Supply chain exception handling method, apparatus, device, storage medium, and program product

Abstract

This application provides a supply chain anomaly handling method, apparatus, equipment, storage medium, and program product. The method includes: determining at least one related indicator corresponding to a target indicator in the supply chain; wherein the target indicator is an indicator corresponding to a target business in the supply chain, and the related indicator is an indicator corresponding to a business associated with the target business in the supply chain; determining an abnormal indicator based on the anomaly degree of the target indicator and each related indicator; wherein the abnormal indicator is the indicator among the at least one related indicator that causes the target indicator to be abnormal; and, for the abnormal indicator, determining the dimensions and elements among the multiple dimensions that cause the target indicator to be abnormal based on multiple preset dimensions, according to the anomaly degree corresponding to each dimension and the contribution degree corresponding to each element under each dimension, thereby quickly and accurately diagnosing the cause of anomalies in the supply chain business and improving the efficiency and accuracy of supply chain diagnosis.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method, apparatus, equipment, storage medium, and program product for handling supply chain anomalies. Background Technology

[0002] With the continuous development of the social economy, the supply chain of goods involves more and more businesses, and the interdependencies between these businesses are becoming increasingly complex. Furthermore, various business processes within the supply chain, such as replenishment plans and sales plans, may experience anomalies for various reasons, affecting the production and distribution of goods. Therefore, monitoring supply chain operations has become a hot topic.

[0003] In practical applications, when certain supply chain operations experience anomalies, R&D personnel often spend a lot of time investigating the possible causes of the anomalies. This not only affects the normal operation of the supply chain operations but may also lead to problems such as product shortages and stockpiles, affecting the normal flow of goods in the supply chain, resulting in low product processing efficiency and high warehousing costs. Summary of the Invention

[0004] The main objective of this application is to provide a supply chain anomaly handling method, apparatus, equipment, storage medium, and program product to quickly and accurately locate the causes of anomalies in supply chain operations and improve the efficiency and accuracy of supply chain anomaly diagnosis.

[0005] In a first aspect, embodiments of this application provide a supply chain anomaly handling method, including:

[0006] Identify at least one related indicator corresponding to the target indicator of the supply chain; wherein the target indicator is the indicator corresponding to the target business in the supply chain, and the related indicator is the indicator corresponding to the business in the supply chain that is associated with the target business.

[0007] Based on the degree of abnormality of the target indicator and each related indicator, an abnormal indicator is determined, wherein the abnormal indicator is the indicator among the at least one related indicator that causes the target indicator to be abnormal.

[0008] For the aforementioned abnormal indicators, based on multiple preset dimensions, and according to the degree of abnormality corresponding to each dimension and the contribution of each element under each dimension, the abnormal dimensions and elements are determined among the multiple dimensions.

[0009] Among them, the degree of anomaly is used to represent the difference between the actual value and the predicted value of the corresponding indicator, and the contribution is used to represent the proportion of the fluctuation change of the element in the fluctuation change of the abnormal indicator.

[0010] Optionally, at least one related indicator corresponding to the target indicator of the supply chain is determined, including:

[0011] The first service on which the target service is determined;

[0012] Based on the first service and the second service on which the first service is produced, identify the associated services, and determine the at least one associated indicator based on the associated services.

[0013] Optionally, the method further includes:

[0014] For the target indicator, based on multiple preset dimensions, and according to the degree of anomaly corresponding to each dimension and the contribution of each element under each dimension, the abnormal dimensions and elements are determined; and / or,

[0015] Display at least one of the following information for each of the target indicators and related indicators: the degree of anomaly of the indicator, the degree of anomaly of each dimension, and the contribution of each element under each dimension.

[0016] Optionally, the method further includes:

[0017] After determining the dimension and element of the anomaly corresponding to any indicator, the corresponding cause comparison tree is found according to the indicator; the cause comparison tree includes multiple nodes forming a binary tree, wherein the leaf nodes are used to represent the candidate causes of the corresponding business anomaly under the dimension and element; and the branch nodes are used to represent the judgment conditions set for the corresponding business.

[0018] Based on the actual completion status of the corresponding business under the dimensions and elements, and the cause comparison tree, select the cause that leads to the abnormality of the indicator from the candidate causes.

[0019] Optionally, the method further includes at least one of the following:

[0020] Update the configuration parameters corresponding to the cause of the abnormality in the indicator;

[0021] If the abnormality of the target indicator exceeds a preset threshold, a first alarm message is output.

[0022] Identify the user responsible for maintaining the metrics of the dimensions and elements of the anomaly, and output a second alarm message to the user.

[0023] Optionally, for the aforementioned anomaly indicators, based on multiple preset dimensions, and according to the degree of anomaly corresponding to each dimension and the contribution of each element under each dimension, the abnormal dimensions and elements are determined, including:

[0024] For each dimension, based on the predicted and actual values ​​of the abnormal indicators corresponding to each element under the dimension, the degree of abnormality and contribution of each element are calculated, and the degree of abnormality of multiple elements under the dimension is added together to obtain the degree of abnormality of the dimension.

[0025] The degree of anomaly in each dimension is ranked, and the dimension of anomaly is determined based on the ranking result.

[0026] Based on the contribution of each element in the anomaly dimension, the anomalous element is determined from multiple elements in the dimension.

[0027] Optionally, the target indicator and the associated indicator are indicators of the same type, wherein the types of indicators include: output, output rate, and accuracy.

[0028] The actual value corresponding to the output is the number of portions actually produced, and the predicted value is the number of portions that should be produced.

[0029] The actual value corresponding to the output rate is the ratio of the actual number of portions produced to the upper limit of the number of portions, and the predicted value is the ratio of the number of portions that should be produced to the upper limit of the number of portions.

[0030] The accuracy rate corresponds to the actual business data and the predicted business data, respectively.

[0031] Secondly, embodiments of this application also provide a supply chain anomaly handling device, comprising:

[0032] The determination module is used to determine at least one related indicator corresponding to the target indicator of the supply chain; wherein the target indicator is the indicator corresponding to the target business in the supply chain, and the related indicator is the indicator corresponding to the business associated with the target business in the supply chain.

[0033] The diagnostic module is used to determine abnormal indicators based on the degree of abnormality of the target indicator and each related indicator, wherein the abnormal indicator is the indicator among the at least one related indicator that causes the target indicator to be abnormal.

[0034] The processing module is used to determine the abnormal dimensions and elements in the multiple dimensions based on the preset multiple dimensions, according to the degree of abnormality corresponding to each dimension and the contribution of each element under each dimension.

[0035] Among them, the degree of anomaly is used to represent the difference between the actual value and the predicted value of the corresponding indicator, and the contribution is used to represent the proportion of the fluctuation change of the element in the fluctuation change of the abnormal indicator.

[0036] Thirdly, embodiments of this application provide an electronic device, including:

[0037] At least one processor; and

[0038] A memory that is communicatively connected to the at least one processor;

[0039] The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, cause the electronic device to perform the method described in any of the above aspects.

[0040] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the method described in any of the above aspects.

[0041] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the methods described in any of the above aspects.

