Intelligent supply chain and logistics optimization management method and system

Abstract

The invention relates to the technical field of supply chains, and discloses an intelligent supply chain and logistics optimization management method and system, and the method comprises the steps: obtaining a real-time order of each supply chain station in a region, and planning a logistics distribution path; data analysis is carried out, the sales volume of orders on the current day is predicted, and the inventory of a supply chain station is predicted in combination with supply volume consumption analysis; and according to calculation of the inventory emergency degree of the supply chain station, a target scheduling station is selected, and an inventory logistics scheduling scheme is generated. According to the invention, full-chain optimization management from order obtaining to logistics distribution path planning, accurate stock prediction and reasonable stock scheduling is realized, the logistics distribution efficiency can be significantly improved, the transportation cost is reduced, refined and scientific management of the stock is realized, the stock cost and the stockout risk are accurately balanced, and the logistics distribution efficiency is improved. The capacity of the supply chain for dealing with market changes and uncertainty is enhanced, the customer satisfaction is improved, the market competitiveness is improved, and efficient, stable and sustainable operation of the supply chain system is guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of supply chain technology, and in particular to a smart supply chain and logistics optimization management method and system. Background Art

[0002] With the acceleration of globalization and the booming development of e-commerce, the complexity of supply chain and logistics management is increasing. Traditional methods are gradually showing their limitations in the face of rapidly changing market demands, volatile external environments (such as weather and traffic conditions), and consumers' high demands for delivery time and service quality.

[0003] Existing technologies have the following problems: they rely on regularly updated data and cannot reflect the latest business conditions in real time, resulting in information asymmetry and affecting the timeliness and accuracy of decision-making. Existing logistics distribution route planning is mostly based on static data and fails to fully consider factors such as real-time traffic conditions and weather changes, which can easily lead to extended transportation time and increased costs. Sales forecasts are not accurate enough and it is difficult to capture complex market trends and consumer behaviors, resulting in large forecast errors. Inventory management optimization capabilities are poor and there is a lack of refined analysis of multiple factors, resulting in excessively high or low inventory levels, increased holding costs or affected sales, as well as irrational resource allocation. Summary of the Invention

[0004] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of this application to avoid obscuring the purpose of this section, the abstract and the title of the invention, and such simplifications or omissions should not be used to limit the scope of the present invention.

[0005] In view of the above existing problems, the present invention is proposed. Therefore, the present invention provides a smart supply chain and logistics optimization management method to solve the above problems.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, the present invention provides a smart supply chain and logistics optimization management method, comprising: Obtain real-time orders from each supply chain station in the region, plan logistics delivery routes based on the real-time orders, and track logistics delivery status in real time; Based on the real-time orders and historical orders of the supply chain station, data analysis is performed regularly to predict the sales volume of orders for the day, and combined with the supply consumption analysis, the inventory level of the supply chain station is predicted; Based on the predicted results of the inventory quantity of the supply chain station, it is compared with the inventory threshold. If it is greater than the inventory threshold, the inventory quantity is continuously monitored. If it is less than the inventory threshold, the target scheduling site is selected based on the calculation of the inventory urgency of the supply chain station to generate an inventory logistics scheduling plan.

[0007] As a preferred solution of the intelligent supply chain and logistics optimization management method of the present invention, obtaining real-time orders of each supply chain station in the region includes: Obtain order data, including order number, order status, user ID, product name, product quantity, product price, recipient, recipient phone number, recipient address, order time, estimated delivery time, supply chain station name, supply chain station address, and logistics information. Order status includes "pending delivery," "delivering," "completed," and "cancelled." Divide the coverage of each supply chain station, determine the supply chain station to which the order address belongs based on the order address, and the supply chain station to which the order address belongs receives the order data; Based on the order data, all orders with a status of "to be delivered" are sorted in chronological order to generate a set of orders to be delivered, and real-time orders are received to update the set of orders to be delivered; Sort the orders of the set of orders to be delivered, extract the first-ranked target order and target similar orders that match the target order in sequence, generate multiple target delivery order sets, and sort them according to the order in which the sets are generated to generate a target set queue; The target similar order is represented as an order to be delivered whose order placement time difference with the target order and the distance between the order addresses of the two orders are both less than a preset threshold.

[0008] As a preferred solution of the intelligent supply chain and logistics optimization management method of the present invention, wherein: planning the logistics distribution route according to the real-time order and tracking the logistics distribution status in real time includes: Obtain the logistics status of all logistics terminals of the supply chain station, extract all logistics terminals with idle status, and sort them in descending order according to the idle time of the logistics terminals to generate a logistics queue; The target set queue and the logistics delivery queue are marked in a one-to-one correspondence according to the order of the two queues, the orders in the target delivery order set are delivered, and the delivery path is generated according to the order addresses in the target delivery order set; Collect the real-time location of the logistics end and mark it on the map, track and record the delivery time of the orders in the target delivery order set of the logistics end, the average order delivery data interval time, the order delivery residence time, and the return time to the supply chain station; If, during the evaluation cycle, the user's order delivery stay time is greater than the stay threshold number of times, and the number of times is greater than the threshold, the user ID will be marked; if, in the next evaluation cycle, the user's order delivery stay time is greater than the stay threshold number of times, and the number of times is less than the threshold, the user ID will be unmarked; when generating a target delivery order set for the set of to-be-delivered orders containing the marked user, the number of orders in the target delivery order set will be less than the preset threshold.

