain during each demand period.

[0029] The unit price excess value refers to the extent to which the unit price of agricultural products corresponding to each demand period in the downstream supply chain exceeds the standard unit price, which can be used to indicate the high or low unit prices of agricultural products in the downstream supply chain. The larger the unit price excess value, the higher the extent to which the unit price of agricultural products corresponding to each demand period in the downstream supply chain exceeds the standard unit price, and the less the demand for agricultural products in the downstream supply chain. The method for collecting the excess unit price is as follows: The minimum and maximum durations of selling a unit of agricultural products in the upstream of the supply chain are calculated, and the minimum duration and the maximum duration are added and averaged to obtain the duration mean; the unit value is the quantity of agricultural products used to calculate the duration mean, and is used as a specific value for determining the sales duration corresponding to a certain quantity of agricultural products. The specific data of the unit value can be set manually or according to the existing weight unit. For example, one unit value is 1 ton; Using a time length mean as the division standard, E interval-distributed sub-intervals are divided within D demand periods; the interval-distributed sub-intervals can mark time ranges of equal length within the demand period to ensure that the data within two adjacent sub-intervals remain independent and prevent the data within two adjacent sub-intervals from being repeated or too similar; Query all the sales unit prices in the E sub-intervals of the upstream supply chain one by one, remove the maximum and minimum sales unit prices, and sort the remaining sales unit prices in ascending order from small to large to generate a unit price queue; If the number of sales unit prices in the unit price queue is an even number, add the two sales unit prices in the middle of the unit price queue and calculate the average to obtain the median sales unit price; If the number of sales unit prices in the unit price queue is an odd number, the sales unit price in the middle of the unit price queue is recorded as the median sales unit price; The first unit price, the median unit price, and the last unit price in the unit price queues of the E sub-intervals are added together and averaged to obtain the E unit price means; The expression for the mean unit price is: ; In the formula, For the Demand period The mean unit price of the sub-intervals, =1,2,...,D, =1,2,...,E, For the Demand period The sales unit price of the first digit of the sub-interval, For the Demand period The median sales price of each sub-range, For the Demand period The last digit of the sales price in each sub-interval; The average of the E unit prices in the D demand periods is accumulated one by one, and the average is subtracted from the standard unit price to obtain D unit price excess values; The expression for the unit price excess value is: ; In the formula, For the The unit price excess value during the demand period, The standard unit price.

[0030] The time period transaction rate refers to the proportion of the sales volume of agricultural products corresponding to each demand period in the downstream supply chain to the current total volume, which can indicate the sales speed of agricultural products in the downstream supply chain. The larger the unit transaction rate, the greater the proportion of the sales volume of agricultural products corresponding to each demand period in the downstream supply chain to the current total volume, and the greater the demand for agricultural products in the downstream supply chain. The collection method of time period transaction rate is: At the first moment of D demand periods, count the quantity of all agricultural products to be sold in the downstream of the supply chain to obtain D total supply values; At the last moment of D demand periods, count the quantity of all agricultural products to be sold in the downstream of the supply chain and obtain D supply surplus values; Set the sales loss rate of agricultural products, and multiply the sales loss rate by D total supply values ​​to obtain D time period loss values; the sales loss rate of agricultural products refers to the loss probability caused by the sale of agricultural products in the downstream of the supply chain, which is used to calculate the basis for the loss quantity in the entire sales process; The expression of the period loss value is: ; In the formula, For the The time loss value of a demand period, is the sales loss rate, For the The total supply value for a demand period; Subtract the D supply surplus values ​​from the D time period loss values, and compare the difference with the D total supply values ​​to obtain the D time period transaction rate; The expression of the period transaction rate is: ; In the formula, For the The transaction rate of the demand period, For the The supply margin value for a demand period.

[0031] After obtaining the demand analysis data of D demand time periods, the demand analysis data of the D demand time periods can be combined one by one, so that all the demand analysis data in each demand time period can be orderly and accurately summarized together, and the summarized demand analysis data are recorded as a demand combination set to obtain D demand combination sets, so that the D demand combination sets can comprehensively represent the data affecting the demand for agricultural products in the downstream and upstream of the supply chain in the D demand time periods. Therefore, each demand combination set contains an independent time period attribute, total output value, unit price excess value and time period transaction rate, that is: after summarizing the D time period attributes, D total output values, D unit price excess values ​​and D time period transaction rates, D demand combination sets are generated.

