Food material distribution intelligent control method and system based on big data

Abstract

The invention provides a food material distribution intelligent control method and system based on big data, and the method comprises the steps: collecting the temperature and humidity of a distribution compartment in a distribution process in real time, collecting the position information and speed of a distribution vehicle, and obtaining the sensitive parameters of distributed food materials in the distribution vehicle; all reachable paths are obtained, a prediction model is used for predicting the remaining fresh-keeping time of the food materials distributed on all the reachable paths according to the current compartment internal environment parameters, the fresh-keeping state data of the food materials and the following path parameters, and the path parameters comprise weather and road conditions; screening out reachable paths of which the difference value between the remaining fresh-keeping time and the arrival time of the reachable paths does not exceed a preset time threshold value from the reachable paths; and obtaining the reachable path with the minimum cost calculated through the environment-sensitive supply chain cost model from the screened reachable paths. By means of the method and the corresponding system, it is ensured that the food materials are delivered at the lowest cost and keep the best freshness after delivery is completed, and the quality loss caused by transportation is reduced.

Description

Technical Field

[0001] The present invention proposes a method and system for intelligent control of food distribution based on big data, and relates to the technical field of artificial intelligence. Background Art

[0002] In the current development context, big data technology has gradually been introduced into the intelligent control system of food distribution to improve efficiency, reduce costs and enhance customer satisfaction. However, the existing intelligent control methods for food distribution still have some significant defects and challenges: the current distribution intelligent control system often lacks the ability to adjust in real time. Once traffic jams, equipment failures or changes in demand occur, the system is difficult to respond and adjust in time. Most systems only rely on plans made in advance, lack flexibility, and lead to insufficient resource utilization; many methods do not effectively predict and control the food loss caused by dynamic changes in environmental factors (such as temperature and humidity), lack of fine environmental monitoring, resulting in insufficient visibility of cold chain management and distribution status, which easily leads to food damage and waste; most existing systems rely on rule-driven decision-making models rather than data-driven intelligent analysis, and cannot fully utilize machine learning and artificial intelligence technologies to optimize route selection. Summary of the invention

[0003] The present invention provides a method and system for intelligent food distribution control based on big data to solve the above-mentioned problems: The present invention proposes a method for intelligently controlling food distribution based on big data, the method comprising: Real-time collection of temperature and humidity in the delivery compartment during the delivery process, collection of the location information and speed of the delivery vehicle, and acquisition of sensitive parameters of the delivered food in the delivery vehicle; Obtain all accessible paths, and use the prediction model to predict the remaining freshness time of the food delivered on all accessible paths based on the current internal environment parameters of the carriage, the freshness status data of the food, and the next path parameters. The path parameters include weather and road conditions, and select the accessible paths whose remaining freshness time and the arrival time of the accessible paths do not exceed the preset time threshold; The accessible path with the minimum cost calculated by the environmentally sensitive supply chain cost model is obtained from the screened accessible paths.

[0004] Furthermore, the temperature and humidity in the delivery compartment during the delivery process are collected in real time, the location information and speed of the delivery vehicle are collected, and sensitive parameters of the food delivered by the delivery compartment are obtained, including: The temperature and humidity of the compartment where food is delivered are collected through temperature sensors and humidity sensors respectively; Collect the location information and speed of delivery vehicles in real time through GPS devices and on-board OBD devices; Get the sensitive parameters of the food delivered by the delivery vehicle.

[0005] Furthermore, sensitive parameters of the food delivered in the delivery vehicle are obtained, including: Determine the range of environmental variables that need to be tested, including: setting the temperature gradient, setting the humidity gradient, and setting the test time period to select the target food sample; The target food samples are stored under different temperature and humidity conditions, and their freshness status is regularly recorded, wherein the freshness status includes weight, nutrient content, water loss rate, color change and texture change, wherein the color change is detected by a colorimeter, and the texture change is detected by a hardness meter; Record the temperature and humidity in the test environment in real time, take samples within a preset time, and save the collected data as a structured data table, which includes the fields: time, temperature, humidity, and freshness status; Performing preliminary data cleaning on the data stored in the structured data table, wherein the preliminary data cleaning includes: removing outliers and averaging repeated experimental data under the same conditions; Substitute the experimental data into the sensitive parameter model. Specifically, the sensitive parameter model is:

[0006] in, represents the rate of change of food freshness, k represents the process rate constant, T is the temperature in the actual operating environment, represents the temperature constant, H represents the relative humidity, the logarithm of the sensitive parameter model is taken, n represents the sensitivity coefficient of the food p loss to temperature, m represents the sensitivity coefficient of the food p loss to humidity, the sensitive parameter model is converted into a linear equation, and multiple linear regression is used to calculate k, n and m in the sensitive parameter model; The least square method is used to optimize the fitting error and verify the residual distribution of the fitting. When the residual distribution does not exceed a preset threshold, the final sensitive parameter model is obtained.

[0007] Furthermore, all accessible paths are obtained, and the prediction model is used to predict the remaining freshness time of the delivered food on all accessible paths according to the current internal environment parameters of the carriage, the freshness status data of the food, and the next path parameters. The path parameters include weather and road conditions, and the accessible paths are screened out, in which the difference between the remaining freshness time in the accessible path and the arrival time of the accessible path does not exceed the preset time threshold, including: Obtaining historical temperature and humidity data inside the carriage, and obtaining historical fresh-keeping status data of food, wherein the historical fresh-keeping status data includes weight, nutrient content, water loss rate, color change, texture change, sugar content, and food sensitive parameters, wherein the food sensitive parameters include a sensitivity coefficient to temperature and a sensitivity coefficient to humidity; Obtain forecast data of weather conditions in the next transportation route, process and clean the collected historical temperature data, historical humidity data, and historical freshness status data, remove outliers and noise, and extract important features from the cleaned historical data, including: the temperature change rate inside the carriage, the change rate of the freshness status data, and the temperature and relative humidity of the forecast data; Construct an LSTM model, combine the extracted important features into the LSTM model, divide the historical data into a training set, a validation set and a test set, train the LSTM model through the training set, so that the model input features are: the temperature change rate inside the compartment, the change rate of the fresh-keeping state data, the temperature and relative humidity of the forecast data, and the output is the remaining insurance time. The validation set verifies the accuracy of the LSTM model, and then adjusts the model parameters. After the adjustment, the performance of the LSTM model is evaluated through the test set; Obtain important features of the current food transportation, input the important features into the trained LSTM model, obtain the remaining shelf life of the food, and screen out the reachable paths whose difference between the remaining shelf life and the arrival time of the reachable paths does not exceed the preset time threshold.

[0008] Furthermore, the accessible path with the minimum cost calculated by the environmentally sensitive supply chain cost model is obtained from the screened accessible paths, including: An environmentally sensitive supply chain cost model is defined. Specifically, the environmentally sensitive supply chain cost model is:

[0009] Among them, R (x, y) represents the transportation volume from the starting point to the point (x, y), C (x, y) represents the cost per unit of transportation volume, and a represents the cost unit of loss. represents the loss rate of food p, represents the end time of a transport operation at a specific location (x,y) and for a specific product p, The constraints of the environmentally sensitive supply chain cost model are: , constraining each customer's location to be visited once and only once; ≤ the preset loss threshold, constraining the total loss not to exceed the acceptable loss; Each screened path is evaluated using the environmentally sensitive supply chain cost model, the total cost of each path is calculated, the costs of all paths are compared, and the path with the minimum cost is found.