[0009] As a preferred solution of the intelligent supply chain and logistics optimization management method of the present invention, wherein: based on the real-time orders and historical orders of the supply chain station, data analysis is performed regularly to predict the sales volume of orders on the same day, including: According to the set forecast period, the real-time orders of the supply chain station, historical orders of different periods, as well as the weather data, date attributes, and promotion activity data of the day are obtained, and the data are pre-processed; Extract features from the data, including order time features, weather features, product features, and user features; Wavelet analysis is used to decompose the historical order time series of each supply chain station, and the order data is decomposed at multiple scales to obtain subsequences of different frequencies; A sales hybrid prediction model is constructed using a deep neural network and random forest. The deep neural network model includes an input layer, a hidden layer, and an output layer. The input layer encodes order time features, weather features, product features, and user features as input, and the number of nodes is determined based on the feature dimension. The hidden layer has multiple layers and defines different activation functions. The output layer outputs the sales volume of orders for the day. The random forest model uses the prediction residuals of the deep network model, order time features, weather features, product features, and user features as input features to determine model parameters and fit the prediction residuals to correct prediction bias. The daily order sales forecast result of the sales hybrid forecasting model includes the sum of the prediction value of the deep neural network model and the correction value of the random forest model.

[0010] As a preferred solution of the intelligent supply chain and logistics optimization management method of the present invention, the forecasting of the inventory of the supply chain station includes: The supply consumption analysis includes a total consumption analysis of product losses, product compensation, consumption of ordered products, and order sales forecasts for future periods; Commodity loss analysis includes collecting commodity storage environment data and logistics transportation environment data to generate environmental factor vectors , and establish a commodity loss model , get the commodity loss value, where, represents the functional relationship obtained by data fitting, Indicates the commodity loss rate; Product compensation analysis includes analyzing user complaint texts and after-sales feedback text data through natural language processing, classifying the texts, identifying texts associated with compensation keywords, and generating compensation text feature vectors. , build a product compensation model , get the product compensation value, where, Represents the mapping relationship between the payout ratio and text features learned through the model, Indicates the product payout ratio; The consumption analysis of ordered goods includes obtaining the consumption values ​​of goods ordered on the day based on real-time order data; The order sales forecast analysis for the future period includes obtaining the order commodity consumption forecast value of each station based on the order sales forecast value of the supply chain station on the same day, and revising the order commodity consumption forecast value in combination with the order cancellation and return forecast value; Based on the initial inventory data of the day, combined with the product loss value, product compensation value, the consumption value of the goods ordered on the day, and the predicted consumption value of the ordered goods, the inventory forecast value of the supply chain station is calculated, which is expressed as: ; in, express Inventory forecast value at time, Indicates the initial inventory value of the product on that day. Indicates the commodity loss value at the previous moment. Indicates the compensation value of the product at the last moment. Indicates the consumption value of the goods ordered at the last moment. express The predicted consumption value of order goods at the moment, Represents the inventory safety factor.

[0011] As a preferred solution of the intelligent supply chain and logistics optimization management method of the present invention, the calculation of the inventory urgency of the supply chain station includes: Based on the forecast result of the inventory quantity of the supply chain station, if it is less than the inventory threshold, the inventory urgency of the supply chain station is calculated according to the available time of the inventory forecast value, the out-of-stock risk coefficient and the commodity demand fluctuation coefficient. , expressed as: ; in, represents the daily average sales forecast value, express The inventory value at the moment, Represents the mean of the order sales forecast value series for the site in the future. Indicates the order sales forecast value sequence of the site in the future period. A sequence value.

[0012] As a preferred solution of the intelligent supply chain and logistics optimization management method of the present invention, selecting a target scheduling site and generating an inventory logistics scheduling plan includes: According to the defined inventory emergency level, the corresponding emergency level is divided according to the inventory emergency value of the supply chain station, and the emergency level includes urgent, general, and non-urgent; If the inventory urgency of the supply chain station is urgent, the station closest to the current station with sufficient inventory is selected as the target scheduling station, and an inventory scheduling plan is generated and sent to the target scheduling station. Based on the current idleness of the logistics terminals of the two stations, the station with higher logistics idleness is selected to complete the scheduling and delivery; If the inventory urgency of the supply chain station is general or non-urgent, the distribution center is selected as the target scheduling site, an inventory scheduling plan is generated and sent to the target scheduling site, and the distribution center completes the scheduling and delivery.

[0013] In a second aspect, the present invention provides a smart supply chain and logistics optimization management system, comprising: The order distribution module is used to obtain real-time orders from each supply chain station in the region, plan logistics distribution routes based on the real-time orders, and track logistics distribution status in real time; A forecasting module is used to perform data analysis based on the real-time orders and historical orders of the supply chain station, forecast the sales volume of orders for the day, and forecast the inventory level of the supply chain station in combination with the supply consumption analysis; The scheduling module is used to compare the predicted results of the inventory level of the supply chain station with the inventory threshold. If the inventory level is greater than the inventory threshold, the inventory level is continuously monitored. If the inventory level is less than the inventory threshold, the target scheduling station is selected based on the calculation of the inventory urgency of the supply chain station to generate an inventory logistics scheduling plan.

