E-commerce missing order distribution method and related equipment
The method addresses e-commerce inventory management challenges by using prepayment verification and dynamic demand analysis to prioritize and allocate bargain hunting orders, enhancing customer satisfaction and operational efficiency.
Patent Information
- Application Number
- CN202411236738.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-04
- Publication Date
- 2024-12-03
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing e-commerce platforms struggle to dynamically allocate inventory and manage bargain hunting orders due to static inventory systems, false demands, and inefficient order prioritization, leading to inventory backlogs, customer dissatisfaction, and increased operational costs.
A method involving prepayment verification, dynamic demand analysis, and intelligent order allocation using machine learning algorithms to prioritize and allocate bargain hunting orders based on customer demand, inventory availability, and order authenticity, ensuring timely replenishment and optimal resource utilization.
Enhances customer satisfaction by accurately meeting demand, optimizing inventory management, reducing waste, and improving operational efficiency by ensuring timely order fulfillment and resource allocation.
Smart Images

Figure CN119067758A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of e-commerce, and in particular to a method for allocating e-commerce bargain hunting orders and related equipment. Background Art
[0002] With the development of e-commerce, consumers' demand for goods has become more and more diversified and immediate. However, on e-commerce platforms, due to the high demand for goods, popular goods often sell out quickly, resulting in consumers being unable to purchase the goods they need in time. In this case, how to effectively allocate remaining inventory or replenish it in time has become one of the key issues for e-commerce platforms to improve customer satisfaction and reduce customer churn. However, most of the existing methods for allocating missed orders rely on static inventory and order management systems, lacking dynamic analysis of customer needs, resulting in inventory backlogs or failure to meet customer needs in a timely manner in actual operations.
[0003] Existing e-commerce discount order allocation systems often ignore the dynamic nature of customer demand and the importance of order authenticity verification when dealing with the replenishment and order allocation of sold-out products. Due to the lack of an effective prepayment verification mechanism, there may be a large number of false demands in discount orders, resulting in a waste of inventory resources and unreasonable order allocation in the actual allocation process. This situation not only increases the operating costs of e-commerce platforms, but may also lead to customer dissatisfaction, further affecting the platform's credibility and customer stickiness.
[0004] In addition, the current methods of allocating bargain hunting orders often use simple logical judgments when dealing with the priority sorting of order cancellations and replenishment needs, ignoring the volatility of actual market demand and the timely adjustment of inventory. In this case, when multiple customers' bargain hunting needs emerge at the same time, the platform cannot make the best order allocation decision in time, which affects the customer experience and the platform's operational efficiency. Therefore, in response to these problems in the allocation process of e-commerce bargain hunting orders, a more intelligent and dynamic allocation method is urgently needed to optimize inventory management and order allocation and improve the overall operational efficiency of the platform. Summary of the invention
[0005] The main purpose of the present invention is to provide an e-commerce bargain hunting order allocation method and related equipment, which solves the technical problem that when multiple customers' bargain hunting demands emerge at the same time, the platform cannot make the best order allocation decision in time, thereby affecting the customer experience and the platform's operating efficiency.
[0006] To achieve the above object, the present invention provides a method for allocating e-commerce bargain hunting orders, comprising the following steps: When the product is sold out, obtain the sold-out product information and customer demand information, compare and analyze the customer demand information and the sold-out product information to obtain the first bargain order; The authenticity of the first bargain hunting order is verified through the prepayment mechanism to obtain a verified bargain hunting order queue; Based on the verified leak-picking order queue, a statistical calculation of leak-picking demand is performed to obtain a dynamic leak-picking order quantity; Obtain the allocable inventory quantity, and when the number of dynamic bargain hunting orders reaches a preset first threshold, notify the purchasing staff to replenish the goods, and obtain first replenishment data; wherein the preset first threshold is the difference between the number of dynamic bargain hunting orders and the allocable inventory quantity; Within a specified time before the e-commerce discount order cut-off time, timely remind customers who need to cancel orders in the first replenishment data, obtain the number of canceled orders, and obtain a second discount order based on the number of canceled orders and the first replenishment data; Determine whether the second bargain hunting order reaches a preset second threshold value. If reached, obtain second replenishment data based on the allocatable inventory quantity and the second bargain hunting order, prioritize and allocate the second bargain hunting order based on the second replenishment data and the allocatable inventory quantity, and obtain an allocation result of the e-commerce bargain hunting order; wherein the second threshold value is the difference between the values of the second bargain hunting order and the allocatable inventory quantity.
[0007] Further, the customer demand information and the sold-out product information are subjected to a comparison analysis to obtain a first bargain hunting order, including: Vectorizing the customer demand information through a preset quantitative evaluation technology to obtain a demand embedding vector; Perform sparse aggregation on the sold-out product information to obtain the sold-out product scarcity vector; Performing a comparison analysis on the demand embedding vector and the sold-out product scarcity vector to obtain a preliminary bargain order; Perform probabilistic deduction on the preliminary bargain hunting order to generate a first bargain hunting order.
[0008] Furthermore, the authenticity of the first bargain hunting order is verified through the prepayment mechanism to obtain a verified bargain hunting order queue, including: Extracting multi-dimensional information from the first bargain hunting order to obtain an order feature vector; wherein the multi-dimensional information includes the customer's historical purchase behavior, product attributes, and order amount; Performing an authenticity score on the order feature vector using a preset XGBoost algorithm to obtain an order authenticity score; The order authenticity score is divided into credibility levels by a preset dynamic threshold segmentation algorithm to obtain a credibility result; wherein the credibility result includes high-credibility orders and low-credibility orders; Through the smart contract prepayment mechanism of the blockchain, the low-credibility order in the credibility result is initiated, and the customer is required to pay a certain proportion of the prepayment to obtain the prepayment response result; Using a long short-term memory network model to perform a time series analysis on the prepayment response results to predict the customer's prepayment willingness and obtain the customer's prepayment willingness probability; By using a priority queue algorithm, the first bargain hunting orders are sorted and screened based on the customer's prepayment willingness probability and high-credibility orders to obtain a verified bargain hunting order queue.
[0009] Furthermore, based on the verified bargain hunting order queue, the demand for bargain hunting is statistically calculated to obtain the number of dynamic bargain hunting orders, including: By using a preset timing capture technology, the verification and leak-picking order queue is mapped to a multi-time demand map to capture the changes in the demand for leak-picking orders at different time points; Performing weighted aggregation on the demand changes of the missing order to form a dynamic demand weight matrix; Perform statistical calculation of the demand for leak picking based on the dynamic demand weight matrix to obtain a demand probability distribution; The demand probability distribution is calculated under different confidence intervals to obtain the number of dynamic missing orders.
[0010] Furthermore, when the number of the dynamic missing orders reaches a preset first threshold, a replenishment notification is sent to the purchasing staff to obtain first replenishment data, including: The real-time dynamic order data flow is monitored through a preset dynamic threshold trigger model. When the number of missed orders is detected to reach a preset first threshold for dynamic order triggering, a corresponding dynamic order trigger signal is generated; Performing calculation and analysis on the dynamic order trigger signal to obtain a replenishment warning matrix; wherein the replenishment warning matrix is used to analyze the current demand intensity, inventory status and supply chain tension of each commodity; Parsing the replenishment warning matrix to obtain a replenishment data stream; wherein the replenishment data stream includes a replenishment quantity, a time window, and a list of commodities with priority replenishment; The replenishment data stream is predicted and analyzed to obtain first replenishment data.
[0011] Furthermore, timely reminding customers who need to cancel orders in the first replenishment data to obtain the number of canceled orders, and obtaining a second missing order based on the number of canceled orders and the first replenishment data, including: Perform real-time analysis on the customer order status in the first replenishment data through a preset LSTM neural network model to identify customers who need to cancel their orders and generate a list of customers who need to cancel their orders; Based on the list of customers who need to cancel their orders, timely remind the customers who need to cancel their orders, and obtain the number of canceled orders; Using a preset dynamic decision tree network, conditionally screening the number of canceled orders and the first replenishment data to obtain a candidate replenishment combination; wherein the condition screening includes logistics capabilities, seasonal factors, promotional activities and product characteristics, and the product characteristics include shelf life and storage conditions; Performing vectorized calculation of a multi-dimensional replenishment strategy set on the candidate replenishment combination to obtain a demand response function; The demand response function is collaboratively calculated to obtain a conditional optimal solution space; wherein the conditional optimal solution space includes an optimal replenishment quantity, an optimal replenishment time, and an optimal distribution plan; Based on the conditional optimal solution space, the number of cancelled orders and the first replenishment data are hierarchically aggregated to obtain a preliminary second missing order; A time-varying weighted analysis is performed on the preliminary second bargain-hunting order to obtain the second bargain-hunting order.