[0010] The present invention proposes a food delivery intelligent control system based on big data, the system comprising: The collection module is used to collect the temperature and humidity in the delivery compartment in real time during the delivery process, collect the location information and speed of the delivery vehicle, and obtain sensitive parameters of the delivered ingredients in the delivery vehicle; The prediction and screening module is used to obtain all accessible paths, and use the prediction model to predict the remaining freshness time of the food delivered on all accessible paths according to the current internal environment parameters of the carriage, the freshness status data of the food, and the next path parameters. The path parameters include weather and road conditions, and screen out the accessible paths whose remaining freshness time and the arrival time of the accessible paths do not exceed the preset time threshold. The module for obtaining the minimum cost path is used to obtain the minimum cost accessible path calculated by the environmentally sensitive supply chain cost model from the screened accessible paths.

[0011] Furthermore, the acquisition module includes: The temperature and humidity collection module is used to collect the temperature and humidity of the compartment where food is delivered through a temperature sensor and a humidity sensor respectively; The vehicle information collection module is used to collect the location information and speed of the delivery vehicle in real time through the GPS device and the on-board OBD device; The module for obtaining sensitive parameters of ingredients is used to obtain sensitive parameters of ingredients delivered by the delivery vehicle.

[0012] Furthermore, the module for obtaining sensitive parameters of ingredients includes: The module for determining the range of environmental variables is used to determine the range of environmental variables to be tested, including: setting the temperature gradient, setting the humidity gradient, and setting the test time period to select the target food sample; A recording module is used to store target food samples under different temperature and humidity conditions and regularly record their freshness status, which includes weight, nutritional components, water loss rate, color change and texture change. The color change is detected by a colorimeter, and the texture change is detected by a hardness meter. The structured data module is used to record the temperature and humidity in the test environment in real time, perform sampling within a preset time, and save the collected data as a structured data table, which includes the fields: time, temperature, humidity, and preservation status; A preliminary cleaning module is used to perform preliminary data cleaning on the data stored in the structured data table, wherein the preliminary data cleaning includes: removing outliers and averaging repeated experimental data under the same conditions; The preliminary calculation module is used to substitute the experimental data into the sensitive parameter model. Specifically, the sensitive parameter model is:

[0013] in, represents the rate of change of food freshness, k represents the process rate constant, T is the temperature in the actual operating environment, represents the temperature constant, H represents the relative humidity, the logarithm of the sensitive parameter model is taken, n represents the sensitivity coefficient of the food p loss to temperature, m represents the sensitivity coefficient of the food p loss to humidity, the sensitive parameter model is converted into a linear equation, and multiple linear regression is used to calculate k, n and m in the sensitive parameter model; The fitting error module is used to optimize the fitting error using the least square method, verify the fitting residual distribution, and obtain the final sensitive parameter model when the residual distribution does not exceed a preset threshold.

[0014] Furthermore, the prediction and screening module includes: A data acquisition module is used to acquire historical temperature and humidity data inside the carriage, and acquire historical fresh-keeping status data of food materials, wherein the historical fresh-keeping status data includes weight, nutrient content, water loss rate, color change, texture change, sugar content, and food material sensitive parameters, wherein the food material sensitive parameters include a sensitivity coefficient to temperature and a sensitivity coefficient to humidity; The feature extraction module is used to obtain the forecast data of weather conditions in the next transportation route, process and clean the collected historical temperature data, historical humidity data, and historical freshness status data, remove outliers and noise, and extract important features from the cleaned historical data, including: the temperature change rate inside the carriage, the change rate of the freshness status data, and the temperature and relative humidity of the forecast data; A training model module is used to construct an LSTM model, combine the extracted important features into the LSTM model, divide the historical data into a training set, a validation set and a test set, and train the LSTM model through the training set, so that the model input features are: the temperature change rate inside the compartment, the change rate of the fresh-keeping state data, the temperature and relative humidity of the forecast data, and the output is the remaining insurance time. The validation set verifies the accuracy of the LSTM model, and then adjusts the model parameters. After the adjustment, the performance of the LSTM model is evaluated through the test set; The prediction module is used to obtain important features of the current food transportation, input the important features into the trained LSTM model, obtain the remaining shelf life of the food, and screen out the reachable paths in which the difference between the remaining shelf life and the arrival time of the reachable paths does not exceed a preset time threshold.

[0015] Furthermore, the module for obtaining the minimum cost path includes: An environmentally sensitive supply chain cost model is defined. Specifically, the environmentally sensitive supply chain cost model is:

[0016] Among them, R (x, y) represents the transportation volume from the starting point to the point (x, y), C (x, y) represents the cost per unit of transportation volume, and a represents the cost unit of loss. represents the loss rate of food p, represents the end time of a transport operation at a specific location (x,y) and for a specific product p, The constraints of the environmentally sensitive supply chain cost model are: , constraining each customer's location to be visited once and only once; ≤ the preset loss threshold, constraining the total loss not to exceed the acceptable loss; Each screened path is evaluated using the environmentally sensitive supply chain cost model, the total cost of each path is calculated, the costs of all paths are compared, and the path with the minimum cost is found.

[0017] The beneficial effects of the present invention are as follows: improving freshness, by accurately controlling and predicting the freshness-keeping time, ensuring that the food maintains the best freshness when the delivery is completed, and reducing the quality loss caused by transportation; reducing costs, optimizing the route selection not only considering time but also cost-effectiveness, ensuring that the quality target is achieved at the lowest cost during transportation; having dynamic adjustment capabilities, which can be flexibly adjusted according to real-time environmental changes and predictions, and improving the response speed of delivery route selection; enhancing decision-making efficiency, by integrating a variety of real-time and predicted data, improving the scientific nature and operability of delivery decisions, and improving logistics efficiency; improving customer satisfaction, ensuring that products are delivered to customers in the best condition and at the lowest cost, and improving customer satisfaction and loyalty; through the above technical scheme, the delivery process is not only optimized in terms of time and cost, but also effectively guarantees product quality, fully reflecting the advantages of intelligent logistics management. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a schematic diagram of a method for intelligent control of food delivery based on big data described in the present invention. DETAILED DESCRIPTION

[0019] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0020] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. The embodiments described are only a part of the embodiments of the present invention, rather than 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.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0022] One embodiment of the present invention provides a method for intelligently controlling food delivery based on big data, the method comprising: Real-time collection of temperature and humidity in the delivery compartment during the delivery process, collection of the location information and speed of the delivery vehicle, and acquisition of sensitive parameters of the delivered food in the delivery vehicle; Obtain all accessible paths, and use the prediction model to predict the remaining freshness time of the food delivered on all accessible paths based on the current internal environment parameters of the carriage, the freshness status data of the food, and the next path parameters. The path parameters include weather and road conditions, and select the accessible paths whose remaining freshness time and the arrival time of the accessible paths do not exceed the preset time threshold; The accessible path with the minimum cost calculated by the environmentally sensitive supply chain cost model is obtained from the screened accessible paths.