[0032] The demand forecasting module takes the demand period of the demand combination set as the target period and predicts the supply demand value in the next demand period of the target period through the demand forecasting model; After obtaining the demand combination set, all demand combination sets can be input into the demand forecasting model, so that the demand forecasting model can predict the supply demand value of the next demand period based on the input demand combination set. Since the supply demand value needs to be predicted at future moments based on the time line of the demand period corresponding to the demand combination set, the demand period of the previous demand combination set of the predicted supply demand value at future moments is recorded as the target period.

[0033] The demand forecasting model is based on the collected demand combination set, combined with artificial intelligence technology, and uses a machine learning model to predict the supply demand value corresponding to the demand combination set, so that the supply demand value at future moments can be predicted based on the demand combination set of the existing demand period; The supply demand value is the total amount of agricultural products required by the downstream supply chain corresponding to the demand combination set in each demand period, and serves as the specific demand value of the agricultural supply chain in each demand period.

[0034] In order to realize the prediction operation of the demand prediction model, it is necessary to train the demand prediction model in advance; The training method of the demand forecasting model is: Collect multiple sets of demand combination sets and supply demand values ​​in advance, and convert the time period attributes of each set of demand combination sets into digital form, for example, convert the peak time period into 1, convert the flat time period into 2, and convert the off-peak time period into 3; The multiple demand combination sets are converted into multiple feature vectors using the sliding window method. The supply demand values ​​are converted into labels corresponding to the demand combination sets according to the sliding step length. One feature vector corresponds to one label and constitutes a set of training data. The demand combination sets are arranged in the order of the demand periods, and the prediction time step length, sliding step length and sliding window length are preset. The feature vector is used as the input of the machine learning model, the last demand period in the sliding window length is taken as the target period, the supply demand value of the demand period after the time step of the target period is predicted as the output, the subsequent supply demand value of each training set is used as the prediction target, and the minimized sum of prediction errors is used as the training target. The machine learning model is trained to train a demand prediction model that predicts the supply demand value of the next demand period of the target period based on the demand combination set.

[0035] Exemplarily, the machine learning model is any one of the CNN neural network model or AlexNet.

[0036] After the demand forecasting model is trained, the collected demand combination set can be used as input data, and the supply demand value of the next demand period of the target period can be used as output data to predict the demand for agricultural products in the downstream of the supply chain in advance, thereby predicting the supply demand value of agricultural products in the downstream of the supply chain in the future.

[0037] The status analysis module collects comprehensive logistics data in the middle of the supply chain, calculates the effective storage value of the next demand period during the target period, and analyzes the demand and supply status of the agricultural supply chain at future times; Comprehensive logistics data refers to the comprehensive data related to the storage and delivery of agricultural products in the corresponding agricultural product warehouses in the middle of the supply chain during each demand period. It can effectively represent the storage quantity of agricultural products in the agricultural product warehouses in the middle of the supply chain during each demand period, thereby calculating the storage quantity of agricultural products in the middle of the supply chain during each demand period and obtaining the effective storage value. Comprehensive logistics data includes inbound quantity value, outbound quantity value, storage loss value and current balance value; The incoming quantity value and the outgoing quantity value refer to the quantity of agricultural products entering the warehouse for storage and being moved out of the warehouse for transportation in the next demand period after the target period, which can be used to represent the incoming and outgoing quantities of agricultural products in the warehouse. Specifically, the incoming quantity value and the outgoing quantity value are obtained by querying the incoming data and the outgoing data of the warehouse management system in the next demand period after the target period.

[0038] The current balance value refers to the total quantity of agricultural products stored in the warehouse during the next demand period after the target period, which can be used to represent the maximum storage quantity of agricultural products in the warehouse during the next demand period after the target period; the current balance value is obtained by querying the warehouse management system.