[0012] Furthermore, the second bargain hunting orders are prioritized and allocated based on the second replenishment data and the allocatable inventory quantity to obtain allocation results of the e-commerce bargain hunting orders, including: Based on the second replenishment data and the allocatable inventory quantity, a pool of orders to be allocated is obtained; Based on the pool of orders to be allocated, priority reasoning is performed on the second leak-picking orders to obtain a priority scoring matrix; wherein the priority scoring matrix is used to reflect the urgency and importance of each second leak-picking order; Performing hierarchical aggregation on the priority scoring matrix to obtain a multi-dimensional priority scoring function; Dynamically weighting the second bargain hunting order based on the multi-dimensional priority scoring function to obtain an allocation plan for the second bargain hunting order; Probabilistically predicting the allocation scheme through a preset entropy weight priority sorting algorithm to obtain an adaptive allocation network; wherein the adaptive allocation network is used to determine the optimal allocation order; The adaptive allocation network is coordinated and optimized to obtain allocation results of e-commerce missed orders.
[0013] The present invention can also provide a device for allocating e-commerce missed orders, comprising: An analysis module is used to obtain sold-out product information and customer demand information when the product is sold out, and compare and analyze the customer demand information and the sold-out product information to obtain a first bargain order; A verification module, used to verify the authenticity of the first bargain hunting order through a prepayment mechanism, and obtain a verified bargain hunting order queue; A calculation module is used to perform statistical calculation of the demand for leak picking based on the verified leak picking order queue to obtain the number of dynamic leak picking orders; A notification module is used to obtain the allocable inventory quantity. When the number of dynamic bargain hunting orders reaches a preset first threshold, a replenishment notification is sent to the purchasing staff to obtain first replenishment data; wherein the preset first threshold is the difference between the number of dynamic bargain hunting orders and the number of allocable inventory; A reminder module is used to timely remind customers who need to cancel orders in the first replenishment data within a specified time before the e-commerce bargain hunting cut-off time, obtain the number of canceled orders, and obtain a second bargain hunting order based on the number of canceled orders and the replenishment data; The allocation module is used to determine whether the second bargain hunting order reaches a preset second threshold value. If reached, second replenishment data is obtained based on the allocatable inventory quantity and the second bargain hunting order, and the second bargain hunting order is prioritized and allocated based on the second replenishment data and the allocatable inventory quantity to obtain the allocation result of the e-commerce bargain hunting order; wherein the second threshold value is the difference between the second bargain hunting order and the allocatable inventory quantity.
[0014] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.
[0015] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned methods are implemented.
[0016] The method for allocating e-commerce bargain hunting orders provided by the present invention comprises the following steps: when a commodity is sold out, obtaining sold-out commodity information and customer demand information, performing bargain hunting comparative analysis on the customer demand information and the sold-out commodity information, and obtaining a first bargain hunting order; verifying the authenticity of the first bargain hunting order through a prepayment mechanism, and obtaining a verified bargain hunting order queue; performing bargain hunting demand statistical calculation based on the verified bargain hunting order queue, and obtaining a dynamic bargain hunting order quantity; obtaining the allocable inventory quantity, and when the dynamic bargain hunting order quantity reaches a preset first threshold, notifying the purchasing staff to replenish the goods, and obtaining the first replenishment data; wherein the preset first threshold is the difference between the dynamic bargain hunting order quantity and the allocable inventory quantity; within a specified time before the e-commerce bargain hunting order cut-off time, timely reminding customers who need to cancel their orders in the first replenishment data, and obtaining the number of canceled orders, and based on the number of canceled orders and The first replenishment data obtains a second bargain hunting order; determines whether the second bargain hunting order reaches a preset second threshold value, and if reached, obtains the second replenishment data based on the allocatable inventory quantity and the second bargain hunting order, and prioritizes and allocates the second bargain hunting order based on the second replenishment data and the allocatable inventory quantity to obtain an allocation result of the e-commerce bargain hunting order; wherein, the second threshold value is the difference between the second bargain hunting order and the allocatable inventory quantity; solves the technical problem that when the bargain hunting demands of multiple customers emerge at the same time, the platform cannot make the best order allocation decision in time, thereby affecting the customer experience and the operating efficiency of the platform, and achieves the beneficial effect of ensuring the optimal utilization of resources and maximizing customer satisfaction by prioritizing and reasonably allocating orders when processing canceled orders and bargain hunting orders. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 It is a schematic diagram of the steps of a method for allocating e-commerce bargain hunting orders in one embodiment of the present invention; Figure 2 This is a structural block diagram of a device for allocating orders for e-commerce bargain hunting in one embodiment of the present invention; Figure 3 It is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.
[0018] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION
[0019] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.
[0020] like Figure 1 As shown, Figure 1 This is a schematic diagram of steps of a method for allocating e-commerce bargain hunting orders in one embodiment of the present invention; An embodiment of the present invention provides a method for allocating e-commerce bargain hunting orders, comprising the following steps: Step S1, when the product is sold out, obtain the sold-out product information and customer demand information, compare and analyze the customer demand information and the sold-out product information to obtain a first bargain hunting order.
[0021] Specifically, when a product is sold out, the first thing to do is to obtain the information about the sold-out product and the customer demand information. This means that when a product on an e-commerce platform is sold out, the system will record and save the detailed information of the product, such as the name, model, price, inventory quantity, etc. At the same time, the system will also obtain customer demand information for the product, such as how many customers have tried to buy it but failed, or how many customers continue to express their intention to buy the product after it is sold out. This customer demand information may be collected through customers clicking the buy button, adding to the shopping cart, or registering for pre-order requirements after the product is sold out. Next, the customer demand information is compared with the sold-out product information for leak-picking comparison analysis. The purpose of this step is to match customer needs with the specific characteristics of the sold-out product. The so-called "leak-picking comparison analysis" can be understood as the system checking which customer needs are still unmet and whether these needs are consistent with the existing sold-out product information. For example, if a product is sold out on the platform, the system will analyze which customers are still looking for this product and what their specific needs are, such as color, storage capacity, etc. Through this analysis, the system can identify these unmet needs and map these needs to the sold-out products, forming a preliminary list of "pick-up orders". This "first pick-up order" list is generated through the above-mentioned pick-up comparative analysis. It contains orders for which customers still express their willingness to buy even though the products are sold out. These orders are generated not only based on customer demand information, but also combined with the information of the sold-out products. Therefore, this order list can help the platform better understand the actual needs of customers and provide a reference for subsequent replenishment and inventory allocation. Through the above steps, the e-commerce platform can more effectively manage the order demand after the sold-out products are sold out, ensuring that customers who still want to buy the sold-out products can be identified by the system and have the opportunity to obtain the products through pick-up. This method improves the accuracy of inventory management, increases customer satisfaction, and avoids the loss of customers due to the sold-out products. It should be noted that the above-mentioned products include but are not limited to mobile phones, sports shoes and fruits. Just to make the solution better understood, mobile phones, sports shoes and fruits are used as examples.
[0022] Step S2, verify the authenticity of the first bargain hunting order through the prepayment mechanism to obtain a verified bargain hunting order queue.