[0014] In a third aspect, the present invention provides an electronic device, comprising: memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the smart supply chain and logistics optimization management method are implemented.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the intelligent supply chain and logistics optimization management method.

[0016] Compared with the existing technology, the beneficial effects of the present invention are as follows: the present invention realizes the optimization management of the entire chain from order acquisition to logistics distribution route planning, accurate inventory forecasting and reasonable inventory scheduling, which can significantly improve logistics distribution efficiency, reduce transportation costs, realize refined and scientific management of inventory, accurately balance inventory costs and out-of-stock risks, enhance the supply chain's ability to cope with market changes and uncertainties, improve customer satisfaction, enhance market competitiveness, and ensure the efficient, stable and sustainable operation of the supply chain system. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0018] Figure 1 This is a schematic diagram of the overall process of the smart supply chain and logistics optimization management method described in one embodiment of the present invention. DETAILED DESCRIPTION

[0019] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0020] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0021] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0022] The present invention is described in detail with reference to schematic diagrams. For ease of illustration, cross-sectional views of device structures may be partially enlarged and not to scale when describing embodiments of the present invention. Furthermore, the schematic diagrams are merely illustrative and should not limit the scope of the present invention. Furthermore, in actual production, the three-dimensional dimensions of length, width, and depth should be included.

[0023] In the description of the present invention, it should be noted that the terms "upper, lower, inner, and outer" and other references to orientations or positional relationships are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first, second, or third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0024] In this disclosure, unless otherwise specified or limited, the terms "mounted," "connected," and "connected" should be interpreted broadly. For example, they may refer to fixed, removable, or integral connections. They may also refer to mechanical, electrical, or direct connections, indirect connections through an intermediary, or internal communication between two components. Those skilled in the art will understand the specific meanings of these terms in this disclosure.

[0025] Reference Figure 1 , as an embodiment of the present invention, provides a smart supply chain and logistics optimization management method, including: S101, obtain the real-time orders of each supply chain station in the region, plan the logistics distribution route based on the real-time orders, and track the logistics distribution status in real time; S102, based on the real-time orders and historical orders of the supply chain station, data analysis is performed regularly to predict the sales volume of orders for the day, and combined with the supply consumption analysis, the inventory level of the supply chain station is predicted; S103, based on the predicted results of the inventory of the supply chain station, compare with the inventory threshold. If it is greater than the inventory threshold, the inventory is continuously monitored. If it is less than the inventory threshold, the target scheduling station is selected according to the calculation of the inventory urgency of the supply chain station to generate an inventory logistics scheduling plan.

[0026] In a preferred embodiment, step S101, obtaining real-time orders of each supply chain station in a region, includes: Obtain order data, including but not limited to order number, order status, user ID, product name, product quantity, product price, recipient, delivery phone number, delivery address, order time, estimated delivery time, supply chain station name, supply chain station address, and logistics information. Order status includes pending delivery, delivering, completed, canceled, etc. Divide the coverage of each supply chain station, determine the supply chain station to which the order address belongs based on the order address, and the supply chain station to which it belongs receives the order data; Based on the order data, all orders with the status of "to be delivered" are sorted in the order of order time, and a set of orders to be delivered is generated. In addition, real-time orders are received and the set of orders to be delivered is updated. Sort the orders in the set of orders to be delivered, extract the first-ranked target order and target similar orders that match the target order, generate multiple target delivery order sets, and sort them according to the order in which the sets are generated to generate a target set queue; Among them, the target similar orders are represented as orders to be delivered whose order time difference with the target order and the distance between the order addresses of the two are less than the preset threshold.

[0027] Specifically, the user coverage of each supply chain station does not overlap, and the corresponding station is automatically matched according to the address where the user places an order. The corresponding station receives the user's order in real time, selects goods and performs logistics distribution according to the order content. The initial state of the confirmed order is to be delivered, and the station allocates and distributes the goods to the order address. Considering the timeliness of order delivery, the orders are sorted according to the order placement time, and the orders with earlier orders are delivered first to improve customer satisfaction. A to-be-delivered order set is generated for all the to-be-delivered orders of the station. Among them, according to the real-time orders of the system, they are added to the to-be-delivered order set in real time, the set is updated, and these orders in the set are divided into groups, and multiple target delivery order sets are generated. The target delivery order set contains at least one to-be-delivered order. The purpose is to divide multiple similar orders into the same target delivery order set to improve logistics efficiency. The target delivery order set is sorted according to the generation time order to determine the logistics distribution order of the order. Among them, the preset thresholds of the order placement time difference and the distance between the two order addresses can be set according to the actual application scenario.

[0028] In a preferred embodiment, step S101, planning a logistics delivery route based on real-time orders and tracking the logistics delivery status in real time, includes: Obtain the logistics status of all logistics terminals of the supply chain station, extract all logistics terminals with idle status, and sort them in descending order according to the idle time of the logistics terminals to generate a logistics queue; The target set queue and the logistics delivery queue are marked in a one-to-one correspondence according to the order of the two queues, the orders in the target delivery order set are delivered, and the delivery route is generated according to the order address in the target delivery order set; Collect the real-time location of the logistics end and mark it on the map, track and record the delivery time of orders in the target delivery order set of the logistics end, the average order delivery data interval time, the order delivery residence time, and the return time to the supply chain station; If, during the evaluation cycle, the user's order delivery stay time is greater than the stay threshold number of times, and the number of times is greater than the threshold, the user ID will be marked. If, in the next evaluation cycle, the user's order delivery stay time is greater than the stay threshold number of times, and the number of times is less than the threshold, the user ID will be unmarked. When generating a target delivery order set for the set of to-be-delivered orders containing the marked user, the number of orders in the target delivery order set will be less than the preset threshold.