[0023] The working principle and effect of the above technical solution are: real-time collection of temperature and humidity data in the delivery compartment to monitor and record environmental conditions that affect the preservation of food, collection of location information and speed of the delivery vehicle to calculate transportation time and evaluate the feasibility of different routes, acquisition and recording of sensitive parameters of the delivered food, such as temperature and humidity sensitivity, and the freshness status of each ingredient; route information acquisition, obtain information on all accessible routes, including weather forecasts, road conditions and estimated driving time, which provide basic data for route selection; use deep learning or other prediction models to comprehensively analyze the current compartment environmental parameters (temperature, humidity), The freshness of ingredients and the environmental conditions of future routes are used to predict the remaining freshness time of ingredients delivered on each route; based on the current situation and predictions, each optional route is analyzed to calculate the freshness of ingredients at the end of delivery; the estimated arrival time of each achievable route and the remaining freshness time of ingredients are compared to select the route where the difference between the freshness time and the arrival time does not exceed the preset threshold. This step ensures that the ingredients still meet the quality standards when they reach the destination; cost optimization, applying the environmentally sensitive supply chain cost model to the selected routes, calculating the comprehensive transportation cost, including logistics costs and possible loss costs; and selecting the achievable route with the lowest comprehensive cost as the final transportation plan. Improve freshness by accurately controlling and predicting the shelf life to ensure that ingredients remain at optimal freshness upon delivery and reduce quality losses caused by transportation; reduce costs by optimizing route selection to consider not only time but also cost-effectiveness to ensure that quality goals are achieved at the lowest cost during transportation; have dynamic adjustment capabilities and can flexibly adjust according to real-time environmental changes and predictions to improve the response speed of delivery route selection; enhance decision-making efficiency by integrating a variety of real-time and predictive data to improve the scientific nature and operability of delivery decisions and improve logistics efficiency; improve customer satisfaction by ensuring that products are delivered to customers in the best condition and at the lowest cost, thereby improving customer satisfaction and loyalty; through the above technical solutions, the delivery process is not only optimized in terms of time and cost, but also effectively guarantees product quality, fully reflecting the advantages of intelligent logistics management.

[0024] In one embodiment of the present invention, the temperature and humidity in the delivery compartment during the delivery process are collected in real time, the location information and speed of the delivery vehicle are collected, and sensitive parameters of the food delivered by the delivery compartment are obtained, including: The temperature and humidity of the compartment where food is delivered are collected through temperature sensors and humidity sensors respectively; Collect the location information and speed of delivery vehicles in real time through GPS devices and on-board OBD devices; Get the sensitive parameters of the food delivered by the delivery vehicle.

[0025] The working principle and effect of the above technical solution are as follows: through the coordinated work of multiple sensors and devices, the food preservation and logistics efficiency during the delivery process are monitored and optimized in real time; the temperature sensor is installed inside the delivery compartment to monitor the temperature changes in the compartment in real time to ensure that the temperature is maintained within a range suitable for food preservation; the humidity sensor is also installed in the compartment to monitor the humidity level in real time. Some food is very sensitive to humidity, and maintaining appropriate humidity can effectively extend the shelf life of the food; GPS devices provide high-precision geographic location information and ensure the accuracy of the route by tracking the vehicle's route and location in real time. The on-board OBD device (On-Board Diagnostics) is connected to the vehicle's OBD interface to collect data on the vehicle's speed and other dynamic conditions in real time, which helps to understand the vehicle's operating status and detect whether there are any problems that may affect transportation efficiency. The database stores and reads sensitive parameters of each delivered ingredient, including the optimal values ​​of temperature and humidity, as well as critical conditions for preservation. These parameters provide specific guidance for environmental control and help predict the potential risk of deterioration of ingredients during the delivery process. Improve the quality of freshness preservation. Through continuous monitoring of temperature and humidity, the stability of the vehicle environment is ensured, and the freshness and shelf life of the food are maximized. Real-time monitoring and response. Combined with the vehicle dynamic information provided by devices such as GPS and OBD, logistics managers can adjust the distribution strategy in real time, such as changing the route or strengthening cold chain measures midway, so as to respond more quickly to unexpected situations such as traffic jams or equipment failures. All collected data can be recorded and analyzed, providing detailed data support for subsequent optimization, making the logistics process more transparent and monitorable. Improve operational efficiency. By knowing and preparing in advance for possible temperature and humidity fluctuations or logistics delays, adjustments and repair measures can be reduced. Unnecessary losses and costs can be reduced, thereby improving overall operational efficiency. Real-time data collection and monitoring can also help identify loopholes or problems in the logistics process (such as vehicle failures or sensor errors) so that corrective measures can be taken quickly. Through the above solutions, enterprises can not only optimize their supply chain and logistics management, but also significantly improve product quality and customer satisfaction. This high-precision, real-time monitoring technology solution is an important part of modern intelligent transportation systems.

[0026] In one embodiment of the present invention, obtaining sensitive parameters of food delivered by a delivery vehicle includes: Determine the range of environmental variables that need to be tested, including: setting the temperature gradient, setting the humidity gradient, and setting the test time period to select the target food sample; The target food samples are stored under different temperature and humidity conditions, and their freshness status is regularly recorded, wherein the freshness status includes weight, nutrient content, water loss rate, color change and texture change, wherein the color change is detected by a colorimeter, and the texture change is detected by a hardness meter; Record the temperature and humidity in the test environment in real time, take samples within a preset time, and save the collected data as a structured data table, which includes the fields: time, temperature, humidity, and freshness status; Performing preliminary data cleaning on the data stored in the structured data table, wherein the preliminary data cleaning includes: removing outliers and averaging repeated experimental data under the same conditions; Substitute the experimental data into the sensitive parameter model. Specifically, the sensitive parameter model is:

[0027] in, represents the rate of change of food freshness, k represents the process rate constant, T is the temperature in the actual operating environment, represents the temperature constant, H represents the relative humidity, the logarithm of the sensitive parameter model is taken, n represents the sensitivity coefficient of the food p loss to temperature, m represents the sensitivity coefficient of the food p loss to humidity, the sensitive parameter model is converted into a linear equation, and multiple linear regression is used to calculate k, n and m in the sensitive parameter model; The least square method is used to optimize the fitting error and verify the residual distribution of the fitting. When the residual distribution does not exceed a preset threshold, the final sensitive parameter model is obtained.

[0028] The working principle and effect of the above technical scheme are as follows: the above technical scheme is used to quantify the freshness changes of food under different temperature and humidity conditions, prepare the test environment, set the variable range, and determine different environmental conditions by adjusting the temperature gradient, humidity gradient and test time period to test the preservation response of food samples; store the selected food samples under these conditions, and use tools such as colorimeters and hardness meters to monitor and record the color and texture changes of the food, and record the preservation status such as weight, nutrients and moisture loss rate; monitor and record temperature and humidity data in real time, as well as the preservation status of the food at the above time points, and store these data in a structured data table format; clean the collected data, remove outliers, and average repeated data under the same experimental conditions to reduce noise; substitute specific sensitive parameter model equations, use multiple linear regression, convert the equations into linear equations by taking the logarithm of the equations, calculate the model parameters, use the least squares method to optimize the fitting error, verify the distribution of the fitting residuals, and ensure that the residuals are within an acceptable preset threshold range to ensure the accuracy of the model. Generating accurate sensitive parameter models can better predict the freshness changes of food ingredients under specific environmental conditions, thereby optimizing storage and transportation strategies; through quantitative analysis of freshness changes, it can assist in formulating temperature and humidity control rules to extend the shelf life and improve product quality; it can help logistics and food supply chain managers optimize cold chain management and reduce food loss and economic losses caused by improper environment; it not only improves the understanding of the food preservation process, but also strengthens the possibility of scientific management and data-driven optimization measures.