[0023] Specifically, in the step of "verifying the authenticity of the first bargain hunting order through the prepayment mechanism and obtaining the verified bargain hunting order queue", the e-commerce platform uses the prepayment mechanism to verify the authenticity of the first bargain hunting order. This means that after the customer submits the bargain hunting order, the platform will require the customer to make a prepayment of a certain amount, which is usually a part of the product price. The main function of the prepayment is to ensure that the customer's purchase intention is real, rather than placing an order at random or testing the system. In this way, the platform can filter out orders that are not really intended to buy, reduce the number of invalid orders, and ensure that subsequent replenishment and inventory allocation are more targeted and efficient. For example, suppose a customer applies for a bargain hunting order for a limited edition sneaker that has been sold out on the e-commerce platform. During the application process, the system will prompt the customer to pay a small prepayment of 10% of the total price of the product. After paying the prepayment, the order will be temporarily retained in the platform's "first bargain hunting order" list. However, if the customer cannot pay the prepayment, the order will not enter the next verification stage and will not be included in the "verified bargain hunting order queue". The "verified bargain hunting order queue" refers to those bargain hunting orders that have been verified through the prepayment mechanism to confirm that the customer actually has the intention to buy. This step is very critical because it ensures that when the platform handles replenishment and inventory allocation, it only considers those orders with actual purchase intentions, avoiding waste of resources and reduced operational efficiency due to false demand. The use of the prepayment mechanism is an effective verification method in the e-commerce field, which can be similar to the deposit payment in our daily life. For example, when we want to book tickets for a popular concert, we usually need to pay part of the fee in advance as a deposit to ensure that our order is given priority. Similarly, the prepayment mechanism plays a similar role in the verification of bargain hunting orders, ensuring that only those customers with real purchasing needs can enter the next order queue, thereby helping e-commerce platforms to manage inventory and resources more accurately and improve overall operational efficiency.
[0024] Step S3, based on the verified bargain hunting order queue, statistical calculation of bargain hunting demand is performed to obtain the dynamic bargain hunting order quantity.
[0025] Specifically, in the step of "calculating the demand for bargain hunting based on the verified bargain hunting order queue to obtain the number of dynamic bargain hunting orders", the e-commerce platform needs to further analyze and count the bargain hunting orders that have been verified by the prepayment mechanism. Specifically, the platform will perform statistical calculations on the demand for each order in the "verified bargain hunting order queue" and finally obtain a "dynamic number of bargain hunting orders". This dynamic number of bargain hunting orders represents the total number of all currently verified bargain hunting orders. It will change in real time according to the actual situation and is an important basis for the platform's subsequent replenishment and inventory allocation. In order to better understand this process, assume that within a certain period of time, the platform has 100 bargain hunting orders that have passed the prepayment verification and entered the verified bargain hunting order queue. These 100 orders correspond to different commodities or the same commodities, and the demand for each commodity is also different. The platform needs to count the demand for each commodity and add up these demands to obtain the final number of dynamic bargain hunting orders. This number is dynamic because over time, new bargain hunting orders may be verified and added to the queue, or existing orders may be cancelled, so the platform needs to continuously update this data to ensure the accuracy of decision-making. The word "dynamic" mentioned here is very important because it means that the order quantity is not fixed, but changes with time and customer demand. This is different from traditional static inventory management systems, which usually only consider fixed inventory and order quantities. The dynamic number of leak-picking orders can help the platform better cope with market fluctuations and changes in customer demand, ensuring that when the goods are re-listed or restocked, reasonable resource allocation can be made according to the latest order situation. For example, suppose that after a certain product is sold out on an e-commerce platform, the number of dynamic leak-picking orders in the verification leak-picking order queue has reached 200. This means that 200 customers have passed the prepayment verification and clearly expressed their desire to purchase this product. The platform will decide whether urgent replenishment is needed and how much the replenishment should be based on this number. This dynamic adjustment method enables the platform to respond to customer needs more flexibly, avoid customer loss due to insufficient supply, or inventory backlog due to excessive supply, thereby optimizing inventory management and sales strategies.
[0026] Step S4, obtaining the allocable inventory quantity. When the number of dynamic bargain hunting orders reaches a preset first threshold, a replenishment notification is sent to the purchasing staff to obtain first replenishment data; wherein the preset first threshold is the difference between the number of dynamic bargain hunting orders and the number of allocable inventory.
[0027] Specifically, in the step of "obtaining the number of allocable inventory, and when the number of dynamic leak-picking orders reaches the preset first threshold, notifying the purchasing staff to replenish the stock and obtaining the first replenishment data", the e-commerce platform first needs to obtain the current number of allocable inventory. The allocable inventory refers to the number of goods that can be immediately used to meet the order requirements in the warehouse. The platform monitors the inventory system in real time to understand the specific situation of the existing inventory, so that timely decisions can be made during the order processing process. Then, the platform will compare and calculate the "number of dynamic leak-picking orders" previously obtained with the current "number of allocable inventory". If the number of dynamic leak-picking orders is greater than or equal to the number of allocable inventory and reaches a preset difference, which is called the "preset first threshold", the platform will trigger a replenishment notification mechanism. This threshold is set to ensure that when the inventory is close to insufficient or about to fail to meet customer demand, the platform can promptly notify the purchasing staff to replenish the stock. For example, suppose an e-commerce platform has a product with a current allocable inventory of 50 pieces, and the number of verified dynamic leak-picking orders has reached 70 pieces. If the first threshold set by the platform is a difference of 20 pieces, then it means that the difference between the number of dynamic leak-picking orders and the number of allocable inventory is exactly 20 pieces, which meets the replenishment trigger condition set by the platform. Therefore, the system will automatically generate a replenishment notice to remind the purchasing staff that they need to replenish at least 20 pieces of inventory to ensure that all the leak-picking orders can be met. "First replenishment data" refers to the specific replenishment quantity and commodity information generated according to this replenishment notice. It is the preliminary replenishment plan made by the purchasing department based on the demand data provided by the system. This data is crucial for the inventory management of e-commerce platforms because it directly affects whether the customer's order needs can be met in a timely manner, and whether it can effectively avoid sales losses caused by insufficient inventory or capital occupation caused by excess inventory. Through this mechanism, e-commerce platforms can dynamically monitor and manage inventory to ensure that when the demand for goods suddenly increases, they can replenish in time and maintain a good inventory turnover rate and customer satisfaction. This approach not only improves the accuracy of inventory management, but also improves the response speed and flexibility of the supply chain, enabling the platform to better respond to market changes and fluctuations in customer demand.
[0028] Step S5, within a specified time before the e-commerce bargain hunting cut-off time, timely remind customers who need to cancel their orders in the first replenishment data, obtain the number of canceled orders, and obtain a second bargain hunting order based on the number of canceled orders and the first replenishment data.
[0029] Specifically, in the step of "within a specified time before the e-commerce platform's order cut-off time, timely remind customers who need to cancel their orders in the first replenishment data, obtain the number of canceled orders, and obtain the second order based on the number of canceled orders and the first replenishment data", the e-commerce platform needs to remind relevant customers within a predetermined time before the order cut-off time, especially those order customers included in the first replenishment data. First, the platform will determine the products that are about to be replenished and the corresponding order customer list based on the first replenishment data generated previously. Then, before the order cut-off time (i.e., the time point when customers can no longer change or cancel orders), the system will actively send reminders to these customers to inform them that if they need to cancel their orders, they can do so within this period of time. This reminder is usually sent to customers in the form of email, SMS or APP notification, in order to give customers the opportunity to make a final decision before the order cut-off time to avoid wasting resources due to unnecessary orders. For example, suppose a customer has booked a popular product on the platform, and this product has been included in the plan of the first replenishment data. One hour before the order cut-off time, the system sends a reminder to the customer, asking him if he still needs this product. If the customer decides to cancel the order, he can do so within the time window provided by the system. Once the cancellation is completed, the system automatically updates the number of canceled orders and removes these canceled orders from the first replenishment data. "Number of canceled orders" refers to the total number of orders that customers actively cancel within a specified time. This number is very important because it directly affects the platform's inventory management and replenishment plan. Based on this number, the platform will adjust the first replenishment data to remove those items that no longer need to be replenished, thereby forming a more accurate second leak-picking order. This second leak-picking order is an optimized order list that includes those orders that are finally confirmed to need replenishment, and it will become the basis for the final replenishment and inventory allocation. This mechanism of early reminder and cancellation of orders can effectively reduce inventory backlogs or waste caused by customers changing their needs or forgetting to cancel. Through this process, the e-commerce platform not only improves operational efficiency, but also enhances customer experience, making the entire leak-picking and replenishment process smoother and more accurate.
[0030] Step S6, determine whether the second bargain hunting order reaches a preset second threshold value. If reached, obtain second replenishment data based on the allocatable inventory quantity and the second bargain hunting order, prioritize and allocate the second bargain hunting order based on the second replenishment data and the allocatable inventory quantity, and obtain the allocation result of the e-commerce bargain hunting order; wherein the second threshold value is the difference between the second bargain hunting order and the allocatable inventory quantity.