[0042] The supply chain anomaly handling method, apparatus, equipment, storage medium, and program products provided in this application can determine at least one related indicator corresponding to a target indicator in the supply chain. The target indicator is an indicator corresponding to a target business in the supply chain, and the related indicator is an indicator corresponding to a business associated with the target business in the supply chain. Based on the anomaly degree of the target indicator and each related indicator, anomaly indicators are determined. The anomaly indicator is the indicator among the at least one related indicator that causes the target indicator to be abnormal. For the anomaly indicator, based on multiple preset dimensions, and according to the anomaly degree corresponding to each dimension and the contribution of each element under each dimension, the dimensions and elements among the multiple dimensions that cause the target indicator to be abnormal are determined. This allows for rapid and accurate diagnosis of data anomalies in the supply chain business, improving the timeliness and accuracy of supply chain business output, enhancing the executability of the supply chain business, facilitating the automation of supply chain business optimization, effectively improving the efficiency and accuracy of supply chain diagnosis, meeting the normal usage needs of operational personnel, reducing problems such as product shortages and backlogs caused by supply chain business anomalies, maintaining the normal flow of goods in the supply chain, improving product processing efficiency, and reducing warehousing costs. Attached Figure Description

[0043] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0044] Figure 1 This is a schematic diagram of an application scenario provided by an embodiment of this application;

[0045] Figure 2 This is a schematic diagram illustrating another application scenario provided by an embodiment of this application;

[0046] Figure 3 A diagnostic diagram of a replenishment plan provided in an embodiment of this application;

[0047] Figure 4 A flowchart illustrating a supply chain anomaly handling method provided in this application embodiment;

[0048] Figure 5 A dependency diagram of a replenishment plan provided for an embodiment of this application;

[0049] Figure 6 A schematic diagram of a cause-and-effect tree provided for an embodiment of this application;

[0050] Figure 7 A system architecture diagram for handling supply chain anomalies is provided in this application embodiment;

[0051] Figure 8 A flowchart illustrating the process of determining abnormal dimensions and elements provided in this application embodiment;

[0052] Figure 9 This is a schematic diagram of the structure of a supply chain anomaly handling device provided in an embodiment of this application;

[0053] Figure 10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0054] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0055] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0056] First, let me explain the terms used in this application:

[0057] Attribution: Attribution theory refers to the process by which people infer the causes of the behavior of others or themselves. In the application scenario of this application, it can refer to the causal explanation and inference of abnormal indicators through reasoning.

[0058] Supply chain operations: Any business involved in the supply chain, which can be planning-related business, such as replenishment plans and sales plans, or other business, such as inventory data.

[0059] The embodiments of this application can be used to process supply chain data in any field, especially for real-time monitoring of supply chain plans. For supply chain planning, the planned output quantity and accuracy are usually affected by multiple data. When anomalies occur, it is necessary to attribute the anomalies in the supply chain plan, quickly locate the root cause of the anomaly, thereby improving the timeliness, accuracy and executability of the supply chain plan output, and enhancing the automation level of supply chain optimization.

[0060] In practical applications, supply chains often involve multiple roles, which may have upstream and downstream relationships, and one or more of these roles may have plans for replenishment, sales, etc. Figure 1 This is a schematic diagram illustrating an application scenario provided by an embodiment of this application. For example... Figure 1 As shown, taking e-commerce as an example, the supply chain can include multiple roles such as consumers, distributors, and suppliers. Distributors can purchase goods from suppliers, replenish their own warehouses with goods from the suppliers' warehouses, and when consumers purchase goods, distributors directly retrieve the goods from their own warehouses and ship them to the consumers. To achieve more precise replenishment, corresponding replenishment plans can be configured for distributors, such as when, from which warehouse, and how many of which types of goods to replenish to their own warehouses.

[0061] Figure 2 This is a schematic diagram illustrating another application scenario provided by an embodiment of this application. For example... Figure 2 As shown, multiple roles in the supply chain can include consumers, distributors, suppliers, and raw material providers. Manufacturers can purchase components from raw material providers and assemble them into marketable goods. For example, they can purchase tires, glass, and other components from multiple raw material providers and manufacture products such as automobiles. The goods then flow to at least one downstream distributor, who then sells them to consumers. In this scenario, corresponding replenishment plans can be configured for both the manufacturer and the distributor.

[0062] In addition to replenishment plans, sales plans and allocation plans can also be set. Optionally, in the supply chain plan, a replenishment plan can be the quantity of goods to be replenished from the upstream warehouse to the current warehouse, specifically the quantity needed for the next week or each day; a sales plan can be the quantity of goods expected to be sold in the next period; multiple warehouses can be set up for the same role, and goods can be transferred between these warehouses. For example, a distributor has two warehouses, A and B, located in different shipping locations. A strategy of shipping from the nearest warehouse is used. If warehouse A is expected to have high sales volume and low inventory in the next period, a shortage of goods may occur. Therefore, a portion of goods can be transferred from warehouse B to warehouse A in advance. Correspondingly, the allocation plan can be the quantity of goods transferred from one warehouse to another within the same role.

[0063] Optionally, each role can purchase or sell multiple types of goods, and each role can also have multiple warehouses. Each supply chain plan can be generated for a warehouse, for a product, or for both products and warehouses.

[0064] For example, a replenishment plan can be generated for each type of product. When there are multiple warehouses, the replenishment plan can include the replenishment quantity for each warehouse. Multiple replenishment plans can be generated for multiple products. Alternatively, replenishment plans can be generated separately for the warehouse and product dimensions. For example, product 1 in warehouse A can have a corresponding replenishment plan, product 2 in warehouse A can also have a corresponding replenishment plan, and product 1 in warehouse B can also have a corresponding replenishment plan.

[0065] For supply chain planning, the output is typically based on multiple data sources, using algorithms for prediction or data processing. This results in a large amount of dependent data and a long computational chain. For example, replenishment plans can be dynamic, with the quantity of goods replenished to the warehouse each time determined in advance based on factors such as predicted sales and inventory. Inaccuracies in replenishment plans may stem from inaccurate sales forecasts, inaccurate inventory data, or inaccurate replenishment parameter configurations. Therefore, replenishment plans are interconnected with sales plans, inventory, and parameter configurations.

[0066] During routine operations, operations personnel can monitor metrics such as replenishment plans, sales plans, and allocation plans. When operations personnel report errors in data for certain products or suppliers, development personnel spend a significant amount of time investigating the data and processes that may cause the anomalies. These anomalies are usually the result of multiple data anomalies across multiple dimensions or dependencies. Due to the long supply chain planning process and the large amount of dependent data, troubleshooting requires checking each node one by one, which is time-consuming and labor-intensive. This also affects the normal operation of operations personnel and may lead to product shortages, stockpiles, and other problems, disrupting the normal flow of goods in the supply chain, resulting in low product processing efficiency and high warehousing costs.

[0067] In view of this, the embodiments of this application can summarize the accuracy and productivity issues in supply chain business into indicators, correlate the various indicators, analyze and process abnormal indicators through multiple dimensions, monitor various indicators in real time, proactively discover anomalies, and quickly and accurately locate the root cause of anomalies. This is especially suitable for anomaly discovery when there are multiple data sources and upstream and downstream dependencies on various data processing and aggregation.

[0068] Optionally, supply chain operations may include supply chain planning or other operations. Each supply chain operation can have multiple types of metrics, including but not limited to: output, output rate, accuracy, etc. Supply chain plans can be generated separately for each product. Taking replenishment planning as an example, output can refer to the quantity of goods produced according to the replenishment plan; output rate can refer to the ratio of the quantity of goods produced according to the replenishment plan to the total quantity of goods in the warehouse; and accuracy can refer to the difference between the replenishment quantity in the replenishment plan and the actual replenishment quantity.

[0069] Figure 3 This is a diagnostic diagram of a replenishment plan provided in an embodiment of this application. Figure 3 As shown, replenishment planning relies on business data such as product pool data, sales plans, product-warehouse relationships, planning parameters, and inventory data. These business data can be regarded as upstream indicators of replenishment planning and are among the factors affecting replenishment planning.

[0070] exist Figure 3 In the scheme shown, the target indicator that needs to be monitored can be the replenishment plan output rate, and its related indicator can be the output rate of the relevant business. These will be explained below.

[0071] Optionally, the product pool data can refer to the identifiers of products that require replenishment plans. Optionally, the product pool contains identifiers of multiple products (which can be all products involved in the current role's warehouse). Among them, only products that meet preset conditions may require replenishment plans. For example, only products with sales exceeding a certain value in the most recent month will have replenishment plans generated, while products with sales not meeting the requirements will not need replenishment plans generated. The preset conditions can be set by the operations staff or use default rules.