[0029] In one embodiment of the present invention, all accessible paths are obtained, and a prediction model is used to predict the remaining freshness time of the food delivered on all accessible paths according to the current internal environment parameters of the carriage, the freshness state data of the food, and the next path parameters, where the path parameters include weather and road conditions, and the accessible paths are screened out, where the difference between the remaining freshness time in the accessible paths and the arrival time of the accessible paths does not exceed a preset time threshold, including: Obtaining historical temperature and humidity data inside the carriage, and obtaining historical fresh-keeping status data of food, wherein the historical fresh-keeping status data includes weight, nutrient content, water loss rate, color change, texture change, sugar content, and food sensitive parameters, wherein the food sensitive parameters include a sensitivity coefficient to temperature and a sensitivity coefficient to humidity; Obtain forecast data of weather conditions in the next transportation route, process and clean the collected historical temperature data, historical humidity data, and historical freshness status data, remove outliers and noise, and extract important features from the cleaned historical data, including: the temperature change rate inside the carriage, the change rate of the freshness status data, and the temperature and relative humidity of the forecast data; Construct an LSTM model, combine the extracted important features into the LSTM model, divide the historical data into a training set, a validation set and a test set, train the LSTM model through the training set, so that the model input features are: the temperature change rate inside the compartment, the change rate of the fresh-keeping state data, the temperature and relative humidity of the forecast data, and the output is the remaining insurance time. The validation set verifies the accuracy of the LSTM model, and then adjusts the model parameters. After the adjustment, the performance of the LSTM model is evaluated through the test set; Obtain important features of the current food transportation, input the important features into the trained LSTM model, obtain the remaining shelf life of the food, and screen out the reachable paths whose difference between the remaining shelf life and the arrival time of the reachable paths does not exceed the preset time threshold.

[0030] The working principle and effect of the above technical solution are: data collection and organization, acquisition of historical data, collection of historical temperature and humidity data inside the carriage, and historical freshness status data of food ingredients. These data include multiple attributes, such as weight, nutrients, water loss, color change, texture change, sugar content, and sensitivity coefficients to temperature and humidity; obtain weather forecast data for the upcoming transportation route, including air temperature and relative humidity; remove outliers and noise in the data to ensure data quality, and extract important features from the cleaned data, such as the rate of change of temperature inside the carriage, the rate of change of freshness status data, and the air temperature and relative humidity in the forecast data. These features will provide input variables for the model; build an LSTM model, and use the extracted important features to build an LSTM model to capture the dynamic changes and long-term dependencies of time series data; The historical data is divided into training set, validation set and test set for model training, parameter adjustment and performance evaluation; the LSTM model is trained, and the input features are the temperature change rate, the freshness state change rate, the predicted temperature and the relative humidity; the output is the remaining freshness time; the accuracy of the model is tested through the validation set, and the model parameters are adjusted according to the results, and the final model performance is evaluated using the test set to ensure that it can work reliably in practical applications; real-time important features are obtained during the current transportation process, and these features are input into the trained LSTM model to predict the remaining freshness time of the food; and the reachable paths whose difference between the remaining freshness time and the path arrival time does not exceed the preset time threshold are screened out. Precise freshness prediction: By using the LSTM model to accurately predict the remaining freshness of food under different environmental conditions, the control of food freshness during the delivery process is improved; the delivery route is optimized, and the most appropriate delivery route is selected based on the information of the remaining freshness time to ensure that food is delivered while maintaining the best freshness; logistics efficiency and customer satisfaction are improved: by optimizing route selection and transportation conditions management, food loss is reduced, delivery efficiency is improved, and customer satisfaction is improved; dynamic adjustment and real-time decision support: during transportation, transportation strategies are adjusted according to real-time data to respond to sudden changes and enhance the flexibility of logistics management. Provide a foundation for supply chain optimization: the insights and data analysis obtained through this technical solution can provide support and basis for further optimization of the cold chain supply chain; by integrating advanced data analysis technology and predictive models, it can not only effectively manage the freshness of food during the delivery process, but also greatly improve the overall efficiency and responsiveness of transportation and logistics management.

[0031] In one embodiment of the present invention, obtaining a reachable path with the minimum cost calculated by an environmentally sensitive supply chain cost model from among the screened reachable paths includes: An environmentally sensitive supply chain cost model is defined. Specifically, the environmentally sensitive supply chain cost model is:

[0032] Among them, R (x, y) represents the transportation volume from the starting point to the point (x, y), C (x, y) represents the cost per unit of transportation volume, and a represents the cost unit of loss. represents the loss rate of food p, represents the end time of a transport operation at a specific location (x,y) and for a specific product p, The constraints of the environmentally sensitive supply chain cost model are: , constraining each customer's location to be visited once and only once; ≤ the preset loss threshold, constraining the total loss not to exceed the acceptable loss; Each screened path is evaluated using the environmentally sensitive supply chain cost model, the total cost of each path is calculated, the costs of all paths are compared, and the path with the minimum cost is found.

[0033] The working principle and effect of the above technical solution are as follows: when delivering food to customers, the delivery order will affect the quality of certain products (such as temperature-sensitive fresh products). In order to better deal with this situation, the environmentally sensitive supply chain cost model is designed to be more refined. The goal is to minimize the direct costs of the transportation route. The route selection and the order in which the ingredients are delivered to customers are not only related to the transportation distance, but also affect the fuel use and driving time, continuous loss The term considers the effects of temperature and humidity on food preservation, and evaluates the change in the state of food through the derivative of the loss rate. The freshness of food directly affects the quality and safety of the final product. Optimizing this part can reduce the risk of food spoilage due to environmental conditions; reduce economic losses, reduced customer satisfaction, and possible return costs caused by food spoilage; by better controlling losses, food waste can be reduced and sustainable development can be promoted; the loss of fresh food does not occur instantaneously, but accumulates during transportation. The integral term can naturally express this cumulative effect by calculating the small losses at each moment and summing them up to get the total loss; environmental parameters (temperature and humidity) change over time, and the loss rate also changes accordingly. The integral formula can clearly express this time dependence and accurately reflect the impact of environmental conditions at different time points on product losses; in actual transportation, such as cold chain transportation, temperature and humidity are not constant. The use of integrals can dynamically evaluate the impact of these environmental parameters on products over time, so that the model can more accurately simulate the actual situation; through integration, It can accurately accumulate the loss at each moment and reflect the precise loss situation; it can flexibly adjust the model according to the temperature and humidity information at different time points, without assuming that the environmental parameters are constant; it is suitable for transportation simulation and optimization in different regions and under different weather conditions; assuming that a specific product p is particularly sensitive to temperature, the loss rate of the product will increase significantly with every 1 degree increase in temperature. If a heat wave occurs during transportation, the loss of this product will increase sharply over a period of time. In this case, the integral formula can well capture the dynamic changes in the loss rate and reflect the actual total loss. Using this integral term to calculate the loss can carefully and accurately evaluate the loss of the product caused by changes in environmental factors such as time, temperature, and humidity during the entire logistics process. It can very effectively simulate and optimize the logistics distribution process in a complex actual environment and improve the accuracy and scientificity of the model. Minimize overall costs, by integrating all relevant costs, to achieve comprehensive optimization of the total cost of the supply chain, so that enterprises can specifically identify and reduce unnecessary expenses in the logistics process; loss control, evaluate and minimize the quality loss costs caused by environmental factors such as temperature and humidity, improve product freshness and customer satisfaction; by supporting large-scale data analysis and the integration of machine learning models (such as LSTM), make decisions more accurate and predictive; improve supply chain efficiency, by comparing the total cost of different transportation options, it helps to reasonably allocate logistics resources and select the optimal path; by considering environmental sensitivity, it can quickly respond to market changes, such as changes in climate conditions and changes in customer demand, etc.; through clear mathematical structures and parametric design, it provides a method for integrating complex data analysis and operation optimization in logistics and supply chain management. This not only enhances the responsiveness and decision-making efficiency of enterprises in market competition, but also improves the cost-effectiveness of the entire supply chain and promotes sustainable development.