[0029] Specifically, if there is no associated target set queue to be delivered or in delivery under the name of the logistics end, the status of the logistics end is idle. If there is a target set queue to be delivered or in delivery on the logistics end, the status of the logistics end is busy. When the last order in the target set queue of the logistics end is delivered and the arrival time returned to the supply chain station is generated, the status of the logistics end will change from busy to idle. When the logistics end completes the association and reception with the target set queue to be delivered, the status of the logistics end will change from idle to busy. First, the logistics end is assigned to the target set queue, and the target set queue and the logistics queue are marked and associated one by one according to the same sorting. That is, the first target set queue is associated with the first logistics end of the logistics queue for delivery. Secondly, according to the delivery address of the order in the target delivery order set in the target set queue, the order closest to the supply chain station is the first delivery order of this delivery task. Combined with other order addresses and considering real-time traffic conditions, the delivery path is planned with the shortest path or shortest time as the goal, and the delivery task is completed according to the delivery path. During logistics distribution, both users and managers can see the real-time delivery location on the map. When logistics is delivering orders, it is necessary to record the delivery time of each order in the target delivery order set, the average order delivery data interval in the set, the delivery stay time of each order, and the time it takes to return to the supply chain station after all orders in the set are delivered.

[0030] Among them, the system needs to conduct regular evaluations of the users of the platform. The evaluation cycle, stay threshold, number threshold, and preset threshold of the number of collective orders can be set according to actual conditions. In this embodiment, the evaluation cycle is preferably one week, the stay threshold is 10 minutes, the number threshold is 3 times, and the preset threshold of the number of collective orders is 3; the service level of the user is judged by the user's order delivery stay time and number of times, which can reflect the frequency of emergencies or service difficulty of the user, so as not to delay the delivery of other orders. The user is marked and the number of orders delivered in the same batch is reduced when planning the delivery of the user's orders.

[0031] It should be noted that this step can grasp the order demand situation of each site in the first time, accurately know the key information such as the type, quantity and order time of the goods ordered by the customer, provide basic data support for subsequent rapid response and delivery arrangements, and avoid delivery delays caused by information delays; by analyzing real-time order information and combining the geographical location, traffic conditions and other factors of each site to plan the best delivery route, it can effectively shorten the delivery distance, reduce transportation time, improve logistics distribution efficiency, reduce transportation costs, reasonably avoid congested sections, and choose a better delivery sequence; it is convenient for supply chain managers and customers to understand the location and estimated arrival time of the goods at any time, and enhance the transparency of the delivery process. For managers, they can promptly discover unexpected situations in the delivery process (such as vehicle breakdowns, traffic jams causing delays, etc.) and make timely adjustments; for customers, they can make preparations for receiving the goods in advance, thereby improving customer satisfaction.

[0032] In a preferred embodiment, step S102, based on the real-time orders and historical orders of the supply chain station, regularly performing data analysis to predict the sales volume of orders for the day includes: According to the set forecast period, real-time orders from the supply chain station, historical orders from different periods, as well as weather data, date attributes, and promotion activity data for the day are obtained, and the data is pre-processed. The forecast period can be set as needed, and in this embodiment, it is preferably 3 hours. Extract features from the data, including order time features, weather features, product features, and user features; Wavelet analysis is used to decompose the historical order time series of each supply chain station, and the order data is decomposed at multiple scales to obtain subsequences of different frequencies; A sales hybrid prediction model is constructed using a deep neural network and random forest. The deep neural network model consists of an input layer, a hidden layer, and an output layer. The input layer encodes order time features, weather features, product features, and user features as input. The number of nodes is determined based on the feature dimension. The hidden layer has multiple layers and defines different activation functions. The output layer outputs the sales volume of orders for the day. The random forest model uses the prediction residuals of the deep network model, order time features, weather features, product features, and user features as input features to determine model parameters and fit the prediction residuals to correct prediction bias. The sales hybrid forecasting model's daily order sales forecast results include the sum of the deep neural network model's prediction value and the random forest model's correction value.

[0033] Specifically, data integration and preprocessing mainly involves obtaining real-time order data from each site, including order time, product type, quantity, ordering user information, etc. At the same time, historical order data for at least the past six months is collected, covering orders in different seasons, months, weeks, and special periods (such as holidays, e-commerce promotions, etc.). The data is cleaned, and missing and outliers are processed. Missing values ​​can be filled based on the mean or median of historical data; outliers, such as data points with order volumes far exceeding the normal range, are identified and corrected by comparing with data from similar sites during the same period or based on statistical principles (such as the 3-times standard deviation method). At the same time, external data is integrated, such as local weather information, date attributes, promotional activity data, etc.

[0034] Extract order data features, including order time features (such as day of the week, whether it is a weekend, whether it is the beginning or end of the month, whether it is a holiday, etc.), product features (product category, price range, whether it is a new product, etc.), and user features (new and old users, user consumption level, etc.).