[0039] The storage loss value refers to the amount of loss incurred when the warehouse stores and manages agricultural products during the next demand period after the target period, so that the lost agricultural products cannot be supplied for external transportation; the storage loss value is obtained by querying the warehouse loss rate and multiplying it by the current balance value.

[0040] After the comprehensive logistics data of the midstream of the supply chain is collected, the effective storage value of the midstream of the supply chain in the next demand period of the target period can be calculated based on the collected data, so that the effective storage value can be used as the maximum quantity of agricultural products supplied to the downstream of the supply chain in the next demand period of the target period; The calculation method of effective storage value is: Add the current surplus value of the next demand period of the target period to the incoming quantity value, and then subtract them from the outgoing quantity value and the storage loss value to obtain the effective storage value of the next demand period of the target period; The expressions for valid storage values ​​are: ; In the formula, is the effective storage value of the next demand period in the target period. is the current margin value of the next demand period in the target period, is the inventory value of the next demand period during the target period, is the outbound quantity value of the next demand period during the target period, It is the storage loss value of the next demand period after the target period.

[0041] After calculating the effective storage value, the effective storage value can be compared with the predicted supply demand value, and based on the comparison results, it can be judged whether the agricultural products stored in the midstream of the supply chain can meet the demand of the downstream supply chain during the next demand period of the target period, thereby representing the demand and supply status of the agricultural supply chain at future times, and judging whether the agricultural supply chain will be under-supply or over-supply in the future.

[0042] The demand and supply status is used to specifically express the demand and supply relationship of the agricultural supply chain in the next demand period after the target period. Specifically, the demand and supply status includes the supply and demand imbalance status, the supply and demand balance status and the supply and demand excess status. The supply and demand imbalance status means that the supply of agricultural products in the next demand period after the target period cannot meet the demand, the supply and demand balance status means that the supply of agricultural products in the next demand period after the target period can meet the demand, and the supply and demand excess status means that the supply of agricultural products in the next demand period after the target period exceeds the demand.

[0043] The analysis methods for supply-demand imbalance, supply-demand balance, and supply-demand excess are as follows: The effective storage value of the next demand period of the target period in the midstream of the supply chain is subtracted from the predicted supply demand value and the preset storage red line value to obtain the supply-demand difference value; the preset storage red line value refers to the minimum quantity of agricultural products that need to be maintained in the warehouse in the midstream of the supply chain, so that agricultural products with the preset storage red line value can be used to meet procedural requirements such as sampling and inspection of agricultural products; The expression for the difference between supply and demand is: ; In the formula, is the supply-demand difference between the target period and the next demand period. Supply the demand value for the next demand period of the predicted target period, It is the storage red line value; when When it is greater than 0, it means that the agricultural products stored in the middle of the supply chain in the next demand period of the target period can exceed the demand of the downstream supply chain, and the agricultural supply chain will have an excess supply of agricultural products in the next demand period of the target period, and the demand-supply state is recorded as the excess supply-demand state; when When it is equal to 0, it means that the agricultural products stored in the middle of the supply chain in the next demand period of the target period can meet the demand of the downstream supply chain, and the agricultural supply chain will not have excess or imbalance in the next demand period of the target period, and the demand-supply state is recorded as the supply-demand balance state; when When it is less than 0, it means that the agricultural products stored in the middle of the supply chain during the next demand period after the target period cannot meet the demand of the downstream supply chain. Then the supply of agricultural products in the agricultural supply chain during the next demand period after the target period will be unbalanced, and the demand-supply status will be recorded as a supply-demand imbalance state.