[0031] Specifically, in the step of "determining whether the second bargain-hunting order reaches the preset second threshold, if so, obtaining the second replenishment data based on the allocatable inventory quantity and the second bargain-hunting order, prioritizing and allocating the second bargain-hunting order based on the second replenishment data and the allocatable inventory quantity, and obtaining the allocation result of the e-commerce bargain-hunting order", the e-commerce platform must first determine the generated second bargain-hunting orders to determine whether the total number of these orders reaches the preset second threshold. The "second threshold" here is a pre-set reference value, which represents the difference between the number of second bargain-hunting orders and the allocatable inventory quantity. In other words, if the number of second bargain-hunting orders is greater than or equal to the allocatable inventory quantity and reaches this difference, the subsequent operation process is triggered. Assume that in a commodity sales activity, the number of second bargain-hunting orders has reached 80 units, and the current allocatable inventory quantity of the platform is 50 units. If the second threshold preset by the platform is a difference of 30 units (i.e. 80 minus 50 equals 30), then when this condition is met, the platform will generate the second replenishment data based on the current allocatable inventory quantity and the second bargain-hunting order. The second replenishment data is the result of a detailed planning of the goods that will need to be replenished, including which goods need to be replenished, how much to replenish, and other information. Next, the platform will prioritize and allocate the second leak-picking orders based on this second replenishment data and the existing allocable inventory quantity. This priority sorting is usually based on multiple factors, such as the customer's purchase history, order submission time, order payment status, etc. The purpose of sorting is to give priority to those more important or higher priority orders when the inventory is insufficient to meet all orders. For example, in the scenario mentioned above, if the system decides to allocate inventory to customers who have completed payment and have a good purchase history after priority sorting, then the orders of these customers will be processed and allocated first. Finally, the platform generates an allocation result of e-commerce leak-picking orders based on the result of priority sorting. This result will determine which customers can successfully purchase the goods they ordered and which customers' orders need to wait for the next replenishment or be cancelled. This process based on threshold judgment, replenishment planning, priority sorting, and inventory allocation ensures that e-commerce platforms can effectively manage inventory and maximize customer satisfaction when dealing with a large number of leak-picking orders, while avoiding order delays or cancellations due to insufficient inventory. Through this refined management, the platform can improve operational efficiency, increase customer satisfaction, and remain flexible and competitive in the highly competitive e-commerce environment.
[0032] In a specific embodiment, the customer demand information and the sold-out product information are subjected to a comparison analysis to obtain a first bargain hunting order, including: Vectorizing the customer demand information through a preset quantitative evaluation technology to obtain a demand embedding vector; Perform sparse aggregation on the sold-out product information to obtain the sold-out product scarcity vector; Performing a comparison analysis on the demand embedding vector and the sold-out product scarcity vector to obtain a preliminary bargain order; Perform probabilistic deduction on the preliminary bargain hunting order to generate a first bargain hunting order.
[0033] Specifically, in the step of "conducting a comparison analysis of the customer demand information and the sold-out product information to obtain the first bargain order", the e-commerce platform must first conduct a detailed comparison analysis of the customer demand information and the sold-out product information. This process includes multiple steps, the first of which is the processing of the customer demand information. The platform vectorizes the customer demand information through a preset quantitative evaluation technology to obtain a demand embedding vector. In simple terms, vectorization is to convert the customer demand data into a mathematical form (i.e., a vector), which can represent the customer's demand characteristics, such as the type, color, and brand of the product that the customer likes. Next, the platform performs sparse aggregation on the information of the sold-out products to obtain a sold-out product scarcity vector. Sparse aggregation is a data processing technology that can integrate various attributes of the product, especially the information of the products with very low inventory or sold out. Through this aggregation, the platform can identify which products are very scarce in the market and express this information in the form of a vector, i.e., a sold-out product scarcity vector. After completing the generation of the above two vectors, the platform will conduct a comparison analysis of the demand embedding vector and the sold-out product scarcity vector. The comparative analysis process is to determine which customers' needs can be met by picking up leaks by calculating the similarity or matching degree between the two vectors. For example, if a customer's demand embedding vector is very close to the scarcity vector of a sold-out product, it means that the customer's demand and the product are highly matched, and the customer's order may be included in the preliminary leak-picking order. After obtaining the preliminary leak-picking order, the platform will not immediately generate the final leak-picking order, but will further probabilistically deduce these preliminary orders. This step is to calculate the probability of success of each preliminary leak-picking order by analyzing various uncertain factors, such as whether the product will be restocked, whether other customers will cancel their orders, etc. Through these probabilistic calculations, the platform can more accurately predict which orders are most likely to be realized and finally generate the first leak-picking order. For example, suppose a customer really wants to buy a sold-out limited edition sneaker. The platform identifies the customer's demand characteristics through vectorization processing and compares it with the existing scarce product vectors on the market. If it is found that the customer's demand is very matched with this limited edition sneaker, the platform will include the customer's order in the preliminary leak-picking order. Next, the platform will make a probability deduction based on factors such as market inventory and other orders, and finally generate the first leak-picking order to ensure that the customer has a chance to get the shoes when the next restocking occurs. Through this systematic and intelligent processing, e-commerce platforms can not only efficiently manage leak-picking orders for sold-out products, but also better meet customers' personalized needs and improve their shopping experience.
[0034] In a specific embodiment, the authenticity of the first bargain hunting order is verified through the prepayment mechanism to obtain a verified bargain hunting order queue, including: Extracting multi-dimensional information from the first bargain hunting order to obtain an order feature vector; wherein the multi-dimensional information includes the customer's historical purchase behavior, product attributes, and order amount; Performing an authenticity score on the order feature vector using a preset XGBoost algorithm to obtain an order authenticity score; The order authenticity score is divided into credibility levels by a preset dynamic threshold segmentation algorithm to obtain a credibility result; wherein the credibility result includes high-credibility orders and low-credibility orders; Through the smart contract prepayment mechanism of the blockchain, the low-credibility order in the credibility result is initiated, and the customer is required to pay a certain proportion of the prepayment to obtain the prepayment response result; Using a long short-term memory network model to perform a time series analysis on the prepayment response results to predict the customer's prepayment willingness and obtain the customer's prepayment willingness probability; By using a priority queue algorithm, the first bargain hunting orders are sorted and screened based on the customer's prepayment willingness probability and high-credibility orders to obtain a verified bargain hunting order queue.
[0035] Specifically, in the step of "verifying the authenticity of the first bargain hunting order through the prepayment mechanism and obtaining the verified bargain hunting order queue", the e-commerce platform must first perform a series of processing on the first bargain hunting order to ensure the authenticity and reliability of the order. The whole process involves multiple technical links, each of which is implemented through specific algorithms and models, and finally obtains a verified bargain hunting order queue. First, the platform will extract multi-dimensional information for each first bargain hunting order to obtain an order feature vector. The multi-dimensional information here refers to various important data extracted from orders and customers, including the customer's historical purchase behavior, the attributes of the goods, and the order amount. For example, the customer's historical purchase behavior may include the type, frequency, and amount of goods purchased by the customer in the past; the attributes of the goods may include the brand, type, popularity, etc. of the goods; and the order amount directly reflects the total value of the goods purchased by the customer. Through the comprehensive extraction of this information, the platform can generate a specific order feature vector, which can express the various features of the order in mathematical form. Next, the platform uses the preset XGBoost algorithm to score the authenticity of these order feature vectors and obtain the authenticity score of the order. XGBoost is a powerful machine learning algorithm that is particularly suitable for processing complex data sets and performing classification tasks. In this scenario, XGBoost will predict the authenticity of each order based on the order feature vector. The prediction result is presented in the form of a score. The higher the score, the higher the authenticity of the order. Conversely, there may be a risk of false or unreliable. After obtaining the authenticity score of the order, the platform will use the preset dynamic threshold segmentation algorithm to divide the scores into two categories: high-credibility orders and low-credibility orders. The dynamic threshold segmentation algorithm flexibly adjusts the division criteria according to the distribution of the current order scores to ensure that orders with high and low credibility can be reasonably distinguished in different situations. For example, in some cases, the general level of order scores is high. At this time, the algorithm may increase the threshold to ensure that only the orders with the highest scores are considered high-credibility orders; while in the case of generally low order scores, the algorithm may lower the threshold to ensure that too many potentially real orders are not rejected. For those orders marked as low-credibility, the platform will require customers to pay a certain proportion of the advance payment through the blockchain's smart contract advance payment mechanism. The application of blockchain technology here ensures the transparency and immutability of the entire prepayment process, while the smart contract automatically executes this mechanism to ensure that relevant operations are automatically triggered when the prepayment conditions are met. The purpose of this step is to further verify the authenticity of the order by asking the customer to pay the prepayment. If the customer is willing to pay the prepayment, it usually indicates that the customer's demand for the order is real and the reliability of the order is high.After the customer pays the advance payment, the platform will further analyze the advance payment response results. This step is achieved through the long short-term memory network model (LSTM). LSTM is a deep learning model that is particularly good at processing time series data, so it is very suitable for analyzing the customer's advance payment response results. Through the time series analysis of these response results, the LSTM model can predict the customer's willingness to pay in advance and obtain a probability value, which indicates whether the customer is likely to continue to pay the advance payment or complete the order in the future. Finally, the platform will sort and filter the first leak-picking orders according to the probability of customer's willingness to pay in advance and high-confidence orders through the priority queue algorithm, and finally obtain the verification leak-picking order queue. The priority queue algorithm is an algorithm for sorting and processing tasks. Here, it is used to sort all orders so that high-confidence orders and orders with high willingness to pay in advance are processed and allocated first. This algorithm ensures that when inventory is limited, the most likely orders are processed first, while those with higher risks or less reliable orders are delayed or eliminated.