[0035] Time series decomposition and trend analysis: Use methods such as wavelet analysis or empirical mode decomposition (EMD) to decompose the historical order time series of each site and split it into components such as long-term trends, seasonal cycles, and short-term fluctuations. For example, when using wavelet analysis, select an appropriate wavelet basis function (such as Haar wavelet) to perform multi-scale decomposition of the order data to obtain sub-sequences of different frequencies, representing trends and fluctuations at different levels. For the long-term trend component, use linear regression or nonlinear regression models for fitting to capture the overall growth or decline trend of order volume over time. For the seasonal cycle component, analyze its cyclical pattern (such as weekly, monthly, and annual periodicity) and calculate the seasonal index. This index can be used to measure the fluctuation of each month's order volume relative to the annual average level.

[0036] Furthermore, a hybrid prediction model is constructed based on deep neural network (DNN) and random forest (RF). DNN is used to learn the complex nonlinear relationships in order data, and RF is used to correct and supplement the prediction results of DNN.

[0037] Deep neural network models include: Input layer: Order time features, product features, user features, and user characteristics are encoded as input. The number of nodes is determined by the feature dimension. Hidden layer: Set up multiple hidden layers and use different activation functions (such as ReLU function). Optimize model performance by adjusting the number of hidden layer nodes and layers. Hyperparameter optimization methods such as grid search or random search can be used. Output layer: Outputs the predicted order volume. The model is trained using the mean squared error (MSE) as the loss function, using a backpropagation algorithm and an optimizer (such as the Adam optimizer). The historical order data is divided into training, validation, and test sets. The validation set loss is monitored during training to prevent overfitting.

[0038] Random forest model construction includes: The prediction residual of DNN is used as one of the input features of RF, and the order time feature, product feature, etc. are used as other input features; Determine the number of RF decision trees, tree depth, node splitting criteria and other parameters. The optimal parameter combination can be determined through hyperparameter optimization methods. The RF model fits the prediction residuals of DNN through ensemble learning of a large number of decision trees to correct possible prediction biases of DNN. The final prediction result is the sum of the DNN prediction value and the RF correction value.

[0039] Model evaluation and dynamic adjustment include: The time series cross-validation method is used to evaluate the model performance. The historical data is divided into multiple training-testing subsets in chronological order. For example, m The data of each time step is used as the training set, the order volume of the next time step is predicted and compared with the actual value, and the evaluation indicators (such as MSE, mean absolute error MAE, mean absolute percentage error MAPE, etc.) are calculated. The model performance is monitored in real time according to the evaluation indicators. If the model performance is found to be declining (such as the prediction error increases), the model update mechanism is triggered. The update mechanism can adopt an incremental learning method to gradually add new order data and related features to the training set, and fine-tune the model to adapt to the dynamic changes in market demand. At the same time, the model is fully retrained regularly (such as monthly or quarterly) to ensure that the model can fully learn the long-term and short-term market changes.

[0040] It should be noted that this step integrates multiple data (real-time orders, historical orders, external data, etc.) and cleans and processes them to ensure the accuracy and completeness of the data, making subsequent data-based analysis and predictions more reliable, reducing prediction biases caused by data quality issues, and laying a good foundation for accurate sales forecasting; extracting rich order features helps to more carefully grasp the order patterns of different products, different time periods, and different user groups, and realize refined grouping forecasts, which can better adapt to the complex and changing market demand characteristics and improve the pertinence and accuracy of forecasts; using advanced time series decomposition methods, it is possible to deeply explore the inherent laws of changes in order volume, not only accurately capture the overall development trend, but also accurately grasp the characteristics of cyclical changes, thereby providing a basis for more accurate sales forecasts; using a hybrid forecasting model to give full play to the advantages of different models, DNN can learn complex nonlinear relationships, and RF can correct the prediction bias of DNN. The combination of the two can effectively improve the accuracy and stability of the forecast, better cope with complex order data patterns and the uncertainty of the market environment, and reduce the risk of overfitting or underfitting that may exist in a single model.

[0041] In a preferred embodiment, step S102, predicting the inventory of the supply chain station includes: Supply consumption analysis includes total consumption analysis of product losses, product compensation, consumption of ordered products, and order sales forecasts for future periods; Commodity loss analysis includes collecting commodity storage environment data and logistics transportation environment data to generate environmental factor vectors , and establish a commodity loss model , get the commodity loss value, where, It represents the functional relationship obtained by data fitting. The function can be constructed by methods such as logistic regression and neural network. Indicates the commodity loss rate; Product compensation analysis includes analyzing user complaint texts and after-sales feedback text data through natural language processing, classifying the texts, identifying texts associated with compensation keywords, and generating compensation text feature vectors. , build a product compensation model , get the product compensation value, where, Represents the mapping relationship between the payout ratio and text features learned through the model, Indicates the product payout ratio; The consumption analysis of ordered goods includes obtaining the consumption values ​​of goods ordered on the day based on real-time order data; The order sales forecast analysis for the future period includes obtaining the order product consumption forecast for each station based on the current day's order sales forecast of the supply chain station, and revising the order product consumption forecast based on the order cancellation and return forecast. Based on the initial inventory data of the day, combined with the product loss value, product compensation value, the consumption value of the goods ordered on the day, and the predicted consumption value of the ordered goods, the inventory forecast value of the supply chain station is calculated, which is expressed as: ; in, express Inventory forecast value at time, Indicates the initial inventory value of the product on that day. Indicates the commodity loss value at the previous moment. Indicates the compensation value of the product at the last moment. Indicates the consumption value of the goods ordered at the last moment. express The predicted consumption value of order goods at the moment, Represents the inventory safety factor.