[0072] When the target metric is the replenishment plan output rate, the product pool data output rate, as a corresponding related metric, can refer to the ratio of the number of products that need to generate a replenishment plan to the total number of products in the warehouse. For example, if there are 100 products in total, and according to the preset rules, only 70 products need to generate a replenishment plan, then the corresponding product pool data output rate can be 0.7.

[0073] Optionally, the product-warehouse relationship refers to the relationship between a product and a warehouse, which can be configured by operations personnel. For example, it can be configured whether a certain product corresponds to a certain warehouse. In practical applications, the same product may need to be replenished in multiple warehouses. For instance, a distributor may have warehouses in multiple cities, and different warehouses may require different quantities of products. Therefore, when calculating replenishment plans, the product-warehouse relationship can be used: if a product corresponds to a certain warehouse, it means that a replenishment plan can be calculated for that product in that warehouse; if a product does not correspond to a certain warehouse, it means that the product does not need to be allocated to that warehouse, and therefore no corresponding replenishment plan needs to be calculated.

[0074] Correspondingly, the product-warehouse relationship output rate can be defined as the ratio of the actual number of goods with product-warehouse relationships to the total number of goods. Similarly, the sales plan output rate can be defined as the ratio of the actual number of products produced according to the sales plan to the total number of goods.

[0075] Optionally, the planning parameters can be those needed when calculating the supply chain plan. For replenishment planning, parameters such as maximum number of days to sell, replenishment lead time, and replenishment calendar can be configured in advance. These parameters represent the maximum number of days a product can be listed for sale, the minimum number of days in advance that replenishment needs to be made, and the specific day for replenishment (e.g., replenishment every Monday). These parameters may affect the output of the replenishment plan.

[0076] Correspondingly, the planned parameter output rate can be statistically analyzed from the perspective of commodities and / or parameters. It can be the ratio of the number of planned parameters actually configured to the number of planned parameters that should be configured, or the ratio of the number of commodities with planned parameters actually configured to the total number of commodities, or the ratio of the sum of the number of planned parameters actually configured for each commodity to the sum of the number of planned parameters for all commodities.

[0077] Optionally, inventory data may include inventory data in warehouses and / or in transit. Inventory data in warehouses may include the quantity of goods currently in the warehouse, and inventory data in transit may include the quantity of goods currently in transit. These two data may affect replenishment plans. For example, if the quantity of goods in warehouses and in transit increases, the corresponding replenishment plan value may decrease, assuming other influencing factors remain unchanged.

[0078] The output rate of inventory data can refer to the ratio of the number of goods actually produced to the total number of goods. For example, if there are 100 kinds of goods, but only the inventory data of 60 kinds of goods is actually produced, then the corresponding inventory data output rate can be 0.6.

[0079] It should be noted that when the indicators are output quantity and output rate, the focus should be on the actual number of units produced. Whether the output data is accurate can be disregarded for the time being. For example, as long as all products have produced inventory data, the inventory data output rate can be considered to be 100%. Whether each unit of inventory data is accurate is not assessed in the output rate indicator, but can be assessed in the accuracy rate indicator.

[0080] After determining the aforementioned related indicators for the supply chain replenishment plan, the indicators causing the abnormal replenishment plan output rate can be identified based on the abnormality of the replenishment plan output rate and the degree of abnormality of each related indicator. For example, the degree of abnormality can be represented by a value from 0 to 100, with higher values ​​indicating a higher degree of abnormality. Optionally, a corresponding threshold can be set, such as 60, where values ​​greater than this threshold are considered abnormal.

[0081] For example, assuming the abnormality of the replenishment plan output rate is 80, the abnormality of the product pool data output rate is 70, and the abnormality of other indicators is relatively small, then it can be considered that the reason for the abnormality of the replenishment plan output rate is the abnormality of the product pool data output rate.

[0082] After identifying the outlier metric as the product pool data output rate, root cause drilling can be performed based on multiple preset dimensions and each element within each dimension. See also Figure 3 Multiple dimensions can include: region, supplier, merchant, etc. Each dimension can contain multiple elements.

[0083] For example, the region dimension can include various countries and regions globally, such as India, Thailand, and Malaysia, represented in the diagram as country A, country B, country C, etc. Suppliers include the major supplier companies that replenish stock for the current role. Merchants can include various merchants corresponding to the current role, such as the manufacturer.

[0084] In the process of root cause drilling down on the product pool data output rate, considering that each element under different dimensions can have corresponding product pool data, for example, A produced 100 pieces of product pool data and B produced 50 pieces of product pool data, the degree of anomaly in each dimension can be further determined based on the product pool data of each element under each dimension.

[0085] After determining the degree of abnormality in each dimension, one or more dimensions with the highest degree of abnormality can be selected. These dimensions are the most likely to be abnormal. Furthermore, one or more elements with the greatest contribution can be selected from the multiple elements of the abnormal dimensions. These elements are the root causes of the abnormal output rate of the replenishment plan.

[0086] For example, in the product pool data output rate, the region dimension shows the greatest anomaly, and within the region dimension, country A contributes the most, indicating that country A's product pool data output rate is the root cause of the abnormal replenishment plan output rate.

[0087] In practical applications, business data such as product pool data and sales plans can be generated offline or in real time. A certain amount of business data may be generated every day or every week. After processing this business data according to certain rules, a replenishment plan can be obtained. However, problems may occur in the production and processing of business data. For example, the system settings may be incorrect, or the server may fail to produce product pool data, resulting in abnormal replenishment plans. Therefore, the solution provided in the embodiments of this application can be used to monitor and analyze the replenishment plan and its dependent upstream data to determine the root cause of the anomaly.

[0088] Once the root cause is identified, further processing can be performed on the corresponding indicators, dimensions, and elements. For example, if there are anomalies in the product pool data of country A, the output rules for that data can be checked or modified. The cause of the anomalies can also be investigated, such as whether a server malfunction prevented the program from running. After modification or rerunning, the degree of anomaly in the replenishment plan can be reassessed until the replenishment plan returns to normal. The warehouse can then be replenished according to the replenishment plan to maintain the normal operation of the supply chain.

[0089] In summary, by monitoring the anomalies of replenishment plans and their corresponding related indicators in the supply chain, identifying the anomalous indicators based on the degree of anomaly, and then drilling down to the root causes of these anomalies to determine the dimensions and elements leading to the replenishment plan, we can quickly and accurately diagnose data anomalies in the supply chain plan. This improves the timeliness and accuracy of supply chain plan outputs, enhances the executability of the supply chain plan, facilitates the automation of supply chain plan optimization, effectively improves the efficiency of supply chain diagnosis, meets the normal usage needs of operations personnel, reduces problems such as product shortages and backlogs caused by supply chain plan anomalies, maintains the normal flow of goods in the supply chain, improves the efficiency of goods handling, and reduces warehousing costs.

[0090] Furthermore, the embodiments of this application can be used to solve the problem of anomaly detection during online operation and maintenance, and realize real-time diagnosis and anomaly handling of the supply chain. In addition, when diagnosing anomalies, this application only needs to import multiple dimensions and elements, without listing all possible specific causes, which is easy to implement. When different data have upstream and downstream dependencies, but there is no clear calculation formula to derive, the method of this application can effectively find the root cause of the anomaly, greatly shortening the time for locating the problem.

[0091] Besides replenishment plans, similar methods can be used to process other supply chain plans, such as allocation plans and sales plans. Furthermore, other supply chain operations can be processed. These operations can be any business involved in the supply chain. For example, inventory data output may be related to inventory parameter configurations, which vary across different countries and regions. Therefore, inventory data can be considered a type of supply chain operation, with inventory parameter configuration as one of its related indicators and region as one of the dimensions. The aforementioned methods can be used to monitor inventory data output, output rate, or accuracy.