[0036] For example, suppose a customer places an order for limited edition sneakers. After extracting multi-dimensional information about the order, the platform finds that the customer has frequently purchased similar products in the past, and the order amount is large, which indicates that the customer may be a real buyer. Next, the platform generates a high authenticity score for the order through the XGBoost algorithm, but due to the large number of fake orders in the market recently, the platform sets a relatively high threshold, and the order is classified as a low-credibility order. The platform then requires the customer to pay a certain proportion of the advance payment through the blockchain smart contract mechanism. If the customer pays the advance payment within the specified time, the LSTM model analysis results show that the customer's willingness to prepay is more likely. Finally, through the priority queue algorithm, the order is included in the high-priority verification leak order queue to ensure priority processing in the next inventory allocation. Through this series of complex and rigorous steps, the e-commerce platform can effectively screen out the most authentic and reliable leak orders, which not only ensures that customer needs are met, but also avoids resource waste and operational risks caused by fake orders.
[0037] In a specific embodiment, based on the verified bargain hunting order queue, the demand for bargain hunting is statistically calculated to obtain the number of dynamic bargain hunting orders, including: By using a preset timing capture technology, the verification and leak-picking order queue is mapped to a multi-time demand map to capture the changes in the demand for leak-picking orders at different time points; Performing weighted aggregation on the demand changes of the missing order to form a dynamic demand weight matrix; Perform statistical calculation of the demand for leak picking based on the dynamic demand weight matrix to obtain a demand probability distribution; The demand probability distribution is calculated under different confidence intervals to obtain the number of dynamic missing orders.
[0038] Specifically, in the step of "calculating the demand for leak-picking based on the verification leak-picking order queue to obtain the number of dynamic leak-picking orders", the e-commerce platform needs to use a series of technical means to accurately predict the demand for leak-picking orders in order to better manage inventory and meet customer needs. This process starts with the verification of the leak-picking order queue and is gradually completed through a variety of analysis techniques. First, the platform uses the preset time series capture technology to map the verification leak-picking order queue to a multi-time series demand map. Time series capture technology is an analysis tool that can capture the changing trend of data at different time points. In this scenario, the platform will map the order data in the verification leak-picking order queue to a multi-time series demand map in chronological order. This map can vividly show how the demand for leak-picking orders changes at different time points. For example, in certain time periods, the order demand may suddenly increase due to holidays or promotional activities, while in other time periods, the order demand may be relatively stable or decrease. Through this mapping, the platform can clearly see the time change pattern of order demand. Next, the platform performs weighted aggregation on the change in the demand for leak-picking orders to form a dynamic demand weight matrix. Weighted aggregation is a data processing method that emphasizes the importance of different data points by assigning different weights to them. The dynamic demand weight matrix is a table or matrix that records order demand data at different points in time and assigns different weights to these data based on importance or priority. For example, the platform may give higher weights to data in the most recent time period because these data better reflect the current market demand trend; while for more distant historical data, the platform may give lower weights. In this way, the platform is able to form a dynamic demand weight matrix that combines the changing trend of order demand over time and the importance of different data. After obtaining the dynamic demand weight matrix, the platform performs statistical calculations on the demand for leaks based on this matrix to obtain the demand probability distribution. The demand probability distribution is a mathematical model that reflects the probability of order demand occurring under different circumstances. By using the data in the dynamic demand weight matrix, the platform is able to calculate the possibility of order demand at different points in time and under different conditions. This process is similar to probability calculation in statistics, which predicts the results of different situations by combining various factors. For example, if the weight of the order demand for leaks at a certain point in time is high, then the probability of high demand at that point in time will also be high. Then, the platform further analyzes the demand probability distribution and calculates the demand under different confidence intervals, thereby obtaining the number of dynamic missing orders. The confidence interval is a concept in statistics that is used to indicate the credibility of an estimate. In this scenario, the platform will set different confidence intervals, such as 90%, 95% or 99%, which indicate the probability that the order demand falls within a certain range at a specific confidence level.By calculating the demand under these confidence intervals, the platform can accurately estimate how many orders should be prepared under various conditions to ensure that the peak order demand can be met without causing inventory backlogs due to preparing too many orders. For example, suppose a platform is selling a limited edition sneaker. After analyzing the time series capture technology, the platform finds that the order demand for this shoe is usually higher on weekends. Through weighted aggregation, the platform gives a higher weight to the weekend data to form a dynamic demand weight matrix. Subsequently, the platform performs statistical calculations based on this matrix and finds that under the 90% confidence interval, the demand for this shoe is between 100-150 pairs, while under the 95% confidence interval, the demand may be between 120-180 pairs. Therefore, the platform can flexibly decide the number of orders to prepare based on different confidence interval results to ensure that in most cases, the customer's demand for bargain hunting can be met. Through the above series of steps, the e-commerce platform can effectively predict and manage the demand for bargain hunting orders. This data-driven dynamic adjustment method not only improves the efficiency of inventory management, but also improves customer satisfaction because the platform can more accurately meet the actual needs of customers. The entire process combines multiple technical means such as time series analysis, probability statistics and dynamic weight calculation to ensure the accuracy and reliability of the prediction results.
[0039] In a specific embodiment, when the number of dynamic missing orders reaches a preset first threshold, a replenishment notification is sent to the purchasing staff to obtain first replenishment data, including: The real-time dynamic order data flow is monitored through a preset dynamic threshold trigger model. When the number of missed orders is detected to reach a preset first threshold for dynamic order triggering, a corresponding dynamic order trigger signal is generated; Performing calculation and analysis on the dynamic order trigger signal to obtain a replenishment warning matrix; wherein the replenishment warning matrix is used to analyze the current demand intensity, inventory status and supply chain tension of each commodity; Parsing the replenishment warning matrix to obtain a replenishment data stream; wherein the replenishment data stream includes a replenishment quantity, a time window, and a list of commodities with priority replenishment; The replenishment data stream is predicted and analyzed to obtain first replenishment data.
[0040] Specifically, in the step of "when the number of dynamic bargain hunting orders reaches a preset first threshold, a replenishment notification is sent to the purchasing staff to obtain the first replenishment data", the e-commerce platform uses a series of complex technologies and algorithms to ensure that inventory can be replenished in time when the demand for bargain hunting orders surges, so as to avoid out-of-stocks and meet customer needs.