[0042] Specifically, the environmental data stored in the commodity loss analysis may include temperature, humidity, light, etc., and the logistics transportation environment data may include vibration, collision, etc. during transportation.

[0043] In product compensation analysis, the text is first preprocessed, including word segmentation, stop word removal, and word vector representation. Then the text is classified to identify text related to compensation caused by product quality issues, transportation damage, etc. The product compensation rate is calculated based on the number and proportion of compensation texts.

[0044] In the order sales forecast analysis for the future period, by analyzing the cancellation and return information in the historical order data, an order cancellation and return prediction model is established. For example, a binary logistic regression model is used to predict the probability of order cancellation. , let the order-related feature vector be ,but ,in, is the Sigmoid function, It is a linear combination function, which modifies the order commodity consumption forecast value according to the predicted cancellation and return probability to obtain a more accurate order commodity consumption forecast.

[0045] It should be noted that the use of IoT technology to establish commodity loss monitoring and NLP-based intelligent analysis of commodity compensation can accurately quantify commodity losses and compensation situations, change the previous extensive management model based on experience-based estimation, accurately grasp the inventory change factors caused by losses and compensation in a data-driven manner, and improve the inventory forecast's ability to respond to these uncertain factors; comprehensively consider factors such as order cancellation rate and return rate to revise the order commodity consumption forecast, so that the forecast results are more in line with actual sales conditions, avoid inaccurate inventory forecasts caused by ignoring these common situations in actual business, and thus more accurately plan inventory levels to prevent inventory backlogs or out-of-stock phenomena; use a multi-model fusion method to predict future order sales and integrate the advantages of different models, which can more comprehensively and accurately grasp the changing trends of future order sales. On this basis, combined with the inventory safety factor and scientific estimates of future purchases to calculate the inventory forecast value, it is helpful to formulate reasonable inventory strategies, balance inventory costs and out-of-stock risks, and ensure the stable operation of the supply chain.

[0046] In a preferred embodiment, step S103, calculating the inventory urgency of the supply chain station includes: Based on the forecast results of the supply chain station inventory, if it is less than the inventory threshold, the inventory urgency of the supply chain station is calculated according to the inventory forecast value available time, out-of-stock risk coefficient and commodity demand fluctuation coefficient. , expressed as: ; in, represents the daily average sales forecast value, express The inventory value at the moment, Represents the mean of the order sales forecast value series for the site in the future. Indicates the order sales forecast value sequence of the site in the future period. A sequence value.

[0047] Specifically, the duration of inventory forecast value availability is used to measure how long the existing inventory can maintain normal sales without new purchases. At the same time, taking into account the impact of uncertain factors such as product loss and compensation on inventory, the out-of-stock risk coefficient is calculated. The higher this coefficient is, the greater the risk of inventory reduction due to non-sales factors, and the higher the inventory urgency. The demand fluctuation coefficient is used to measure the fluctuation of the site's order sales forecast value. It can be obtained by calculating the ratio of the standard deviation of the sales forecast value to the mean. Suppose the order sales forecast value sequence of the site in the future period is ,The larger the value of the demand fluctuation coefficient is, the greater the demand fluctuation of the site is, and the inventory urgency is relatively high.

[0048] Select the target scheduling site and generate the inventory logistics scheduling plan including: According to the defined inventory emergency level, the corresponding emergency level is divided according to the inventory emergency value of the supply chain station. The emergency levels include urgent, general, and non-urgent; If the inventory urgency of the supply chain station is urgent, the station closest to the current station with sufficient inventory is selected as the target scheduling station. An inventory scheduling plan is generated and sent to the target scheduling station. Based on the current idleness of the logistics terminals of the two stations, the station with higher logistics idleness is selected to complete the scheduling and delivery. If the inventory urgency of the supply chain station is general or non-urgent, the distribution center is selected as the target scheduling site, and an inventory scheduling plan is generated and sent to the target scheduling site, and the distribution center completes the scheduling and delivery.

[0049] Specifically, the inventory emergency level is divided into different numerical ranges according to the urgency setting, which can be set according to actual needs. The numerical ranges correspond to urgent, general, and non-urgent in order of high to low. High logistics idleness mainly refers to sites with a large number of idle logistics terminals.

[0050] It should be noted that by comparing with the inventory threshold and adopting different response strategies, continuous monitoring of situations where inventory is above the threshold can grasp the dynamic changes of inventory in real time, ensure that inventory is always within a reasonable monitoring range, avoid increased management costs due to excessive attention, and at the same time prevent abnormal inventory situations from being discovered in time due to negligence; accurately select the target scheduling site and generate a scheduling plan based on the urgency of the inventory, which changes the previous situation of blind replenishment or scheduling, and realizes the scientific and reasonable allocation of inventory resources among multiple sites. It can quickly respond to the needs of sites with insufficient inventory, effectively avoid out-of-stock phenomena, improve the inventory turnover rate of the entire supply chain, reduce inventory costs, ensure the efficient and stable operation of the supply chain, and improve customer service levels.