[0034] One embodiment of the present invention is a food delivery intelligent control system based on big data, the system comprising: The collection module is used to collect the temperature and humidity in the delivery compartment in real time during the delivery process, collect the location information and speed of the delivery vehicle, and obtain sensitive parameters of the delivered ingredients in the delivery vehicle; The prediction and screening module is used to obtain all accessible paths, and use the prediction model to predict the remaining freshness time of the food delivered on all accessible paths according to the current internal environment parameters of the carriage, the freshness status data of the food, and the next path parameters. The path parameters include weather and road conditions, and screen out the accessible paths whose remaining freshness time and the arrival time of the accessible paths do not exceed the preset time threshold. The module for obtaining the minimum cost path is used to obtain the minimum cost accessible path calculated by the environmentally sensitive supply chain cost model from the screened accessible paths.

[0035] The working principle and effect of the above technical solution are: real-time collection of temperature and humidity data in the delivery compartment to monitor and record environmental conditions that affect the preservation of food, collection of location information and speed of the delivery vehicle to calculate transportation time and evaluate the feasibility of different routes, acquisition and recording of sensitive parameters of the delivered food, such as temperature and humidity sensitivity, and the freshness status of each ingredient; route information acquisition, obtain information on all accessible routes, including weather forecasts, road conditions and estimated driving time, which provide basic data for route selection; use deep learning or other prediction models to comprehensively analyze the current compartment environmental parameters (temperature, humidity), The freshness of ingredients and the environmental conditions of future routes are used to predict the remaining freshness time of ingredients delivered on each route; based on the current situation and predictions, each optional route is analyzed to calculate the freshness of ingredients at the end of delivery; the estimated arrival time of each achievable route and the remaining freshness time of ingredients are compared to select the route where the difference between the freshness time and the arrival time does not exceed the preset threshold. This step ensures that the ingredients still meet the quality standards when they reach the destination; cost optimization, applying the environmentally sensitive supply chain cost model to the selected routes, calculating the comprehensive transportation cost, including logistics costs and possible loss costs; and selecting the achievable route with the lowest comprehensive cost as the final transportation plan. Improve freshness by accurately controlling and predicting the shelf life to ensure that ingredients remain at optimal freshness upon delivery and reduce quality losses caused by transportation; reduce costs by optimizing route selection to consider not only time but also cost-effectiveness to ensure that quality goals are achieved at the lowest cost during transportation; have dynamic adjustment capabilities and can flexibly adjust according to real-time environmental changes and predictions to improve the response speed of delivery route selection; enhance decision-making efficiency by integrating a variety of real-time and predictive data to improve the scientific nature and operability of delivery decisions and improve logistics efficiency; improve customer satisfaction by ensuring that products are delivered to customers in the best condition and at the lowest cost, thereby improving customer satisfaction and loyalty; through the above technical solutions, the delivery process is not only optimized in terms of time and cost, but also effectively guarantees product quality, fully reflecting the advantages of intelligent logistics management.

[0036] In one embodiment of the present invention, the acquisition module includes: The temperature and humidity collection module is used to collect the temperature and humidity of the compartment where food is delivered through a temperature sensor and a humidity sensor respectively; The vehicle information collection module is used to collect the location information and speed of the delivery vehicle in real time through the GPS device and the on-board OBD device; The module for obtaining sensitive parameters of ingredients is used to obtain sensitive parameters of ingredients delivered by the delivery vehicle.

[0037] The working principle and effect of the above technical solution are as follows: through the coordinated work of multiple sensors and devices, the food preservation and logistics efficiency during the delivery process are monitored and optimized in real time; the temperature sensor is installed inside the delivery compartment to monitor the temperature changes in the compartment in real time to ensure that the temperature is maintained within a range suitable for food preservation; the humidity sensor is also installed in the compartment to monitor the humidity level in real time. Some food is very sensitive to humidity, and maintaining appropriate humidity can effectively extend the shelf life of the food; GPS devices provide high-precision geographic location information and ensure the accuracy of the route by tracking the vehicle's route and location in real time. The on-board OBD device (On-Board Diagnostics) is connected to the vehicle's OBD interface to collect data on the vehicle's speed and other dynamic conditions in real time, which helps to understand the vehicle's operating status and detect whether there are any problems that may affect transportation efficiency. The database stores and reads sensitive parameters of each delivered ingredient, including the optimal values ​​of temperature and humidity, as well as critical conditions for preservation. These parameters provide specific guidance for environmental control and help predict the potential risk of deterioration of ingredients during the delivery process. Improve the quality of freshness preservation. Through continuous monitoring of temperature and humidity, the stability of the vehicle environment is ensured, and the freshness and shelf life of the food are maximized. Real-time monitoring and response. Combined with the vehicle dynamic information provided by devices such as GPS and OBD, logistics managers can adjust the distribution strategy in real time, such as changing the route or strengthening cold chain measures midway, so as to respond more quickly to unexpected situations such as traffic jams or equipment failures. All collected data can be recorded and analyzed, providing detailed data support for subsequent optimization, making the logistics process more transparent and monitorable. Improve operational efficiency. By knowing and preparing in advance for possible temperature and humidity fluctuations or logistics delays, adjustments and repair measures can be reduced. Unnecessary losses and costs can be reduced, thereby improving overall operational efficiency. Real-time data collection and monitoring can also help identify loopholes or problems in the logistics process (such as vehicle failures or sensor errors) so that corrective measures can be taken quickly. Through the above solutions, enterprises can not only optimize their supply chain and logistics management, but also significantly improve product quality and customer satisfaction. This high-precision, real-time monitoring technology solution is an important part of modern intelligent transportation systems.