[0041] First, the platform uses the preset dynamic threshold trigger model to monitor the real-time dynamic order data flow. The dynamic threshold trigger model is an intelligent monitoring tool that can monitor and analyze the flow of order data in real time. Specifically, this model will determine whether the preset first threshold is reached based on the real-time changes in the number of missing orders. The so-called first threshold is a baseline set by the platform based on historical data, market demand forecasts and inventory management strategies. When the number of missing orders reaches or exceeds this threshold, it means that the market demand for certain commodities may have increased suddenly, and the inventory may be tight, so the replenishment procedure needs to be started. When it is detected that the number of missing orders reaches the preset first threshold of the dynamic order trigger, the dynamic threshold trigger model will immediately generate a corresponding dynamic order trigger signal. This signal is equivalent to an alarm, prompting the system to further analyze and process the current order demand and inventory situation. The dynamic order trigger signal is the key starting point of the entire replenishment process, which directly determines the subsequent replenishment analysis and decision-making. After generating the dynamic order trigger signal, the platform will perform calculations and analysis on the signal to obtain a replenishment warning matrix. The replenishment warning matrix is a comprehensive data analysis tool that uses computational analysis to comprehensively consider the current demand intensity, inventory status, and supply chain tension of each product. For example, if the demand for a certain product suddenly increases, while inventory is low and the supply chain is also under great pressure, then the replenishment priority of this product will be high in the replenishment warning matrix. This step ensures that the platform can fully understand the current market and inventory status through weighted analysis of multiple factors, so as to make reasonable replenishment decisions. Next, the platform will further analyze the replenishment warning matrix to obtain the replenishment data stream. The replenishment data stream contains specific replenishment information, such as replenishment quantity, time window, and priority replenishment product list. The replenishment quantity refers to the specific number of products that need to be replenished, the time window refers to the time range of replenishment, and the priority replenishment product list lists the types of products that need to be replenished first in the current situation. For example, if the demand for a limited edition sneaker surges and the inventory is in short supply, the replenishment data stream will clearly indicate how much of this shoe needs to be replenished within how long and list it as a priority replenishment product. After receiving the replenishment data stream, the platform will not immediately perform the replenishment operation, but will predict and analyze the replenishment data stream to finally obtain the first replenishment data. This step predicts changes in market demand in the future through in-depth analysis of the replenishment data stream, thereby determining the optimal replenishment strategy. The first replenishment data not only includes specific replenishment quantities and product lists, but may also include more detailed replenishment plans, such as specific replenishment schedules, logistics arrangements, and supply chain collaboration strategies. For example, suppose an e-commerce platform is selling a limited edition pair of sneakers. As the promotion progresses, the platform detects that the number of orders for this pair of sneakers has risen rapidly, reaching the preset first threshold.The dynamic threshold trigger model immediately generated an order trigger signal, and the system then performed an operation and analysis on the signal, finding that the current inventory of this shoe is close to a low level and the supply chain is also in a relatively tight state. Through the analysis of the replenishment warning matrix, the platform obtained a replenishment data stream that required 500 pairs of shoes to be replenished within 24 hours, and listed this shoe as a priority replenishment product. After further prediction and analysis of the replenishment data stream, the platform decided to immediately notify the purchasing staff to perform the replenishment operation, and finally obtained the first replenishment data, which detailed the specific quantity, time schedule, and coordination plan with suppliers. Through this series of steps, the platform can effectively respond to sudden changes in market demand, ensure that inventory can be replenished in a timely manner when the demand for missing orders surges, thereby avoiding out-of-stock situations and improving customers' shopping experience. At the same time, this intelligent replenishment mechanism also greatly improves the efficiency of inventory management and reduces the risks caused by insufficient or excessive inventory.
[0042] In a specific embodiment, timely reminding customers who need to cancel orders in the first replenishment data to obtain the number of canceled orders, and obtaining a second missing order based on the number of canceled orders and the first replenishment data includes: Perform real-time analysis on the customer order status in the first replenishment data through a preset LSTM neural network model to identify customers who need to cancel their orders and generate a list of customers who need to cancel their orders; Based on the list of customers who need to cancel their orders, timely remind the customers who need to cancel their orders, and obtain the number of canceled orders; Using a preset dynamic decision tree network, conditionally screening the number of canceled orders and the first replenishment data to obtain a candidate replenishment combination; wherein the condition screening includes logistics capabilities, seasonal factors, promotional activities and product characteristics, and the product characteristics include shelf life and storage conditions; Performing vectorized calculation of a multi-dimensional replenishment strategy set on the candidate replenishment combination to obtain a demand response function; The demand response function is collaboratively calculated to obtain a conditional optimal solution space; wherein the conditional optimal solution space includes an optimal replenishment quantity, an optimal replenishment time, and an optimal distribution plan; Based on the conditional optimal solution space, the number of cancelled orders and the first replenishment data are hierarchically aggregated to obtain a preliminary second missing order; A time-varying weighted analysis is performed on the preliminary second bargain-hunting order to obtain the second bargain-hunting order.
[0043] Specifically, in the step of "promptly reminding customers who need to cancel orders in the first replenishment data, obtaining the number of canceled orders, and obtaining the second missing orders based on the number of canceled orders and the first replenishment data", the e-commerce platform manages and optimizes the order and replenishment process through a series of advanced data analysis and machine learning models to ensure that after customers cancel their orders, the platform can quickly adjust its strategy to reduce inventory pressure and optimize supply chain management. First, the platform uses a preset LSTM (long short-term memory) neural network model to perform real-time analysis on the customer order status in the first replenishment data. LSTM is a neural network model that can process time series data. It can remember information in long time series and identify patterns that may affect future events. In this scenario, the LSTM model analyzes the customer's order status, such as order creation time, payment status, delivery time and other data, and predicts which orders may be canceled based on historical data. Through this prediction, the platform can generate a list of customers who need to cancel their orders, which includes all customers who may cancel their orders. Next, the platform promptly reminds these customers based on the list of customers who need to cancel their orders. The reminder can be in the form of SMS, email or in-app notification to inform customers that their orders are about to be canceled or need to be confirmed. After receiving the reminder, some customers may confirm to keep the order, while others may choose to cancel the order. The platform monitors these feedbacks in real time and finally obtains the number of canceled orders. After obtaining the number of canceled orders, the platform will use the preset dynamic decision tree network to analyze these data. The dynamic decision tree network is an algorithm that can automatically generate decision paths based on conditions. Here, the platform will conditionally screen possible replenishment combinations based on the number of canceled orders and the first replenishment data. Conditional screening includes multiple aspects such as logistics capabilities, seasonal factors, promotional activities, and product characteristics. For example, if the platform finds that some products have a short shelf life and the current logistics capacity is limited, the system may give priority to replenishing these products to avoid waste and inventory backlogs. Through conditional screening of candidate replenishment combinations, the platform will obtain a list of candidate replenishment combinations. Subsequently, the platform will perform vectorized calculations on these candidate replenishment combinations for multi-dimensional replenishment strategy sets to obtain demand response functions. Vectorized calculation is a mathematical processing method that integrates multi-dimensional data into a vector for subsequent calculation and analysis. The demand response function reflects the platform's ability to respond to changes in market demand under different conditions. For example, if customer demand for a certain product rises sharply over a period of time, the demand response function will help the platform quickly adjust its replenishment strategy to cope with this change in demand. Then, the platform will collaboratively calculate the demand response function to obtain the conditional optimal solution space. The conditional optimal solution space includes the optimal replenishment quantity, optimal replenishment timing, and optimal delivery plan.Collaborative computing here means that the platform will consider multiple factors, such as the number of canceled orders, inventory status, logistics capabilities, etc., to find a balance point, that is, how to achieve the best replenishment effect at the lowest cost under these conditions. Based on the conditional optimal solution space, the platform will hierarchically aggregate the number of canceled orders and the first replenishment data to obtain preliminary second-pickup orders. Hierarchical aggregation is a data processing method that combines data at different levels to form a more comprehensive view. For example, the platform may separate high-priority replenishment items from low-priority items to ensure that key items can be replenished first. Through hierarchical aggregation, the platform can form preliminary second-pickup orders, which are generated based on the latest data and analysis results and can better meet current market demand. Finally, the platform will perform time-varying weighted analysis on the preliminary second-pickup orders to obtain the final second-pickup orders. Time-varying weighted analysis is a dynamic data analysis method that adjusts the importance of data over time. For example, if the demand for a certain product is expected to increase in the future, the platform may give a higher weight to the replenishment demand for this product. Through time-varying weighted analysis, the platform can ensure that the final second-time leak-picking order is more accurate and that the right goods can be replenished into the inventory at the right time. For example, suppose an e-commerce platform is selling a batch of seasonal fruits. As the weather changes and promotional activities are carried out, the platform monitors that some customers may cancel their orders. The LSTM neural network model identifies these orders and generates a list of customers who need to cancel their orders. Subsequently, the platform reminds these customers and finally confirms the number of canceled orders. Next, the dynamic decision tree network analyzes these data and finds which fruits need to be replenished first under the current limited logistics capacity. After vectorization and collaborative calculation of the multi-dimensional replenishment strategy set, the platform obtains the optimal replenishment quantity and replenishment timing, and generates a preliminary second-time leak-picking order based on these data. After performing time-varying weighted analysis, the platform determines the final replenishment plan to ensure that these seasonal fruits can be replenished into the inventory at the best time to meet market demand. Through the above steps, the platform can effectively manage orders and inventory, ensure that when cancellations occur, the replenishment strategy can be quickly adjusted, optimize supply chain management, reduce inventory pressure, and improve customer satisfaction.