[0092] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Where there is no conflict between the embodiments, the following embodiments and features can be combined with each other. Furthermore, the timing of the steps in the following method embodiments is merely an example and not a strict limitation.

[0093] Figure 4 This is a flowchart illustrating a supply chain anomaly handling method provided in an embodiment of this application. The execution entity of this method can be any device with data processing capabilities, such as a terminal device or a server. Figure 4 As shown, the method may include:

[0094] Step 401: Determine at least one related indicator corresponding to the target indicator of the supply chain; wherein, the target indicator is the indicator corresponding to the target business in the supply chain, and the related indicator is the indicator corresponding to the business in the supply chain that is associated with the target business.

[0095] Optionally, the target business can be any one or more businesses in the supply chain business. The business can include planning-related businesses, such as replenishment plans, sales plans, and allocation plans. It can also include other types of businesses, such as product pool data, sales plans, product-warehouse relationships, planning parameters, and inventory data.

[0096] The related business can be other businesses in the supply chain that affect the target business. For example, if the replenishment plan is affected by the sales plan, then the sales plan can be considered a related business of the replenishment plan.

[0097] Optionally, business can be understood as business data; for example, the replenishment plan for output can be considered as replenishment plan data for output.

[0098] Step 402: Based on the degree of abnormality of the target indicator and each related indicator, determine the abnormal indicator, wherein the abnormal indicator is the indicator among the at least one related indicator that causes the target indicator to be abnormal.

[0099] Optionally, the degree of anomaly can be used to represent the difference between the actual value and the predicted value. The actual value of the indicator can be a real numerical value, and the predicted value can be obtained through a certain algorithm. For example, if 80 replenishment plans were actually produced today, then the actual value of the replenishment plan output can be 80; if, based on the replenishment plan output over a past period, it is predicted that 100 replenishment plans should be produced today, then the predicted value of the replenishment plan output is 100.

[0100] Understandably, the greater the difference between the actual value and the predicted value, the higher the degree of anomaly; the smaller the difference between the actual value and the predicted value, the lower the degree of anomaly.

[0101] In one example, the absolute value of the difference between the true value and the predicted value can be used as the degree of anomaly, or the absolute value of the ratio of the true value to the predicted value minus 1 can be used as the degree of anomaly, or there can be other ways to calculate the degree of anomaly, as long as they can reflect the difference between the true value and the predicted value.

[0102] In another example, the anomaly level of each indicator can be normalized. Optionally, the anomaly level can be set to a value between 0 and 100. A mapping table can be pre-set to store the correspondence between the absolute values ​​and the anomaly level values. In actual use, the anomaly level can be positioned between 0 and 100 based on the actual and predicted values, making it easy for users to view.

[0103] After determining the correlation between the various indicators, if the abnormality of the target indicator is greater than the first preset threshold, then the target indicator is determined to be abnormal.

[0104] After determining that the target indicator is abnormal, if one or more of the related indicators of the target indicator have an abnormality level greater than a second preset threshold, then the one or more related indicators are determined to be abnormal indicators. The first preset threshold and the second preset threshold may be the same or different.

[0105] Alternatively, the correlation indicators can be sorted according to their degree of abnormality, and one or more correlation indicators with the highest degree of abnormality can be selected as abnormal indicators.

[0106] Step 403: For the abnormal indicators, based on multiple preset dimensions, and according to the degree of abnormality corresponding to each dimension and the contribution of each element under each dimension, determine the abnormal dimensions and elements among the multiple dimensions.

[0107] Here, the degree of anomaly is used to represent the difference between the actual value and the predicted value of the corresponding indicator. For each dimension, the degree of anomaly represents the difference between the actual value and the predicted value of the corresponding indicator under that dimension. The contribution rate represents the proportion of the fluctuation change of an element in the fluctuation change of the abnormal indicator.

[0108] For example, when the anomaly indicator is the output of product pool data, for the region dimension, we can first calculate the actual value and predicted value of the output of product pool data in each region, and then further calculate the degree of anomaly and contribution of each region based on the actual value and predicted value. Based on the degree of anomaly in each region, we can obtain the degree of anomaly corresponding to the region dimension.

[0109] In practical applications, supply chain business data can guide daily workflows, such as stocking inventory in warehouses according to replenishment plans and preparing for sales based on sales targets. Furthermore, the methods provided in this embodiment can be used to monitor various indicators in the supply chain, promptly identifying the dimensions and elements causing the anomalies when they occur. For example, if today's replenishment plan shows an anomaly, this embodiment can quickly pinpoint which related indicator and which dimension or element is causing the anomaly.

[0110] In summary, the supply chain anomaly handling method provided in this embodiment can determine at least one related indicator corresponding to a target indicator in the supply chain. The target indicator is the indicator corresponding to a target business in the supply chain, and the related indicator is the indicator corresponding to a business associated with the target business in the supply chain. Based on the anomaly degree of the target indicator and each related indicator, anomaly indicators are determined. The anomaly indicator is the indicator among the at least one related indicator that causes the target indicator to be abnormal. For the anomaly indicator, based on multiple preset dimensions, according to the anomaly degree corresponding to each dimension and the contribution of each element under each dimension, the dimensions and elements among the multiple dimensions that cause the target indicator to be abnormal are determined. This allows for rapid and accurate diagnosis of data anomalies in the supply chain business, improving the timeliness and accuracy of supply chain business output, enhancing the executability of the supply chain business, facilitating the automation of supply chain business optimization, effectively improving the efficiency of supply chain diagnosis, meeting the normal usage needs of operational personnel, reducing problems such as product shortages and backlogs caused by supply chain business anomalies, maintaining the normal flow of goods in the supply chain, improving product processing efficiency, and reducing warehousing costs.

[0111] In one or more embodiments of this application, optionally, the target indicator and the associated indicator are indicators of the same type, wherein the type of indicator includes: output, output rate, and accuracy.

[0112] Optionally, the true value corresponding to the output quantity is the actual number of portions produced, and the predicted value is the number of portions that should be produced; the true value corresponding to the output rate is the ratio of the actual number of portions produced to the upper limit of the number of portions, and the predicted value is the ratio of the number of portions that should be produced to the upper limit of the number of portions; the true value and the predicted value corresponding to the accuracy rate are the actual business data and the predicted business data, respectively.

[0113] Optionally, the number of units that should be produced can be the average number of units actually produced within a historical period. When there are multiple products, corresponding supply chain business data can be generated for each SKU (Stock Keeping Unit). Products belonging to the same SKU can be considered as the same product. The actual number of units produced can be the number of SKUs that actually generated supply chain business, while the number of units that should be produced can be the average number of SKUs that actually generated supply chain business within a historical period.

[0114] For example, the actual value of the replenishment plan output can be the number of SKUs actually produced in the replenishment plan, while the predicted value can be the number of SKUs that should be produced in the replenishment plan, typically represented by the average of the past 30 days. For instance, if 70 SKUs were actually replenished today, the actual value is 70; if the average daily replenishment plan output over the past 30 days was 80 SKUs, the predicted value is 80.

[0115] Optionally, the output rate for supply chain operations can be the ratio of the number of output units to the upper limit of the number of units. The upper limit of the number of units can be set according to actual needs, for example, it can be the total number of SKUs in the warehouse on that day. Correspondingly, the number of output units can be the number of SKUs produced for supply chain operations.

[0116] In the output rate, the true value of the numerator can be the number of SKUs actually produced in the supply chain business, while the predicted value can be the number of SKUs that should be produced in the supply chain business, specifically the average number of SKUs actually produced in the supply chain business over the historical period. The true value of the denominator, i.e., the upper limit of the number of parts, can be the actual number of all SKUs in the warehouse, while the predicted value can be the average number of all SKUs in the warehouse over the historical period.