[0038] In one embodiment of the present invention, the module for obtaining sensitive parameters of ingredients includes: The module for determining the range of environmental variables is used to determine the range of environmental variables to be tested, including: setting the temperature gradient, setting the humidity gradient, and setting the test time period to select the target food sample; A recording module is used to store target food samples under different temperature and humidity conditions and regularly record their freshness status, which includes weight, nutritional components, water loss rate, color change and texture change. The color change is detected by a colorimeter, and the texture change is detected by a hardness meter. The structured data module is used to record the temperature and humidity in the test environment in real time, perform sampling within a preset time, and save the collected data as a structured data table, which includes the fields: time, temperature, humidity, and preservation status; A preliminary cleaning module is used to perform preliminary data cleaning on the data stored in the structured data table, wherein the preliminary data cleaning includes: removing outliers and averaging repeated experimental data under the same conditions; The preliminary calculation module is used to substitute the experimental data into the sensitive parameter model. Specifically, the sensitive parameter model is:

[0039] in, represents the rate of change of food freshness, k represents the process rate constant, T is the temperature in the actual operating environment, represents the temperature constant, H represents the relative humidity, the logarithm of the sensitive parameter model is taken, n represents the sensitivity coefficient of the food p loss to temperature, m represents the sensitivity coefficient of the food p loss to humidity, the sensitive parameter model is converted into a linear equation, and multiple linear regression is used to calculate k, n and m in the sensitive parameter model; The fitting error module is used to optimize the fitting error using the least square method, verify the fitting residual distribution, and obtain the final sensitive parameter model when the residual distribution does not exceed a preset threshold.

[0040] The working principle and effect of the above technical scheme are as follows: the above technical scheme is used to quantify the freshness changes of food under different temperature and humidity conditions, prepare the test environment, set the variable range, and determine different environmental conditions by adjusting the temperature gradient, humidity gradient and test time period to test the preservation response of food samples; store the selected food samples under these conditions, and use tools such as colorimeters and hardness meters to monitor and record the color and texture changes of the food, and record the preservation status such as weight, nutrients and moisture loss rate; monitor and record temperature and humidity data in real time, as well as the preservation status of the food at the above time points, and store these data in a structured data table format; clean the collected data, remove outliers, and average repeated data under the same experimental conditions to reduce noise; substitute specific sensitive parameter model equations, use multiple linear regression, convert the equations into linear equations by taking the logarithm of the equations, calculate the model parameters, use the least squares method to optimize the fitting error, verify the distribution of the fitting residuals, and ensure that the residuals are within an acceptable preset threshold range to ensure the accuracy of the model. Generating accurate sensitive parameter models can better predict the freshness changes of food ingredients under specific environmental conditions, thereby optimizing storage and transportation strategies; through quantitative analysis of freshness changes, it can assist in formulating temperature and humidity control rules to extend the shelf life and improve product quality; it can help logistics and food supply chain managers optimize cold chain management and reduce food loss and economic losses caused by improper environment; it not only improves the understanding of the food preservation process, but also strengthens the possibility of scientific management and data-driven optimization measures.

[0041] In one embodiment of the present invention, the prediction and screening module comprises: A data acquisition module is used to acquire historical temperature and humidity data inside the carriage, and acquire historical fresh-keeping status data of food materials, wherein the historical fresh-keeping status data includes weight, nutrient content, water loss rate, color change, texture change, sugar content, and food material sensitive parameters, wherein the food material sensitive parameters include a sensitivity coefficient to temperature and a sensitivity coefficient to humidity; The feature extraction module is used to obtain the forecast data of weather conditions in the next transportation route, process and clean the collected historical temperature data, historical humidity data, and historical freshness status data, remove outliers and noise, and extract important features from the cleaned historical data, including: the temperature change rate inside the carriage, the change rate of the freshness status data, and the temperature and relative humidity of the forecast data; A training model module is used to construct an LSTM model, combine the extracted important features into the LSTM model, divide the historical data into a training set, a validation set and a test set, and train the LSTM model through the training set, so that the model input features are: the temperature change rate inside the compartment, the change rate of the fresh-keeping state data, the temperature and relative humidity of the forecast data, and the output is the remaining insurance time. The validation set verifies the accuracy of the LSTM model, and then adjusts the model parameters. After the adjustment, the performance of the LSTM model is evaluated through the test set; The prediction module is used to obtain important features of the current food transportation, input the important features into the trained LSTM model, obtain the remaining shelf life of the food, and screen out the reachable paths in which the difference between the remaining shelf life and the arrival time of the reachable paths does not exceed a preset time threshold.

[0042] The working principle and effect of the above technical solution are: data collection and organization, acquisition of historical data, collection of historical temperature and humidity data inside the carriage, and historical freshness status data of food ingredients. These data include multiple attributes, such as weight, nutrients, water loss, color change, texture change, sugar content, and sensitivity coefficients to temperature and humidity; obtain weather forecast data for the upcoming transportation route, including air temperature and relative humidity; remove outliers and noise in the data to ensure data quality, and extract important features from the cleaned data, such as the rate of change of temperature inside the carriage, the rate of change of freshness status data, and the air temperature and relative humidity in the forecast data. These features will provide input variables for the model; build an LSTM model, and use the extracted important features to build an LSTM model to capture the dynamic changes and long-term dependencies of time series data; The historical data is divided into training set, validation set and test set for model training, parameter adjustment and performance evaluation; the LSTM model is trained, and the input features are the temperature change rate, the freshness state change rate, the predicted temperature and the relative humidity; the output is the remaining freshness time; the accuracy of the model is tested through the validation set, and the model parameters are adjusted according to the results, and the final model performance is evaluated using the test set to ensure that it can work reliably in practical applications; real-time important features are obtained during the current transportation process, and these features are input into the trained LSTM model to predict the remaining freshness time of the food; and the reachable paths whose difference between the remaining freshness time and the path arrival time does not exceed the preset time threshold are screened out. Precise freshness prediction: By using the LSTM model to accurately predict the remaining freshness of food under different environmental conditions, the control of food freshness during the delivery process is improved; the delivery route is optimized, and the most appropriate delivery route is selected based on the information of the remaining freshness time to ensure that food is delivered while maintaining the best freshness; logistics efficiency and customer satisfaction are improved: by optimizing route selection and transportation conditions management, food loss is reduced, delivery efficiency is improved, and customer satisfaction is improved; dynamic adjustment and real-time decision support: during transportation, transportation strategies are adjusted according to real-time data to respond to sudden changes and enhance the flexibility of logistics management. Provide a foundation for supply chain optimization: the insights and data analysis obtained through this technical solution can provide support and basis for further optimization of the cold chain supply chain; by integrating advanced data analysis technology and predictive models, it can not only effectively manage the freshness of food during the delivery process, but also greatly improve the overall efficiency and responsiveness of transportation and logistics management.