[0044] The supply chain adjustment module determines the supply chain adjustment strategy corresponding to the demand and supply status, and dynamically adjusts the demand and supply status of the agricultural supply chain; After analyzing the demand and supply status of the agricultural supply chain, the agricultural supply chain can be deeply analyzed and adjusted and optimized according to the actual demand and supply status, so as to determine specific adjustment and optimization plans and record them as supply chain adjustment strategies, so that the supply chain adjustment strategies can be used as guiding measures for adjusting the specific demand and supply status of the agricultural supply chain; Supply chain adjustment strategies include supply increase strategy, supply maintenance strategy and supply reduction strategy; The methods for determining the supply increase strategy, supply maintenance strategy, and supply reduction strategy are as follows: When the demand-supply state of the agricultural supply chain in the next demand period of the target period is in a supply-demand imbalance state, it means that the supply of agricultural products in the downstream of the supply chain will not meet the demand in the future, and the strategy of increasing supply is determined; When the demand-supply state of the agricultural supply chain in the next demand period of the target period is in a supply-demand balance state, it means that there will be no imbalance or excess supply of agricultural products in the downstream of the supply chain in the future, and the supply maintenance strategy is determined; When the demand-supply status of the agricultural supply chain in the next demand period of the target period is in a state of excess supply and demand, it means that there will be an excess supply of agricultural products in the downstream of the supply chain in the future, and a strategy to reduce the supply is determined.

[0045] After determining the corresponding supply chain adjustment strategy, it is necessary to dynamically adjust the demand and supply status of the agricultural supply chain based on the specific supply chain adjustment strategy, so that the demand and supply status of the agricultural supply chain can be in or infinitely close to the supply and demand balance state, thereby avoiding the phenomenon of supply and demand mismatch in the agricultural supply chain in the future and ensuring the supply and demand relationship of the agricultural supply chain; Specifically, when dynamically adjusting the demand and supply status of the agricultural supply chain, if a strategy of increasing supply is determined, it is necessary to increase the storage volume of agricultural products in the midstream of the supply chain until the imbalanced supply and demand state in the next demand period after the target period is at or close to a supply and demand balance state; if a strategy of reducing supply is determined, it is necessary to reduce the storage volume of agricultural products in the midstream of the supply chain until the excess supply and demand state in the next demand period after the target period is at or close to a supply and demand balance state; if a strategy of maintaining supply is determined, there is no need to increase or decrease the storage volume of agricultural products in the midstream of the supply chain.

[0046] In this embodiment, the agricultural supply chain cycle can be obtained based on the transportation scheduling parameters, and the analysis and prediction mechanism is introduced to formulate the demand time period. The overall cycle of the agricultural supply chain can be refined and divided, and the overall cycle can be divided into multiple segments and continuously and orderly, which is convenient for the subsequent accurate collection of relevant data within the demand period.

[0047] By collecting demand analysis data during the demand period, aggregating the demand analysis data into a demand combination set, and predicting the supply demand value of the next demand period during the target period through the demand forecasting model, it is possible to comprehensively collect multi-dimensional data that affects the final demand for agricultural products in the upstream, midstream and downstream of the supply chain, avoiding the limitations of single-dimensional data collection and local supply chain data collection. At the same time, combined with the forecasting method of the demand forecasting model, it is possible to conduct early and accurate analysis and forecast of the demand for agricultural products downstream of the supply chain, thereby effectively avoiding the inefficiency and delay of traditional real-time data collection and analysis and forecasting, and improving the effectiveness of demand analysis and forecasting.

[0048] By calculating the effective storage value to analyze the demand and supply status in the future, and determining the supply chain adjustment strategy, the demand and supply status of the agricultural supply chain can be dynamically adjusted. Based on the results of advance demand analysis and prediction, the balance between supply and demand of agricultural products in the agricultural supply chain can be accurately analyzed and judged. Before the phenomenon of insufficient or excessive supply occurs, optimization and adjustment measures for the agricultural supply chain can be formulated in advance to prevent the imbalance of supply and demand in the agricultural supply chain in the future, thereby ensuring that the agricultural supply chain can always be in a dynamic supply and demand balance state.

[0049] Example 2: Please refer to Figure 2 As shown, the part not described in detail in this embodiment is described in the first embodiment, and a method for analyzing and predicting agricultural supply chain demand based on digital drive is provided, which is implemented based on a system for analyzing and predicting agricultural supply chain demand based on digital drive, including: S1: Collect the transportation scheduling parameters of the agricultural supply chain, introduce the analysis and prediction mechanism, and formulate the demand time periods with serial numbers; S2: Collect demand analysis data from the downstream and upstream supply chains during the demand period, and aggregate the demand analysis data into a demand combination set based on the sequence number; S3: Taking the demand period of the demand combination set as the target period, the supply demand value of the next demand period of the target period is predicted by the demand forecasting model; S4: Collect comprehensive logistics data in the middle of the supply chain, calculate the effective storage value of the next demand period during the target period, and analyze the demand and supply status of the agricultural supply chain at future times; S5: Determine the supply chain adjustment strategy corresponding to the demand and supply status, and dynamically adjust the demand and supply status of the agricultural supply chain.