[0117] For example, for a replenishment plan, if there are 100 SKUs in the warehouse on a given day, the actual value of the denominator is 100. If the warehouse has an average of 120 SKUs per day over the past 30 days, the predicted value of the denominator is 120. If 80 SKUs in the warehouse generate replenishment plans on a given day, the actual value of the numerator is 80. If the warehouse has an average of 90 SKUs in the past 30 days generate replenishment plans, the predicted value of the numerator is 90.

[0118] Optionally, the actual value and predicted value corresponding to the accuracy of a certain supply chain business are the actual business data and the predicted business data, respectively. The predicted value can be calculated using a preset prediction algorithm, while the actual value can be the actual value that occurred.

[0119] For example, for a replenishment plan corresponding to a specific SKU, the business data refers to the replenishment quantity in the replenishment plan, that is, how many units need to be replenished for that SKU. When replenishment plans are generated for multiple SKUs, the accuracy of the replenishment plan can be calculated by combining the actual and predicted values ​​of multiple SKUs. For example, first calculate the accuracy of each SKU based on its actual and predicted values, and then weight them to obtain the final replenishment plan accuracy.

[0120] In this application embodiment, each supply chain business can have multiple types of indicators. For example, the actual supply chain plan indicators may include: replenishment plan output, replenishment plan output rate, replenishment plan accuracy rate, sales plan output, sales plan output rate, sales plan accuracy rate, allocation plan output, allocation plan output rate, allocation plan accuracy rate, etc.

[0121] After defining various supply chain plans and different types of indicators, you can choose at least one supply chain plan, use at least one type of indicator as the target indicator, and set corresponding related indicators.

[0122] In practical applications, target metrics and related metrics can be of the same type. For example, when calculating replenishment plan output rate, product pool data output rate and sales plan output rate can be used as related metrics, while accuracy can be temporarily disregarded.

[0123] In summary, by setting indicators such as output, output rate, and accuracy, the actual output quantity and accuracy of supply chain operations can be calculated, providing a more comprehensive measure of the health of the supply chain. Furthermore, since the target indicators and related indicators are of the same type, the causes of anomalies can be more accurately located, reducing interference from irrelevant factors and improving the accuracy of supply chain diagnosis.

[0124] In one or more embodiments of this application, optionally, determining at least one related indicator corresponding to the target indicator of the supply chain includes: determining a first business that is dependent on when producing the target business; determining related businesses based on the first business and a second business that is dependent on when producing the first business; and determining the at least one related indicator based on the related businesses.

[0125] Figure 5 This is a schematic diagram illustrating the dependencies of a replenishment plan provided in an embodiment of this application. For example... Figure 5As shown, when generating a replenishment plan, it relies on business data such as product pool data, sales plans, product-warehouse relationships, plan parameters, and inventory data. Therefore, when the target business is a replenishment plan, these business data can serve as the primary business data upon which the replenishment plan depends, and the primary business data can further depend on other business data. For example, the sales plan depends on business data such as warehouse allocation ratio, day-by-day ratio, historical sales, and algorithmic predictions. Therefore, warehouse allocation ratio, day-by-day ratio, historical sales, and algorithmic predictions can serve as the secondary business data corresponding to the replenishment plan.

[0126] Optional. Warehouse distribution ratio can refer to the proportion of historical sales in each warehouse. For example, if the front-end historical sales are 100, of which 10 are from warehouse A and 20 are from warehouse B, then the warehouse distribution ratios for warehouses A and B are 10% and 20%, respectively.

[0127] The day-to-day ratio refers to the percentage of historical sales within a given period allocated to each day. For example, if the front-end statistics show a week's historical sales of 100, and Monday and Tuesday's sales are 30 and 40 respectively, then Monday and Tuesday's day-to-day ratios could be 30% and 40% respectively.

[0128] An entity refers to a SKU corresponding to a product. For example, when calculating the historical sales of a brand of mobile phones, it includes the historical sales of all models and colors of that brand. Different models and colors can belong to different SKUs. Entity data is used to illustrate how many SKUs correspond to a certain historical sales volume. The entity ratio can refer to the historical sales volume corresponding to each SKU in the calculated historical sales volume.

[0129] The historical sales figures in the graph can refer to the aforementioned front-end historical sales. The algorithm prediction can be the algorithm used to predict sales, or the sales predicted by a certain algorithm, which is then calculated with factors such as warehouse allocation ratio and day allocation ratio to obtain the final sales plan. When calculating the sales plan, in addition to the front-end historical sales figures, information such as day allocation ratio and warehouse allocation ratio can be added to calculate the sales plan more accurately.

[0130] In practical applications, a sales plan can be calculated first based on historical sales figures, followed by a replenishment plan. Once the plan is in place, supply chain operations can be executed accordingly. For example, the replenishment plan determines how much merchandise needs to be purchased from suppliers today, thus replenishing the warehouse. When monitoring the replenishment plan, in addition to monitoring the sales plan itself, it's also possible to monitor historical sales figures and daily sales ratios upon which the sales plan depends.

[0131] For example, related metrics for replenishment plan output rate may include sales plan output rate, historical sales output rate, and daily output rate. Insufficient output of historical sales data and daily output data may lead to insufficient sales plan output, thereby affecting the output of the replenishment plan. Therefore, when analyzing the causes of anomalies in the replenishment plan, in addition to analyzing the sales plan on which the replenishment plan depends, it is also possible to further analyze the historical sales and daily output data on which the sales plan depends, and perform root cause drill-down.

[0132] In summary, by first identifying the businesses that produce the target business, and then, based on that business and other businesses that produce it, identifying related businesses, we can not only analyze the businesses that the target business depends on when it is abnormal, but also deeply pinpoint the actual cause of the abnormality, achieve root cause drilling, find the source of the abnormality for handling, realize fine-grained diagnosis of the supply chain, save investigation efficiency, and improve the speed and accuracy of abnormality handling.

[0133] In one or more embodiments of this application, optionally, in addition to performing dimension and element analysis on the correlation indicators, dimension and element analysis can also be performed on the target indicators.

[0134] Optionally, for the target indicator, based on multiple preset dimensions, the abnormal dimensions and elements can be determined according to the degree of abnormality corresponding to each dimension and the contribution of each element under each dimension.

[0135] For example, when the target indicator is the replenishment plan output rate, if the abnormal indicator is the product pool data output rate, then in addition to determining the dimensions and elements of the abnormal product pool data output rate, it is also possible to further determine the dimensions and elements of the abnormal replenishment plan output rate.

[0136] Optionally, at least one of the following information may also be displayed for each of the target indicators and related indicators: the degree of anomaly of the indicator, the degree of anomaly corresponding to each dimension, and the contribution of each element under each dimension.

[0137] For example, a pie chart can be displayed based on the degree of anomaly of each associated indicator; wherein the sector angle of each associated indicator in the pie chart is related to the degree of anomaly of the associated indicator; a distribution chart can be displayed based on the degree of anomaly of each dimension; wherein the horizontal axis of the distribution chart is the dimension, and the vertical axis is the degree of anomaly of the target indicator under each dimension; and a list of element contribution values ​​for each dimension can be displayed based on the degree of anomaly of each dimension and the contribution value of each element under each dimension.

[0138] In summary, by calculating the degree of anomaly for each dimension of the target indicator and the contribution of each element under each dimension, it is possible to identify the dimensions and elements of the target indicator that are abnormal among the multiple dimensions. This makes it easier for users to understand which dimensions and elements of the target indicator have problems. By displaying the results corresponding to each indicator, users can more intuitively understand the information of each indicator, dimension, and element, making it easier for users to monitor the target indicator and improve the user experience.