[0044] In a specific embodiment, the second bargain hunting orders are prioritized and allocated based on the second replenishment data and the allocatable inventory quantity to obtain allocation results of the e-commerce bargain hunting orders, including: Based on the second replenishment data and the allocatable inventory quantity, a pool of orders to be allocated is obtained; Based on the pool of orders to be allocated, priority reasoning is performed on the second leak-picking orders to obtain a priority scoring matrix; wherein the priority scoring matrix is used to reflect the urgency and importance of each second leak-picking order; Performing hierarchical aggregation on the priority scoring matrix to obtain a multi-dimensional priority scoring function; Dynamically weighting the second bargain hunting order based on the multi-dimensional priority scoring function to obtain an allocation plan for the second bargain hunting order; Probabilistically predicting the allocation scheme through a preset entropy weight priority sorting algorithm to obtain an adaptive allocation network; wherein the adaptive allocation network is used to determine the optimal allocation order; The adaptive allocation network is coordinated and optimized to obtain allocation results of e-commerce missed orders.
[0045] Specifically, in the order allocation system of the e-commerce platform, the process of prioritizing and allocating the second missing orders based on the second replenishment data and the allocable inventory quantity is intended to ensure that orders can be reasonably prioritized when inventory is limited through a series of steps and algorithms, thereby maximizing the overall operational efficiency and customer satisfaction of the platform. First, the platform obtains a "pool of orders to be allocated" by analyzing the second replenishment data and the current allocable inventory quantity. This pool contains all orders that need to be processed, that is, those orders that were not met during the initial replenishment. The second replenishment data provides specific demand information for the order, such as the type and quantity of goods required for each order, while the allocable inventory quantity refers to the actual number of goods currently available in the warehouse to meet these orders. For a simple example, assuming that there are 100 items of a certain item in the warehouse, and there are currently 150 orders that require this item, then 50 orders may not be met in the first round of allocation, and these orders will be placed in the pool of orders to be allocated. The orders in this pool are the objects of the next priority sorting and allocation. Next, the platform needs to perform priority reasoning on these orders to be allocated to obtain a "priority scoring matrix". Priority reasoning is the process of determining the priority of each order by analyzing its urgency and importance. The priority scoring matrix is a table used to record the priority of each order. This table lists the score of each order, and the higher the score, the more urgent and important the order. For example, if a customer's historical purchase record on the platform shows that he is a high-value customer, and the customer's order has a large number of items and a high urgency, then this order may receive a high priority score. In the scoring matrix, the priority of such orders will appear particularly high, and the system will give priority to satisfying such orders. With the priority scoring matrix, the platform will then perform "hierarchical aggregation" to obtain a multi-dimensional priority scoring function. The purpose of hierarchical aggregation is to integrate data from different levels into a comprehensive scoring function. The different levels here refer to different priority evaluation criteria, such as customer importance, order urgency, market demand for goods, etc. By integrating these different levels together, the platform can obtain a multi-dimensional priority scoring function, which can be understood as a complex mathematical formula used to calculate the comprehensive priority of each order. The significance of this step is that it can unify the information of each dimension, making the priority score of each order more accurate and comprehensive. Next, the platform will dynamically weight the second-level leak-picking orders based on this multi-dimensional priority scoring function to obtain a preliminary allocation plan. Dynamic weighting refers to the process of adjusting the priority score based on actual conditions. For example, if the demand for a certain commodity in the market suddenly increases, the system will give higher weights to the related orders in the allocation plan, thereby giving priority to these orders.This dynamic weighting ensures that the allocation plan is flexible and adaptable, and can respond quickly to real-time market changes to avoid improper order processing due to market changes. After obtaining the preliminary allocation plan, the platform will use the "entropy weight priority sorting algorithm" to make a probabilistic prediction of the plan and obtain an "adaptive allocation network". The entropy weight priority sorting algorithm is an algorithm that determines the priority by calculating information entropy. Information entropy can be understood as a tool to measure uncertainty. Here, it is used to measure the processing priority of each order. The adaptive allocation network is a network structure based on these priority sortings, which can automatically adjust and optimize the allocation order of orders. For example, when inventory changes or the number of orders suddenly increases, the adaptive allocation network can automatically adjust the allocation order according to these changes to ensure that the system can cope with various emergencies and maintain the efficiency of allocation. Finally, the platform will "coordinate and optimize" this adaptive allocation network to obtain the final allocation result of the e-commerce leak-picking order. Coordinated optimization refers to the process of finding a balance point between multiple optimization objectives. For e-commerce platforms, these goals may include maximizing customer satisfaction, optimizing inventory utilization, and minimizing overall operating costs. Through coordinated optimization, the platform can find an optimal balance between these goals, ensuring that the final allocation results can both meet customer needs and optimize inventory management. For example, if the system recognizes that the inventory of a certain type of goods is insufficient to meet all orders, it may prioritize customer orders that are most important to the platform while minimizing the impact on other orders, thereby controlling operating costs while maintaining high customer satisfaction. Through these complex steps, e-commerce platforms can ensure that they make the best order allocation decisions when faced with multiple uncertainties and complex situations, thereby improving overall operational efficiency and customer satisfaction. This process is not just simple data processing, but also the comprehensive application of complex algorithms and optimization techniques, which enables the platform to maintain its leading position in the fierce market competition.
[0046] The above describes the method for allocating e-commerce bargain hunting orders in the embodiment of the present invention. The following describes the device for allocating e-commerce bargain hunting orders in the embodiment of the present invention. Figure 2 In one embodiment of the present invention, an apparatus for distributing e-commerce bargain hunting orders includes: The analysis module 21 is used to obtain the sold-out product information and customer demand information when the product is sold out, and compare and analyze the customer demand information and the sold-out product information to obtain the first bargain order; A verification module 22, used to verify the authenticity of the first bargain hunting order through a prepayment mechanism, and obtain a verified bargain hunting order queue; A calculation module 23 is used to perform statistical calculation of the demand for leak picking based on the verified leak picking order queue to obtain the number of dynamic leak picking orders; The notification module 24 is used to obtain the allocable inventory quantity. When the number of dynamic bargain hunting orders reaches a preset first threshold, a replenishment notification is sent to the purchasing staff to obtain first replenishment data; wherein the preset first threshold is the difference between the number of dynamic bargain hunting orders and the number of allocable inventory; The reminder module 25 is used to timely remind customers who need to cancel orders in the first replenishment data within a specified time before the e-commerce bargain hunting cut-off time, obtain the number of canceled orders, and obtain the second bargain hunting order based on the number of canceled orders and the replenishment data; The allocation module 26 is used to determine whether the second bargain hunting order reaches a preset second threshold value. If reached, second replenishment data is obtained based on the allocatable inventory quantity and the second bargain hunting order, and the second bargain hunting order is prioritized and allocated based on the second replenishment data and the allocatable inventory quantity to obtain the allocation result of the e-commerce bargain hunting order; wherein the second threshold value is the difference between the second bargain hunting order and the allocatable inventory quantity.
[0047] In this embodiment, for the specific implementation of each unit in the above system embodiment, please refer to the description in the above method embodiment, which will not be repeated here.
[0048] Reference Figure 3 The present invention also provides a computer device in an embodiment, wherein the internal structure of the computer device can be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a display screen, an input device, a network interface and a database connected through a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the above method is implemented.
[0049] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied.
[0050] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the above method is implemented. It can be understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.
[0051] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided by the present invention and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM.
[0052] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.
[0053] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.