[0050] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A digital-driven agricultural supply chain demand analysis and forecasting system, applied to a demand forecasting platform, characterized in that: include: The time period formulation module collects the transportation scheduling parameters of the agricultural supply chain, introduces the analysis and prediction mechanism, and formulates the demand time periods with serial numbers. The analysis and prediction mechanism is that two adjacent demand time periods are in a continuous state on the timeline; The data collection module collects demand analysis data from the downstream and upstream supply chains during the demand period, and aggregates the demand analysis data into a demand combination set based on the sequence number; The demand forecasting module takes the demand period of the demand combination set as the target period and predicts the supply demand value of the next demand period of the target period through the demand forecasting model; The status analysis module collects comprehensive logistics data in the middle of the supply chain, calculates the effective storage value of the next demand period during the target period, and analyzes the demand and supply status of the agricultural supply chain at future times; The supply chain adjustment module determines the supply chain adjustment strategy corresponding to the demand and supply status, and dynamically adjusts the demand and supply status of the agricultural supply chain.

2. According to claim 1, a digitally driven agricultural supply chain demand analysis and forecasting system is characterized in that: The agricultural supply chain includes the upstream supply chain, the midstream supply chain and the downstream supply chain, and the transportation scheduling parameters include the first transportation time, the warehousing scheduling time and the second transportation time.

3. The agricultural supply chain demand analysis and forecasting system based on digital drive according to claim 2 is characterized in that: The first transportation time is collected as follows: Find out the locations of the upstream and midstream of the supply chain through electronic maps, and record them as upstream locations and midstream locations respectively; Determine the vehicle type and transportation date of the transportation vehicle from the upstream location to the midstream location, input the vehicle type and transportation date into the navigation software, and query A first transportation routes from the upstream location to the midstream location; Query the tolls and travel times of A first transport routes, combine the A tolls and the A travel times to obtain A first transport suitability values, and record the travel time corresponding to the minimum of the first transport suitability values ​​as the first transport time.

4. The agricultural supply chain demand analysis and forecasting system based on digital drive according to claim 3 is characterized in that: The second transportation time is collected as follows: Find out the location of the downstream of the supply chain through electronic maps and record it as the downstream location; Determine the vehicle type and transportation date of the transportation vehicle from the midstream location to the downstream location, input the vehicle type and transportation date into the navigation software, and query B second transportation routes from the midstream location to the downstream location; The tolls and travel times of B second transport routes are queried, and after combining the B tolls and the B travel times, B second transport suitability values ​​are obtained, and the travel time corresponding to the minimum value of the second transport suitability value is recorded as the second transport time.

5. The agricultural supply chain demand analysis and forecasting system based on digital drive according to claim 4 is characterized in that: The method for formulating the demand period is as follows: The supply chain cycle is obtained by adding the first transportation time, the warehouse scheduling time, and the second transportation time; Mark C intervals of equal length in the supply chain cycle, and record the period between two adjacent intervals as the demand period, and obtain D demand periods; The D demand periods are numbered in ascending order from front to back along the timeline, so as to formulate D demand periods with serial numbers.