[0043] In one embodiment of the present invention, the module for obtaining the minimum cost path includes: An environmentally sensitive supply chain cost model is defined. Specifically, the environmentally sensitive supply chain cost model is:

[0044] Among them, R (x, y) represents the transportation volume from the starting point to the point (x, y), C (x, y) represents the cost per unit of transportation volume, and a represents the cost unit of loss. represents the loss rate of food p, represents the end time of a transport operation at a specific location (x,y) and for a specific product p, The constraints of the environmentally sensitive supply chain cost model are: , constraining each customer's location to be visited once and only once; ≤ the preset loss threshold, constraining the total loss not to exceed the acceptable loss; Each screened path is evaluated using the environmentally sensitive supply chain cost model, the total cost of each path is calculated, the costs of all paths are compared, and the path with the minimum cost is found.

[0045] The working principle and effect of the above technical solution are as follows: when delivering food to customers, the delivery order will affect the quality of certain products (such as temperature-sensitive fresh products). In order to better deal with this situation, the environmentally sensitive supply chain cost model is designed to be more refined. The goal is to minimize the direct costs of the transportation route. The route selection and the order in which the ingredients are delivered to customers are not only related to the transportation distance, but also affect the fuel use and driving time, continuous loss The term considers the effects of temperature and humidity on food preservation, and evaluates the change in the state of food through the derivative of the loss rate. The freshness of food directly affects the quality and safety of the final product. Optimizing this part can reduce the risk of food spoilage due to environmental conditions; reduce economic losses, reduced customer satisfaction, and possible return costs caused by food spoilage; by better controlling losses, food waste can be reduced and sustainable development can be promoted; the loss of fresh food does not occur instantaneously, but accumulates during transportation. The integral term can naturally express this cumulative effect, and the total loss is obtained by calculating the small losses at each moment and summing them up; environmental parameters (temperature and humidity) change over time, and the loss rate changes accordingly. The integral formula can clearly express this time dependence and accurately reflect the impact of environmental conditions at different time points on product losses; in actual transportation, such as cold chain transportation, temperature and humidity are not constant. The use of integrals can dynamically evaluate the impact of these environmental parameters on products over time, so that the model can more accurately simulate the actual situation; through integration, it can be accurately Accurately accumulate the loss at each moment to reflect the precise loss situation; the model can be flexibly adjusted according to the temperature and humidity information at different time points, without assuming that the environmental parameters are constant; it is suitable for transportation simulation and optimization in different regions and under different weather conditions; assuming that a specific product p is particularly sensitive to temperature, the loss rate of the product will increase significantly with every 1 degree increase in temperature. If a heat wave is encountered on the transportation route during transportation, the loss of this product will increase sharply over a period of time. In this case, the integral formula can well capture the dynamic changes in the loss rate and reflect the actual total loss. Using this integral term to calculate the loss can carefully and accurately evaluate the loss of the product caused by changes in environmental factors such as time, temperature, and humidity during the entire logistics process. It can very effectively simulate and optimize the logistics distribution process in a complex actual environment and improve the accuracy and scientificity of the model. Minimize overall costs, by integrating all relevant costs, to achieve comprehensive optimization of the total cost of the supply chain, so that enterprises can specifically identify and reduce unnecessary expenses in the logistics process; loss control, evaluate and minimize the quality loss costs caused by environmental factors such as temperature and humidity, improve product freshness and customer satisfaction; by supporting large-scale data analysis and the integration of machine learning models (such as LSTM), make decisions more accurate and predictive; improve supply chain efficiency, by comparing the total cost of different transportation options, it helps to reasonably allocate logistics resources and select the optimal path; by considering environmental sensitivity, it can quickly respond to market changes, such as changes in climate conditions and changes in customer demand, etc.; through clear mathematical structures and parametric design, it provides a method for integrating complex data analysis and operation optimization in logistics and supply chain management. This not only enhances the responsiveness and decision-making efficiency of enterprises in market competition, but also improves the cost-effectiveness of the entire supply chain and promotes sustainable development.

[0046] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for intelligent control of food distribution based on big data, characterized in that: The method comprises: Real-time collection of temperature and humidity in the delivery compartment during the delivery process, collection of the location information and speed of the delivery vehicle, and acquisition of sensitive parameters of the delivered food in the delivery vehicle; Obtain all accessible paths, and use the prediction model to predict the remaining freshness time of the food delivered on all accessible paths based on the current internal environment parameters of the carriage, the freshness status data of the food, and the next path parameters. The path parameters include weather and road conditions, and select the accessible paths whose remaining freshness time and the arrival time of the accessible paths do not exceed the preset time threshold; The accessible path with the minimum cost calculated by the environmentally sensitive supply chain cost model is obtained from the screened accessible paths.

2. According to the method of intelligent food distribution control based on big data in claim 1, it is characterized in that: Real-time collection of temperature and humidity in the delivery compartment during the delivery process, collection of the location information and speed of the delivery vehicle, and acquisition of sensitive parameters of the food delivered by the delivery compartment, including: The temperature and humidity of the compartment where food is delivered are collected through temperature sensors and humidity sensors respectively; Collect the location information and speed of delivery vehicles in real time through GPS devices and on-board OBD devices; Get the sensitive parameters of the food delivered by the delivery vehicle.

3. According to the method of intelligent control of food distribution based on big data as claimed in claim 2, it is characterized in that: Get the sensitive parameters of the food delivered by the delivery vehicle, including: Determine the range of environmental variables that need to be tested, including: setting the temperature gradient, setting the humidity gradient, and setting the test time period to select the target food sample; The target food samples are stored under different temperature and humidity conditions, and their freshness status is regularly recorded, wherein the freshness status includes weight, nutrient content, water loss rate, color change and texture change, wherein the color change is detected by a colorimeter, and the texture change is detected by a hardness meter; Record the temperature and humidity in the test environment in real time, take samples within a preset time, and save the collected data as a structured data table, which includes the fields: time, temperature, humidity, and freshness status; Performing preliminary data cleaning on the data stored in the structured data table, wherein the preliminary data cleaning includes: removing outliers and averaging repeated experimental data under the same conditions; Substitute the experimental data into the sensitive parameter model. Specifically, the sensitive parameter model is: in, represents the rate of change of food freshness, k represents the process rate constant, T is the temperature in the actual operating environment, represents the temperature constant, H represents the relative humidity, the logarithm of the sensitive parameter model is taken, n represents the sensitivity coefficient of the food p loss to temperature, m represents the sensitivity coefficient of the food p loss to humidity, the sensitive parameter model is converted into a linear equation, and multiple linear regression is used to calculate k, n and m in the sensitive parameter model; The least square method is used to optimize the fitting error and verify the residual distribution of the fitting. When the residual distribution does not exceed a preset threshold, the final sensitive parameter model is obtained.