Claims
1. A method for allocating e-commerce bargain hunting orders, characterized in that: The following steps are involved: When the product is sold out, obtain the sold-out product information and customer demand information, compare and analyze the customer demand information and the sold-out product information to obtain the first bargain order; The authenticity of the first bargain hunting order is verified through the prepayment mechanism to obtain a verified bargain hunting order queue; Based on the verified leak-picking order queue, a statistical calculation of leak-picking demand is performed to obtain a dynamic leak-picking order quantity; Obtain the allocable inventory quantity, and when the number of dynamic bargain hunting orders reaches a preset first threshold, notify the purchasing staff to replenish the goods, and obtain first replenishment data; wherein the preset first threshold is the difference between the number of dynamic bargain hunting orders and the allocable inventory quantity; Within a specified time before the e-commerce discount order cut-off time, timely remind customers who need to cancel orders in the first replenishment data, obtain the number of canceled orders, and obtain a second discount order based on the number of canceled orders and the first replenishment data; Determine whether the second bargain hunting order reaches a preset second threshold value. If reached, obtain second replenishment data based on the allocatable inventory quantity and the second bargain hunting order, prioritize and allocate the second bargain hunting order based on the second replenishment data and the allocatable inventory quantity, and obtain an allocation result of the e-commerce bargain hunting order; wherein the second threshold value is the difference between the values of the second bargain hunting order and the allocatable inventory quantity.
2. The method for allocating e-commerce bargain hunting orders according to claim 1, characterized in that: The customer demand information and the sold-out product information are compared and analyzed to obtain a first bargain order, including: Vectorizing the customer demand information through a preset quantitative evaluation technology to obtain a demand embedding vector; Perform sparse aggregation on the sold-out product information to obtain the sold-out product scarcity vector; Performing a comparison analysis on the demand embedding vector and the sold-out product scarcity vector to obtain a preliminary bargain order; Perform probabilistic deduction on the preliminary bargain hunting order to generate a first bargain hunting order.
3. The method for allocating e-commerce bargain hunting orders according to claim 1, characterized in that: The authenticity of the first bargain hunting order is verified through the prepayment mechanism to obtain a verified bargain hunting order queue, including: Extracting multi-dimensional information from the first bargain hunting order to obtain an order feature vector; wherein the multi-dimensional information includes the customer's historical purchase behavior, product attributes, and order amount; Performing an authenticity score on the order feature vector using a preset XGBoost algorithm to obtain an order authenticity score; The order authenticity score is divided into credibility levels by a preset dynamic threshold segmentation algorithm to obtain a credibility result; wherein the credibility result includes high-credibility orders and low-credibility orders; Through the smart contract prepayment mechanism of the blockchain, the low-credibility order in the credibility result is initiated, and the customer is required to pay a certain proportion of the prepayment to obtain the prepayment response result; Using a long short-term memory network model to perform a time series analysis on the prepayment response results to predict the customer's prepayment willingness and obtain the customer's prepayment willingness probability; By using a priority queue algorithm, the first bargain hunting orders are sorted and screened based on the customer's prepayment willingness probability and high-credibility orders to obtain a verified bargain hunting order queue.
4. The method for allocating e-commerce bargain hunting orders according to claim 1, characterized in that: Based on the verified leak-picking order queue, the leak-picking demand quantity is statistically calculated to obtain the dynamic leak-picking order quantity, including: By using a preset timing capture technology, the verification order queue is mapped to a multi-time demand map to capture the change in demand for the order at different time points; Performing weighted aggregation on the demand changes of the missing order to form a dynamic demand weight matrix; Perform statistical calculation of the demand for leak picking based on the dynamic demand weight matrix to obtain a demand probability distribution; The demand probability distribution is calculated under different confidence intervals to obtain the number of dynamic missing orders.
5. The method for allocating e-commerce bargain hunting orders according to claim 1, characterized in that: When the number of dynamic missing orders reaches a preset first threshold, a replenishment notification is sent to the purchasing staff to obtain first replenishment data, including: The real-time dynamic order data flow is monitored through a preset dynamic threshold trigger model. When the number of missed orders is detected to reach a preset first threshold for dynamic order triggering, a corresponding dynamic order trigger signal is generated; Performing calculation and analysis on the dynamic order trigger signal to obtain a replenishment warning matrix; wherein the replenishment warning matrix is used to analyze the current demand intensity, inventory status and supply chain tension of each commodity; Parsing the replenishment warning matrix to obtain a replenishment data stream; wherein the replenishment data stream includes a replenishment quantity, a time window, and a list of commodities to be replenished first; The replenishment data stream is predicted and analyzed to obtain first replenishment data.
6. The method for allocating e-commerce bargain hunting orders according to claim 1, characterized in that: Promptly reminding customers who need to cancel orders in the first replenishment data to obtain the number of canceled orders, and obtaining a second missing order based on the number of canceled orders and the first replenishment data, including: Perform real-time analysis on the customer order status in the first replenishment data through a preset LSTM neural network model to identify customers who need to cancel their orders and generate a list of customers who need to cancel their orders; Based on the list of customers who need to cancel their orders, timely remind the customers who need to cancel their orders, and obtain the number of canceled orders; Using a preset dynamic decision tree network, conditionally screening the number of canceled orders and the first replenishment data to obtain a candidate replenishment combination; wherein the condition screening includes logistics capabilities, seasonal factors, promotional activities and product characteristics, and the product characteristics include shelf life and storage conditions; Performing vectorized calculation of a multi-dimensional replenishment strategy set on the candidate replenishment combination to obtain a demand response function; The demand response function is collaboratively calculated to obtain a conditional optimal solution space; wherein the conditional optimal solution space includes an optimal replenishment quantity, an optimal replenishment time, and an optimal distribution plan; Based on the conditional optimal solution space, the number of cancelled orders and the first replenishment data are hierarchically aggregated to obtain a preliminary second missing order; A time-varying weighted analysis is performed on the preliminary second bargain-hunting order to obtain the second bargain-hunting order.
7. The method for allocating e-commerce bargain hunting orders according to claim 6, characterized in that: Prioritizing and allocating the second bargain hunting orders based on the second replenishment data and the allocable inventory quantity to obtain an allocation result of the e-commerce bargain hunting orders, including: Based on the second replenishment data and the allocatable inventory quantity, a pool of orders to be allocated is obtained; Based on the pool of orders to be allocated, priority reasoning is performed on the second leak-picking orders to obtain a priority scoring matrix; wherein the priority scoring matrix is used to reflect the urgency and importance of each second leak-picking order; Performing hierarchical aggregation on the priority scoring matrix to obtain a multi-dimensional priority scoring function; Dynamically weighting the second bargain hunting order based on the multi-dimensional priority scoring function to obtain an allocation plan for the second bargain hunting order; Probabilistically predicting the allocation scheme through a preset entropy weight priority sorting algorithm to obtain an adaptive allocation network; wherein the adaptive allocation network is used to determine the optimal allocation order; The adaptive allocation network is coordinated and optimized to obtain allocation results of e-commerce missed orders.
8. A device for allocating orders for e-commerce bargain hunting, characterized in that: include: An analysis module is used to obtain sold-out product information and customer demand information when the product is sold out, and compare and analyze the customer demand information and the sold-out product information to obtain a first bargain order; A verification module, used to verify the authenticity of the first bargain hunting order through a prepayment mechanism, and obtain a verified bargain hunting order queue; A calculation module is used to perform statistical calculation of the demand for leak picking based on the verified leak picking order queue to obtain the number of dynamic leak picking orders; A notification module is used to obtain the allocable inventory quantity, and when the number of dynamic bargain hunting orders reaches a preset first threshold, a replenishment notification is sent to the purchasing staff to obtain first replenishment data; wherein the preset first threshold is the difference between the number of dynamic bargain hunting orders and the number of allocable inventory; A reminder module is used to timely remind customers who need to cancel orders in the first replenishment data within a specified time before the e-commerce discount order cut-off time, obtain the number of canceled orders, and obtain a second discount order based on the number of canceled orders and the replenishment data; The allocation module is used to determine whether the second bargain hunting order reaches a preset second threshold value. If reached, second replenishment data is obtained based on the allocatable inventory quantity and the second bargain hunting order, and the second bargain hunting order is prioritized and allocated based on the second replenishment data and the allocatable inventory quantity to obtain the allocation result of the e-commerce bargain hunting order; wherein the second threshold value is the difference between the second bargain hunting order and the allocatable inventory quantity.
9. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.
Citation Information
Cited By
Order payment management method and device, equipment and storage medium
CN120125311A
Order payment management method and device, equipment and storage medium
CN120125311B