[0051] The present invention realizes the optimization management of the entire chain from order acquisition to logistics distribution route planning, accurate inventory forecasting and reasonable inventory scheduling, which can significantly improve logistics distribution efficiency, reduce transportation costs, realize refined and scientific management of inventory, accurately balance inventory costs and out-of-stock risks, enhance the supply chain's ability to cope with market changes and uncertainties, improve customer satisfaction, enhance market competitiveness, and ensure the efficient, stable and sustainable operation of the supply chain system.

[0052] The above is a schematic diagram of a smart supply chain and logistics optimization management method according to this embodiment. It should be noted that the technical solution of this smart supply chain and logistics optimization management system is based on the same concept as the technical solution of the smart supply chain and logistics optimization management method described above. For details not described in detail in the technical solution of the smart supply chain and logistics optimization management system according to this embodiment, please refer to the description of the technical solution of the smart supply chain and logistics optimization management method described above.

[0053] The intelligent supply chain and logistics optimization management system in this embodiment includes: The order distribution module is used to obtain real-time orders from each supply chain station in the region, plan logistics distribution routes based on real-time orders, and track logistics distribution status in real time; The forecasting module is used to analyze the real-time and historical order data of the supply chain station, predict the sales volume of the order on the same day, and combine the supply consumption analysis to predict the inventory level of the supply chain station; The scheduling module is used to compare the predicted results of the inventory level of the supply chain station with the inventory threshold. If it is greater than the inventory threshold, the inventory level is continuously monitored. If it is less than the inventory threshold, the target scheduling station is selected based on the calculation of the inventory urgency of the supply chain station to generate an inventory logistics scheduling plan.

[0054] This embodiment further provides an electronic device suitable for smart supply chain and logistics optimization management, including: Memory and processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the smart supply chain and logistics optimization management method proposed in the above embodiment.

[0055] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for realizing intelligent supply chain and logistics optimization management proposed in the above embodiment is implemented.

[0056] The storage medium proposed in this embodiment and the method for realizing intelligent supply chain and logistics optimization management proposed in the above embodiment belong to the same inventive concept. The technical details not described in detail in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0057] Through the above description of the embodiments, those skilled in the art will clearly understand that the present invention can be implemented using software and necessary general-purpose hardware. Of course, it can also be implemented using hardware, but in many cases the former is a more preferred embodiment. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk, or optical disk, and includes a number of instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of various embodiments of the present invention.

[0058] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A smart supply chain and logistics optimization management method, characterized in that: include: Obtain real-time orders from each supply chain station in the region, plan logistics distribution routes based on the real-time orders, and track logistics distribution status in real time; Based on the real-time orders and historical orders of the supply chain station, data analysis is performed regularly to predict the sales volume of orders on the day, and combined with the supply consumption analysis, the inventory of the supply chain station is predicted; Based on the predicted results of the inventory quantity of the supply chain station, the inventory quantity is compared with the inventory threshold. If it is greater than the inventory threshold, the inventory quantity is continuously monitored. If it is less than the inventory threshold, the target scheduling site is selected based on the calculation of the inventory urgency of the supply chain station to generate an inventory logistics scheduling plan.

2. The intelligent supply chain and logistics optimization management method according to claim 1, characterized in that: Get real-time orders for each supply chain station in the region including: Obtain order data, including order number, order status, user ID, product name, product quantity, product price, recipient, recipient phone number, recipient address, order time, estimated delivery time, supply chain station name, supply chain station address, and logistics information, wherein the order status includes pending delivery, delivering, completed, and canceled; Divide the coverage of each supply chain station, determine the supply chain station to which the order address belongs according to the order address, and let the supply chain station to which the order address belongs receive the order data; Based on the order data, all orders whose order status is to be delivered are sorted in the order of order time, a set of orders to be delivered is generated, and real-time orders are received to update the set of orders to be delivered; According to the order sorting of the to-be-delivered order set, the first-ranked target order and target similar orders matching the target order are sequentially extracted to generate multiple target delivery order sets, and the sets are sorted according to the generation order of the sets to generate a target set queue; The target similar order is represented as an order to be delivered whose order time difference with the target order and the distance between the order addresses of the two orders are both less than a preset threshold.

3. The intelligent supply chain and logistics optimization management method according to claim 2, characterized in that: Planning the logistics delivery route according to the real-time order and tracking the logistics delivery status in real time include: Obtain the logistics status of all logistics terminals of the supply chain station, extract all logistics terminals with idle status, and sort them in descending order according to the idle time of the logistics terminals to generate a logistics queue; The target set queue and the logistics delivery queue are marked in a one-to-one correspondence according to the order of the two queues, the orders in the target delivery order set are delivered, and the delivery path is generated according to the order addresses in the target delivery order set; Collect the real-time location of the logistics end and mark it on the map, track and record the delivery time of the orders in the target delivery order set of the logistics end, the average order delivery data interval time, the order delivery stay time, and the return time to the supply chain station; If during the evaluation period, the user's order delivery stay time is greater than the stay threshold number of times, and if it is greater than the number threshold, the user ID will be marked; if in the next evaluation period, the user's order delivery stay time is greater than the stay threshold number of times, and is less than the number threshold, the user ID will be unmarked; when generating a target delivery order set for the set of to-be-delivered orders containing the marked user, the number of orders in the target delivery order set is less than the preset threshold.