[0139] In one or more embodiments of this application, optionally, after determining the dimension and element of the anomaly corresponding to any indicator, a corresponding cause comparison tree can be found based on the indicator; the cause comparison tree includes multiple nodes forming a binary tree, wherein leaf nodes are used to represent candidate causes of the corresponding business anomaly under the dimension and element; branch nodes are used to represent judgment conditions set for the corresponding business; and the cause of the indicator anomaly is selected from the candidate causes based on the actual completion status of the corresponding business under the dimension and element and the cause comparison tree.

[0140] Figure 6 This is a schematic diagram of a cause-and-effect tree provided in an embodiment of this application. For example... Figure 6 As shown, an attribution platform can be used to list all possible causes and identify the root cause by analyzing the percentage of problems associated with each cause.

[0141] Taking an abnormal replenishment plan output rate in region A as an example, a cause-and-effect tree can be used to find the corresponding cause of the abnormal replenishment plan output rate. The branch nodes of the cause-and-effect tree are the judgment conditions, and there can be one or more branch nodes. The leaf nodes can be candidate causes for the abnormal replenishment plan output rate. After the replenishment plan output rate becomes abnormal, based on the actual completion status, the corresponding leaf nodes can be found through the judgment conditions of the branch nodes, thereby determining the cause of the abnormality.

[0142] The actual completion status can refer to the actual completion status corresponding to the judgment condition. For example... Figure 5 As shown, we can first determine whether there are any replenishment order records in region A in the past 60 days. If not, it indicates a historical omission problem: high dead inventory due to historical omission. If so, we can further determine whether the order quantity of the most recent replenishment order is less than the system's recommended quantity, and continue processing downwards based on the judgment result until the corresponding leaf node is located to obtain the cause of the anomaly.

[0143] Optionally, supply chain operations can be generated separately for each SKU. When analyzing the causes of anomalies in region A, each anomalous SKU can be analyzed separately. For example, according to the cause-and-effect tree, SKUs with no production replenishment plan or inaccurate production replenishment plan can be analyzed to determine the cause corresponding to that SKU. If more than a preset number of SKUs have the same cause, then that cause can be considered the ultimate cause of the supply chain operation anomaly.

[0144] In summary, after identifying an anomaly in a certain dimension and element, the actual completion status of the corresponding business under that dimension and element can be analyzed based on the cause comparison tree corresponding to the indicator. Since the rules and configuration parameters of the corresponding business under each dimension and element are different, the causes of anomalies under different dimensions and elements may also be different. The cause comparison tree can serve as a fine-grained diagnostic solution to further locate the cause of the anomaly under the abnormal dimension and element, realize root cause drilling down under the dimension and element, and further improve the accuracy of diagnosis.

[0145] After locating the cause using the cause-and-effect tree, optionally, the configuration parameters corresponding to the cause of the abnormal indicator can be updated. Figure 6 For example, if the final cause is that the automatic review limit is set too high, the automatic review limit can be adjusted. If the final cause is that there is a high level of dead inventory in the past, the high level of dead inventory can be found and processed.

[0146] In one or more embodiments of this application, optionally, if the abnormality level of the target indicator exceeds a preset threshold, a first alarm message is output. The first alarm message can be output to the user maintaining the target indicator, such as operations and maintenance personnel.

[0147] Figure 7 This is a system architecture diagram for handling supply chain anomalies, provided as an embodiment of this application. Figure 7 As shown, the business metric definition module allows for the definition of business metrics, which can include: metric name, such as replenishment plan output rate; configurable source data tables, indicating which data tables are used to obtain data for calculating the replenishment plan; calculation dimension, indicating the smallest granularity of the calculation, such as calculation at the SKU level, with each SKU corresponding to one replenishment plan; drill-down dimension, such as region, supplier, etc.; and metric algorithm, used to define the algorithm for the metric, such as the algorithm for replenishment plan output rate being the actual number of replenishment plans produced divided by the upper limit of the number of plans.

[0148] The indicator calculation module can include two modules: offline calculation and real-time calculation, used to implement offline and real-time calculations respectively. Daily summary data for indicators can include: indicator name, such as replenishment plan output rate; indicator value, such as today's replenishment plan output rate being 60%. Daily dimensional data for indicators can include: indicator name, dimension, element, and indicator value, used to record the indicator value corresponding to each dimension and element, for example, today's replenishment plan output rate for each country.

[0149] The metric monitoring module may include: an alarm judgment module, used to determine whether the metric exceeds the corresponding threshold; and an alarm sending module, used to issue an alarm when the corresponding threshold is exceeded. The alarm judgment module is related to the metric threshold rules, which may include: name, rule, threshold, etc.

[0150] Optionally, a user can be identified as responsible for maintaining the metrics of the anomaly's dimensions and elements, and a second alarm message can be output to that user so that the user can process the metrics corresponding to the anomaly's dimensions and elements. For example, different users can maintain different regions. If an anomaly is found in the sales plan output rate of region A, a corresponding second alarm message can be sent to the user responsible for maintaining region A, allowing the user maintaining region A to handle the situation in region A and resolve the anomaly in a timely manner.

[0151] Business metric definitions can also be correlated with metric correlation. Metric correlation can be used to determine the correlation between various metrics, such as the correlation between metric 1 and metric 2. For example, the replenishment plan output rate is correlated with the product pool data output rate and the sales plan output rate. A correlation can be set between the replenishment plan output rate and the product pool data output rate, and also between the replenishment plan output rate and the sales plan output rate. These two correlations can be the same or different. Based on the correlation, the most relevant influencing factors can be determined. For example, when the abnormality levels of the product pool data output rate and the sales plan output rate are equal, the one with a higher correlation can be considered more likely to be the cause of the abnormality in the replenishment plan output rate.

[0152] The attribution module may include: a contribution calculation module for calculating contribution; and a contribution page module for displaying the calculated contribution. The drill-down module may include: a dimension statistics module for calculating the degree of anomaly in each dimension; and a drill-down detail list module for displaying a detailed list of the degree of anomaly in each dimension and the contribution of each element.

[0153] In summary, by updating the configuration parameters corresponding to the causes of the abnormal indicators, the system can automatically handle the abnormalities, saving manpower and resources. Furthermore, when the abnormality of the target indicator exceeds a preset threshold, an alarm message can be output, facilitating user processing of the target indicator and enabling comprehensive handling of indicator abnormalities, thus assisting in the normal operation of the supply chain.

[0154] Figure 8 This is a flowchart illustrating a process for determining abnormal dimensions and elements, provided as an embodiment of this application. Figure 8 As shown, in one or more embodiments of this application, optionally, for the abnormal indicator, based on a preset set of multiple dimensions, and according to the degree of abnormality corresponding to each dimension and the contribution of each element under each dimension, determining the abnormal dimension and element among the multiple dimensions may include:

[0155] Step 801: For each dimension, calculate the degree of anomaly and contribution of each element based on the predicted and actual values ​​of the anomaly indicators corresponding to each element under the dimension, and add up the degree of anomaly of multiple elements under the dimension to obtain the degree of anomaly of the dimension.

[0156] It should be noted that, in addition to analyzing the indicators to be analyzed, this embodiment can also be used to analyze target indicators or other indicators.

[0157] For the indicators to be analyzed, the indicator values ​​for each element in each dimension can be calculated first. These indicator values ​​can include predicted and actual values, and the degree of anomaly can be calculated based on the predicted and actual values. Indicators can be divided into quantitative indicators and rate indicators. Output can be used as a quantitative indicator, while output rate and accuracy rate can be used as rate indicators. For quantitative indicators, the S-value and EP-value can be calculated using the following formulas:

[0158] p ij (m)=F ij (m) / F(m) (1)

[0159] q ij (m)=A ij (m) / A(m) (2)

[0160]

[0161] EP ij (m)=(A ij (m)-F ij (m)) / (A(m)-F(m)) (4)

[0162] Where the subscripts i and j represent the dimension and element respectively, m represents the indicator, F represents the predicted value, and A represents the actual value.ij (m) represents the predicted value of index m of the j-th element in the i-th dimension, A ij (m) represents the true value of index m of the j-th element in the i-th dimension, where p and q are intermediate results, and S ij (m) represents the S-value corresponding to the index m of the j-th element in the i-th dimension, EP ij (m) is the EP value corresponding to the index m of the j-th element in the i-th dimension.