4. According to the method of intelligent food distribution control based on big data in claim 1, it is characterized in that: Obtain all accessible paths, and use the prediction model to predict the remaining freshness time of the delivered ingredients on all accessible paths based on the current internal environment parameters of the carriage, the freshness status data of the ingredients, and the next path parameters. The path parameters include weather and road conditions, and select the accessible paths whose remaining freshness time and the arrival time of the accessible paths do not exceed the preset time threshold, including: Obtaining historical temperature and humidity data inside the carriage, and obtaining historical fresh-keeping status data of food, wherein the historical fresh-keeping status data includes weight, nutrient content, water loss rate, color change, texture change, sugar content, and food sensitive parameters, wherein the food sensitive parameters include a sensitivity coefficient to temperature and a sensitivity coefficient to humidity; Obtain forecast data of weather conditions in the next transportation route, process and clean the collected historical temperature data, historical humidity data, and historical freshness status data, remove outliers and noise, and extract important features from the cleaned historical data, including: the temperature change rate inside the carriage, the change rate of the freshness status data, and the temperature and relative humidity of the forecast data; Construct an LSTM model, combine the extracted important features into the LSTM model, divide the historical data into a training set, a validation set and a test set, train the LSTM model through the training set, so that the model input features are: the temperature change rate inside the compartment, the change rate of the fresh-keeping state data, the temperature and relative humidity of the forecast data, and the output is the remaining insurance time. The validation set verifies the accuracy of the LSTM model, and then adjusts the model parameters. After the adjustment, the performance of the LSTM model is evaluated through the test set; Obtain important features of the current food transportation, input the important features into the trained LSTM model, obtain the remaining shelf life of the food, and screen out the reachable paths whose difference between the remaining shelf life and the arrival time of the reachable paths does not exceed the preset time threshold.

5. According to the method of intelligent control of food distribution based on big data in claim 1, it is characterized in that: From the selected accessible paths, obtain the accessible path with the minimum cost calculated by the environmentally sensitive supply chain cost model, including: An environmentally sensitive supply chain cost model is defined. Specifically, the environmentally sensitive supply chain cost model is: Among them, R (x, y) represents the transportation volume from the starting point to the point (x, y), C (x, y) represents the cost per unit of transportation volume, and a represents the cost unit of loss. represents the loss rate of food p, represents the end time of a transport operation at a specific location (x,y) and for a specific product p, The constraints of the environmentally sensitive supply chain cost model are: , constraining each customer's location to be visited once and only once; ≤ the preset loss threshold, constraining the total loss not to exceed the acceptable loss; Each screened path is evaluated using the environmentally sensitive supply chain cost model, the total cost of each path is calculated, the costs of all paths are compared, and the path with the minimum cost is found.

6. A food delivery intelligent control system based on big data, characterized in that: The system comprises: The collection module is used to collect the temperature and humidity in the delivery compartment in real time during the delivery process, collect the location information and speed of the delivery vehicle, and obtain sensitive parameters of the delivered ingredients in the delivery vehicle; The prediction and screening module is used to obtain all accessible paths, and use the prediction model to predict the remaining freshness time of the food delivered on all accessible paths according to the current internal environment parameters of the carriage, the freshness status data of the food, and the next path parameters. The path parameters include weather and road conditions, and screen out the accessible paths whose remaining freshness time and the arrival time of the accessible paths do not exceed the preset time threshold. The module for obtaining the minimum cost path is used to obtain the minimum cost accessible path calculated by the environmentally sensitive supply chain cost model from the screened accessible paths.

7. According to claim 6, a food distribution intelligent control system based on big data is characterized in that: The acquisition module comprises: The temperature and humidity collection module is used to collect the temperature and humidity of the compartment where food is delivered through a temperature sensor and a humidity sensor respectively; The vehicle information collection module is used to collect the location information and speed of the delivery vehicle in real time through the GPS device and the on-board OBD device; The module for obtaining sensitive parameters of ingredients is used to obtain sensitive parameters of ingredients delivered by the delivery vehicle.

8. According to claim 7, a food distribution intelligent control system based on big data is characterized in that: The module for obtaining sensitive parameters of ingredients includes: The module for determining the range of environmental variables is used to determine the range of environmental variables to be tested, including: setting the temperature gradient, setting the humidity gradient, and setting the test time period to select the target food sample; A recording module is used to store target food samples under different temperature and humidity conditions and regularly record their freshness status, which includes weight, nutritional components, water loss rate, color change and texture change. The color change is detected by a colorimeter, and the texture change is detected by a hardness meter. The structured data module is used to record the temperature and humidity in the test environment in real time, perform sampling within a preset time, and save the collected data as a structured data table, which includes the fields: time, temperature, humidity, and preservation status; A preliminary cleaning module is used to perform preliminary data cleaning on the data stored in the structured data table, wherein the preliminary data cleaning includes: removing outliers and averaging repeated experimental data under the same conditions; The preliminary calculation module is used to substitute the experimental data into the sensitive parameter model. Specifically, the sensitive parameter model is: in, represents the rate of change of food freshness, k represents the process rate constant, T is the temperature in the actual operating environment, Represents the temperature constant, H represents the relative humidity, takes the logarithm of the sensitive parameter model, n represents the sensitivity coefficient of food p loss to temperature, m represents the sensitivity coefficient of food p loss to humidity, converts the sensitive parameter model into a linear equation, and uses multiple linear regression to calculate k, n and m in the sensitive parameter model; a fitting error module is used to optimize the fitting error using the least squares method and verify the residual distribution of the fitting. When the residual distribution does not exceed the preset threshold, the final sensitive parameter model is obtained.

9. The food distribution intelligent control system based on big data according to claim 6 is characterized in that: The prediction and screening module comprises: A data acquisition module is used to acquire historical temperature and humidity data inside the carriage, and acquire historical fresh-keeping status data of food materials, wherein the historical fresh-keeping status data includes weight, nutrient content, water loss rate, color change, texture change, sugar content, and food material sensitive parameters, wherein the food material sensitive parameters include a sensitivity coefficient to temperature and a sensitivity coefficient to humidity; The feature extraction module is used to obtain the forecast data of weather conditions in the next transportation route, process and clean the collected historical temperature data, historical humidity data, and historical freshness status data, remove outliers and noise, and extract important features from the cleaned historical data, including: the temperature change rate inside the carriage, the change rate of the freshness status data, and the temperature and relative humidity of the forecast data; A training model module is used to construct an LSTM model, combine the extracted important features into the LSTM model, divide the historical data into a training set, a validation set and a test set, and train the LSTM model through the training set, so that the model input features are: the temperature change rate inside the compartment, the change rate of the fresh-keeping state data, the temperature and relative humidity of the forecast data, and the output is the remaining insurance time. The validation set verifies the accuracy of the LSTM model, and then adjusts the model parameters. After the adjustment, the performance of the LSTM model is evaluated through the test set; The prediction module is used to obtain important features of the current food transportation, input the important features into the trained LSTM model, obtain the remaining shelf life of the food, and screen out the reachable paths in which the difference between the remaining shelf life and the arrival time of the reachable paths does not exceed a preset time threshold.

10. The food delivery intelligent control system based on big data according to claim 6, characterized in that: The module for obtaining the minimum cost path includes: An environmentally sensitive supply chain cost model is defined. Specifically, the environmentally sensitive supply chain cost model is: Among them, R (x, y) represents the transportation volume from the starting point to the point (x, y), C (x, y) represents the cost per unit of transportation volume, and a represents the unit of loss cost. represents the loss rate of food p, represents the end time of a transport operation at a specific location (x,y) and for a specific product p, The constraints of the environmentally sensitive supply chain cost model are: , constraining each customer's location to be visited once and only once; ≤ the preset loss threshold, constraining the total loss not to exceed the acceptable loss; Each screened path is evaluated using the environmentally sensitive supply chain cost model, the total cost of each path is calculated, the costs of all paths are compared, and the path with the minimum cost is found.

Images