6. The agricultural supply chain demand analysis and forecasting system based on digital drive according to claim 5 is characterized in that: Demand analysis data includes time period attributes, total output value, unit price excess value and time period transaction rate; The method for collecting the excess unit price is as follows: Calculate the minimum and maximum durations for selling a unit of agricultural products in the upstream supply chain, add the minimum and maximum durations together and calculate the average to obtain the mean duration; Using a duration mean as the division standard, E interval distribution sub-intervals are divided in D demand periods; Query all the sales unit prices in the E sub-intervals of the upstream supply chain one by one, remove the maximum and minimum sales unit prices, and sort the remaining sales unit prices in ascending order from small to large to generate a unit price queue; If the number of sales unit prices in the unit price queue is an even number, add the two sales unit prices in the middle of the unit price queue and calculate the average to obtain the median sales unit price; If the number of sales unit prices in the unit price queue is an odd number, the sales unit price in the middle of the unit price queue is recorded as the median sales unit price; The first unit price, the median unit price, and the last unit price in the unit price queues of the E sub-intervals are added together and averaged to obtain the E unit price means; The E unit price means of D demand periods are added up one by one and averaged, and the average is subtracted from the standard unit price to obtain D unit price excess values.

7. The digital-driven agricultural supply chain demand analysis and forecasting system according to claim 6 is characterized in that: The collection method of time period transaction rate is: At the first moment of D demand periods, count the quantity of all agricultural products to be sold in the downstream of the supply chain to obtain D total supply values; At the last moment of D demand periods, count the quantity of all agricultural products to be sold in the downstream of the supply chain and obtain D supply surplus values; Set the sales loss rate of agricultural products, and multiply the sales loss rate by D total supply values ​​to obtain D time period loss values; Subtract the D supply surplus values ​​from the D time period loss values, and compare the difference with the D total supply values ​​to obtain the D time period transaction rate.

8. The digital-driven agricultural supply chain demand analysis and forecasting system according to claim 7 is characterized in that: After summarizing the D time period attributes, D total output values, D unit price excess values ​​and D time period transaction rates, D demand combination sets are generated; The training method of the demand forecasting model is: Collect multiple sets of demand combinations and supply demand values ​​in advance, convert peak hours into 1, flat hours into 2, and off-peak hours into 3; Convert multiple sets of demand combinations into multiple feature vectors using a sliding window method, convert supply demand values ​​into labels corresponding to the demand combination sets according to the sliding step length, one feature vector corresponds to one label, and form a set of training data, presetting the prediction time step length, sliding step length, and sliding window length; The feature vector is used as the input of the machine learning model, the last demand period in the sliding window length is taken as the target period, the supply demand value of the demand period after the time step of the target period is predicted as the output, the subsequent supply demand value of each training set is used as the prediction target, and the minimized sum of prediction errors is used as the training target. The machine learning model is trained to train a demand forecasting model.

9. The digital-driven agricultural supply chain demand analysis and forecasting system according to claim 8 is characterized in that: Comprehensive logistics data includes inbound quantity value, outbound quantity value, storage loss value and current balance value; The calculation method of effective storage value is: Add the current surplus value of the next demand period of the target period to the incoming quantity value, and then subtract them from the outgoing quantity value and the storage loss value to obtain the effective storage value of the next demand period of the target period; The demand and supply status includes the supply and demand imbalance status, the supply and demand balance status and the supply and demand excess status. The analysis methods for the supply and demand imbalance status, the supply and demand balance status and the supply and demand excess status are as follows: The effective storage value of the next demand period of the target period in the midstream of the supply chain is subtracted from the predicted supply demand value and the preset storage red line value to obtain the supply-demand difference; When the difference between supply and demand is greater than 0, the demand-supply status is recorded as excess supply and demand status; When the difference between supply and demand is equal to 0, the demand-supply state is recorded as the supply-demand balance state; When the supply-demand difference is less than 0, the demand-supply status is recorded as a supply-demand imbalance state.

10. The digital-driven agricultural supply chain demand analysis and forecasting system according to claim 9, characterized in that: Supply chain adjustment strategies include increasing supply, maintaining supply, and reducing supply. The methods for determining increasing supply, maintaining supply, and reducing supply are as follows: When the demand-supply state of the agricultural supply chain in the next demand period of the target period is in a supply-demand imbalance state, a strategy to increase supply is determined; When the demand-supply state of the agricultural supply chain in the next demand period of the target period is in a supply-demand balance state, a strategy for maintaining supply is determined; When the demand-supply status of the agricultural supply chain in the next demand period of the target period is in a state of excess supply and demand, a strategy to reduce the supply is determined.

Images