[0163] Taking m as the planned sales output as an example, if there are three dimensions: region, warehouse type, and supplier; and the elements corresponding to the region include: region A, region B, ..., then S 12 (m) can be the S value corresponding to the replenishment plan output of region A.

[0164] Optionally, the actual value can be the actual planned sales output, while the predicted value can be determined using indicators from historical periods. That is, the current day's indicator can be predicted using past indicators, since indicators generally shouldn't change too much. For example, the average planned sales output over the past 30 days can be used as the predicted value for the planned sales output.

[0165] The S-value can be viewed as the degree of anomaly corresponding to an element, also known as the deviation, used to represent the unexpectedness of the element's indicator and to measure the difference between the actual and predicted values. The EP value can be viewed as the contribution of an element, specifically representing the proportion of a certain element's fluctuation in the fluctuation of anomaly indicators, and can measure the element's explanatory power for anomalies. After obtaining the S-values, for each dimension, the S-values ​​of multiple elements under that dimension can be added together to obtain the S-value for that dimension, i.e., the degree of anomaly in that dimension.

[0166] For rate-based indicators, such as output rate and accuracy rate, the S-value and EP-value can be calculated using the following formulas.

[0167] S = sum(Sf, Sg) (5)

[0168] EP ij =((A) ij (m1)-F ij (m1))*F(m2)-(A ij (m2)-F ij (m2))*F(m1)) / (F(m2)*(F(m2)+A ij (m2)-F ij (m2))) (6)

[0169] Among them, the rate value index can be broken down into the ratio of two indicators. For example, the output rate can be equal to the ratio of the number of output units to the upper limit of the number of units, and both the number of output units and the upper limit of the number of units can be dynamically changed.

[0170] For example, taking the sales plan output rate as an example, the numerator is the number of units produced, which can be equal to the number of SKUs produced according to the sales plan, and the denominator is the upper limit of the number of units, which can be the total number of SKUs. The predicted value of the number of units produced can be the average of the number of units produced over the past 30 days, and the upper limit of the number of units can be the average of the upper limit of the number of units produced over the past 30 days. The actual values ​​of the number of units produced and the upper limit of the number of units can be the number of units produced on the current day and the upper limit of the number of units on the current day, respectively.

[0171] In formula (5), Sf is the numerator index, Sg is the denominator index, and S is the sum of the numerator and denominator indices. In formula (6), m1 is the numerator index, and m2 is the denominator index.

[0172] Taking sales plan output rate as an example, A ij (m1) represents the number of units of the daily production and sales plan under the j-th element of the i-th dimension; A ij (m2) represents the maximum number of portions for the day under the j-th element of the i-th dimension; F ij (m1) is the average number of output units of the j-th element in the i-th dimension over the past 30 days; F ij F(m2) is the average of the upper limit of the number of units of the j-th element in the i-th dimension over the past 30 days; F(m1) is the average of the number of units of the sales plan produced in the past 30 days (the sum of all dimensions); F(m2) is the average of the upper limit of the number of units in the past 30 days (the sum of all dimensions).

[0173] Optionally, for each dimension, after obtaining the degree of anomalousness of each element under that dimension, the degree of anomalousness of all elements under that dimension can be summed to obtain the degree of anomalousness for that dimension. Alternatively, the degree of anomalousness of some important elements under that dimension can be summed to obtain the degree of anomalousness for that dimension. Considering that the number of dimensions and elements is large in practical applications, summing all elements may affect processing efficiency. Therefore, the following method can be used to determine the degree of anomalousness corresponding to each dimension.

[0174] Optionally, for a certain dimension, each element under that dimension is processed as follows: determine whether the EP value is greater than the individual threshold (e.g., 1%), and if so, add it to the suspicious list; when the sum of the EP values ​​in the suspicious list is greater than the overall threshold (e.g., 90%), it is considered that all abnormal elements under this dimension have been found.

[0175] Only elements with EP values ​​greater than a single-item threshold are added to the list, thus eliminating the need to consider elements with poor explanatory power or small percentage changes. If the sum of EP values ​​for a certain dimension equals 100%, processing of other elements within that dimension ceases once the sum exceeds the overall threshold. This is because once the sum exceeds the overall threshold, the list of suspected elements can be considered to largely explain or reflect the abnormal fluctuations in the indicator, and there's no need to consider remaining elements with low explanatory power.

[0176] After determining the suspicious list, the S values ​​of all elements in the suspicious list under that dimension are summed to obtain the S value of the dimension.

[0177] Step 802: Sort the degree of abnormality of each dimension and determine the dimension of abnormality based on the sorting results.

[0178] Optionally, multiple dimensions can be sorted from largest to smallest according to the S value, and the top N dimensions can be taken as the abnormal dimensions, where N is an integer greater than or equal to 1, which can be set according to actual needs.

[0179] Step 803: Determine the abnormal elements from the multiple elements of the dimension based on the contribution of each element in the abnormal dimension.

[0180] Optionally, after determining the dimensions of the anomaly, one or more elements with the highest contribution from those dimensions can be selected as the anomaly elements under that dimension. The resulting anomaly elements under the anomaly dimensions can be output as a root cause set. For example, this can be shown to the user that the main dimensions and elements causing the sales plan output rate anomaly are the sales plan output rate anomaly in region A.

[0181] In summary, by using the predicted and actual values ​​of the indicators corresponding to each element under each dimension, the EP value and S value corresponding to each element are calculated. Based on the EP value and S value, the set of abnormal elements under each dimension is located. Finally, the root cause set is output by summarizing the total S value of each dimension. This can decompose the multidimensional root cause analysis problem into multiple single-dimensional root cause analysis problems, quickly and accurately identify the dimensions and elements of anomalies, and improve the efficiency and accuracy of supply chain anomaly diagnosis.

[0182] Optionally, embodiments of this application also provide another supply chain anomaly handling method, which may include steps a to f below.

[0183] Step a: Define target metrics.

[0184] For example, target metrics could be: replenishment plan output rate, replenishment plan accuracy rate, sales plan output rate, sales plan accuracy rate, etc.

[0185] Step b: Set up the relevant metrics.

[0186] For example, if replenishment plans and sales plans are related as upstream and downstream, then the two can be linked.

[0187] Step c: Calculate all defined indicators through offline calculation, and calculate the degree of abnormality of each indicator.

[0188] For example, target indicators and their related indicators can be calculated based on various business data obtained from the database, such as replenishment plan output rate and sales plan output rate. The degree of anomaly for each indicator can be calculated according to the formulas in the aforementioned embodiments.

[0189] Optionally, for each indicator, the degree of abnormality corresponding to each element of each dimension can be calculated using the above formula, and the degree of abnormality can be added together to obtain the degree of abnormality corresponding to that indicator.

[0190] Alternatively, other algorithms can be used to calculate the degree of anomaly for each indicator, as long as they can represent the difference between the actual value and the predicted value of the indicator.

[0191] Step d: Calculate the indicator value for each dimension based on different dimensions (e.g., region, supplier, category, etc.), and use the S-value and EP-value calculation formulas to calculate the degree of anomaly and contribution respectively.

[0192] Specifically, for each indicator in the target indicator and related indicators, the degree of anomaly and contribution of the indicator under each dimension and element can be calculated.

[0193] Step e: Locate the root cause of the problem.