4. The intelligent supply chain and logistics optimization management method according to claim 1 or 3, characterized in that: Based on the real-time orders and historical orders of the supply chain station, data analysis is performed regularly to predict the sales volume of orders for the day, including: According to the set forecast cycle, the real-time orders of the supply chain station, historical orders of different periods, weather data, date attributes, and promotion activity data of the day are obtained, and the data are pre-processed; Extract features from the data, including order time features, weather features, product features, and user features; Wavelet analysis is used to decompose the historical order time series of each supply chain station, and the order data is decomposed at multiple scales to obtain subsequences of different frequencies; A sales hybrid prediction model is constructed by using a deep neural network and a random forest. The deep neural network model includes an input layer, a hidden layer, and an output layer. The input layer encodes order time features, weather features, product features, and user features as input, and determines the number of nodes according to the feature dimension. The hidden layer is configured with multiple layers and defines different activation functions. The output layer outputs the sales volume of orders on the day. The random forest model uses the prediction residuals of the deep network model, order time features, weather features, product features, and user features as input features, determines model parameters, and fits the prediction residuals to correct the prediction deviation. The same-day order sales forecast result of the sales hybrid forecasting model includes the sum of the deep neural network model prediction value and the random forest model correction value.

5. The intelligent supply chain and logistics optimization management method according to claim 4, characterized in that: The forecasting of the inventory quantity of the supply chain station includes: The supply consumption analysis includes a total consumption analysis of commodity loss, commodity compensation, consumption of ordered commodities, and order sales forecast for future periods; Commodity loss analysis includes collecting commodity storage environment data and logistics transportation environment data to generate environmental factor vectors , and establish a commodity loss model , get the commodity loss value, where, represents the functional relationship obtained by data fitting, Indicates the commodity loss rate; Product compensation analysis includes analyzing user complaint texts and after-sales feedback text data through natural language processing, classifying texts, identifying texts associated with compensation keywords, and generating compensation text feature vectors. , build a commodity compensation model , get the product compensation value, where, It represents the mapping relationship between the payout ratio and text features obtained through model learning. Indicates the product payout ratio; The consumption analysis of ordered goods includes obtaining the consumption values ​​of ordered goods on the day based on real-time order data; The order sales forecast analysis for the future period includes obtaining the order commodity consumption forecast value of each station based on the order sales forecast value of the supply chain station on the same day, and revising the order commodity consumption forecast value in combination with the order cancellation and return forecast value; According to the initial inventory data of the day, combined with the commodity loss value, commodity compensation value, commodity consumption value ordered on the day and the order commodity consumption forecast value, the inventory forecast value of the supply chain station is calculated, which is expressed as: ； in, express The inventory forecast value at the time, Indicates the initial inventory value of the product on that day. Indicates the commodity loss value at the previous moment. Indicates the compensation value of the product at the last moment. Indicates the consumption value of the goods ordered at the last moment. express The predicted consumption value of the order goods at the moment, Represents the inventory safety factor.

6. The intelligent supply chain and logistics optimization management method according to claim 1 or 5, characterized in that: The calculation of inventory urgency based on the supply chain station includes: Based on the forecast result of the inventory quantity of the supply chain station, if it is less than the inventory threshold, the inventory urgency of the supply chain station is calculated according to the available time of the inventory forecast value, the out-of-stock risk coefficient and the commodity demand fluctuation coefficient. , expressed as: ； in, represents the daily average sales forecast value, express The inventory value at the moment, Represents the mean of the order sales forecast value sequence of the site in the future period. Indicates the order sales forecast value sequence of the site in the future. A sequence value.

7. The intelligent supply chain and logistics optimization management method according to claim 6, characterized in that: Select the target scheduling site and generate the inventory logistics scheduling plan including: According to the defined inventory emergency level, the corresponding emergency level is divided according to the inventory emergency value of the supply chain station, and the emergency level includes emergency, general, and non-urgent; If the inventory urgency of the supply chain station is in an emergency state, the station closest to the current station and with sufficient inventory is selected as the target scheduling station, and an inventory scheduling plan is generated and sent to the target scheduling station. According to the idleness of the logistics ends of the two stations, the station with high logistics idleness is selected to complete the scheduling and delivery; If the inventory urgency of the supply chain station is general or non-urgent, the distribution center is selected as the target scheduling site, an inventory scheduling plan is generated and sent to the target scheduling site, and the distribution center completes the scheduling and delivery.

8. A smart supply chain and logistics optimization management system, characterized in that: include, The order distribution module is used to obtain the real-time orders of each supply chain station in the region, plan the logistics distribution route according to the real-time orders, and track the logistics distribution status in real time; A prediction module is used to perform data analysis based on the real-time orders and historical orders of the supply chain station, predict the sales volume of orders on the same day, and predict the inventory of the supply chain station in combination with the supply consumption analysis; The scheduling module is used to compare the predicted results of the inventory quantity of the supply chain station with the inventory threshold. If it is greater than the inventory threshold, the inventory quantity is continuously monitored. If it is less than the inventory threshold, the target scheduling site is selected based on the calculation of the inventory urgency of the supply chain station to generate an inventory logistics scheduling plan.

9. An electronic device, characterized in that: include: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the smart supply chain and logistics optimization management method described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: It stores computer executable instructions, which, when executed by a processor, can implement the steps of the smart supply chain and logistics optimization management method described in any one of claims 1 to 7.

Images