[0194] For example, if the target metric is the replenishment plan output rate, when the target metric is abnormal, we can check the degree of abnormality of various related metrics such as product pool data output rate and sales plan output rate to identify the abnormal metric. If the abnormal metric is the product pool data output rate, we can further check which dimension in the product pool data output rate has the highest S value. Assuming that the S value corresponding to the region dimension is the highest, we can find the element with the greatest contribution from the region dimension, such as country A. Then we can determine that the abnormal product pool data output rate of country A caused the abnormal replenishment plan output rate.

[0195] Step f: Monitor the indicators to identify problems in advance.

[0196] By monitoring target and related metrics, abnormal dimensions and elements can be identified in a timely manner, eliminating the need to wait for feedback from operations and maintenance personnel to locate problems. This improves the efficiency of supply chain monitoring and enables more comprehensive and accurate identification of anomalies.

[0197] The above methods can be used to process each type of supply chain business, thereby achieving automatic monitoring of target indicators and related indicators, and improving the stability of supply chain operations.

[0198] Corresponding to the above-mentioned supply chain anomaly handling method, this application embodiment also provides a supply chain anomaly handling device. Figure 9 This is a schematic diagram of a supply chain anomaly handling device provided in an embodiment of this application. Figure 9 As shown, the device includes:

[0199] The determining module 901 is used to determine at least one related indicator corresponding to the target indicator of the supply chain; wherein, the target indicator is the indicator corresponding to the target business in the supply chain, and the related indicator is the indicator corresponding to the business associated with the target business in the supply chain;

[0200] The diagnostic module 902 is used to determine abnormal indicators based on the degree of abnormality of the target indicator and each associated indicator, wherein the abnormal indicator is the indicator among the at least one associated indicator that causes the target indicator to be abnormal.

[0201] The processing module 903 is used to determine the abnormal dimensions and elements in the multiple dimensions based on the preset multiple dimensions, according to the degree of abnormality corresponding to each dimension and the contribution of each element under each dimension.

[0202] Among them, the degree of anomaly is used to represent the difference between the actual value and the predicted value of the corresponding indicator, and the contribution is used to represent the proportion of the fluctuation change of the element in the fluctuation change of the abnormal indicator.

[0203] In one or more embodiments of this application, optionally, the determining module 901 is specifically used for:

[0204] The first service on which the target service is determined;

[0205] Based on the first service and the second service on which the first service is produced, identify the associated services, and determine the at least one associated indicator based on the associated services.

[0206] In one or more embodiments of this application, optionally, the processing module 903 is further configured to:

[0207] For the target indicator, based on multiple preset dimensions, and according to the degree of anomaly corresponding to each dimension and the contribution of each element under each dimension, the abnormal dimensions and elements are determined; and / or,

[0208] Display at least one of the following information for each of the target indicators and related indicators: the degree of anomaly of the indicator, the degree of anomaly of each dimension, and the contribution of each element under each dimension.

[0209] In one or more embodiments of this application, optionally, the processing module 903 is further configured to:

[0210] After determining the dimension and element of the anomaly corresponding to any indicator, the corresponding cause comparison tree is found according to the indicator; the cause comparison tree includes multiple nodes forming a binary tree, wherein the leaf nodes are used to represent the candidate causes of the corresponding business anomaly under the dimension and element; and the branch nodes are used to represent the judgment conditions set for the corresponding business.

[0211] Based on the actual completion status of the corresponding business under the dimensions and elements, and the cause comparison tree, select the cause that leads to the abnormality of the indicator from the candidate causes.

[0212] In one or more embodiments of this application, optionally, the processing module 903 is further configured to perform at least one of the following:

[0213] Update the configuration parameters corresponding to the cause of the abnormality in the indicator;

[0214] If the abnormality of the target indicator exceeds a preset threshold, a first alarm message is output.

[0215] Identify the user responsible for maintaining the metrics of the dimensions and elements of the anomaly, and output a second alarm message to the user.

[0216] In one or more embodiments of this application, optionally, the processing module 903 is specifically used for:

[0217] For each dimension, based on the predicted and actual values ​​of the abnormal indicators corresponding to each element under the dimension, the degree of abnormality and contribution of each element are calculated, and the degree of abnormality of multiple elements under the dimension is added together to obtain the degree of abnormality of the dimension.

[0218] The degree of anomaly in each dimension is ranked, and the dimension of anomaly is determined based on the ranking result.

[0219] Based on the contribution of each element in the anomaly dimension, the anomalous element is determined from multiple elements in the dimension.

[0220] In one or more embodiments of this application, optionally, the target indicator and the associated indicator are indicators of the same type, wherein the type of indicator includes: output, output rate, and accuracy.

[0221] The actual value corresponding to the output is the number of portions actually produced, and the predicted value is the number of portions that should be produced.

[0222] The actual value corresponding to the output rate is the ratio of the actual number of portions produced to the upper limit of the number of portions, and the predicted value is the ratio of the number of portions that should be produced to the upper limit of the number of portions.

[0223] The accuracy rate corresponds to the actual business data and the predicted business data, respectively.

[0224] The supply chain anomaly handling device provided in this application embodiment can be used to perform the above-mentioned... Figures 1 to 8 The technical solutions of the embodiments shown are similar in principle and in effect, and will not be described again here.

[0225] Figure 10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 10 As shown, the electronic device in this embodiment may include:

[0226] At least one processor 1001; and

[0227] Memory 1002 communicatively connected to the at least one processor;

[0228] The memory 1002 stores instructions that can be executed by the at least one processor 1001, which, when executed by the at least one processor 1001, cause the electronic device to perform the method described in any of the above embodiments.

[0229] Alternatively, the memory 1002 can be either standalone or integrated with the processor 1001.

[0230] The implementation principle and technical effects of the electronic device provided in this embodiment can be found in the foregoing embodiments, and will not be repeated here.

[0231] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the method described in any of the foregoing embodiments.

[0232] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the methods described in any of the foregoing embodiments.

[0233] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed.

[0234] The integrated modules implemented as software functional modules described above can be stored in a computer-readable storage medium. These software functional modules, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods described in the various embodiments of this application.

[0235] It should be understood that the aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the application can be directly manifested as being executed by a hardware processor, or executed by a combination of hardware and software modules within the processor. The memory may include high-speed RAM, and may also include non-volatile memory (NVM), such as at least one disk storage device, and may also be a USB flash drive, external hard drive, read-only memory, disk, or optical disc, etc.

[0236] The aforementioned storage medium can be implemented from any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The storage medium can be any available medium accessible to general-purpose or special-purpose computers.

[0237] An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Alternatively, the storage medium can be an integral part of the processor. Both the processor and the storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and storage medium can exist as discrete components in an electronic device or host device.

[0238] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0239] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0240] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0241] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A supply chain anomaly handling method, characterized in that, include: Identify at least one related indicator corresponding to the target indicator of the supply chain; wherein the target indicator is the indicator corresponding to the target business in the supply chain, and the related indicator is the indicator corresponding to the business in the supply chain that is associated with the target business. Based on the degree of abnormality of the target indicator and each related indicator, an abnormal indicator is determined, wherein the abnormal indicator is the indicator among the at least one related indicator that causes the target indicator to be abnormal. For the aforementioned abnormal indicators, based on multiple preset dimensions, and according to the degree of abnormality corresponding to each dimension and the contribution of each element under each dimension, the abnormal dimensions and elements are determined among the multiple dimensions. Among them, the degree of anomaly is used to represent the difference between the actual value and the predicted value of the corresponding indicator, and the contribution is used to represent the proportion of the fluctuation of the element in the fluctuation of the abnormal indicator. Identify at least one related indicator corresponding to the target indicators of the supply chain, including: The first service on which the target service is determined; Based on the first service and the second service on which the first service is produced, determine the associated services, and determine the at least one associated indicator based on the associated services; Also includes: After determining the dimension and element of the anomaly corresponding to any indicator, the corresponding cause comparison tree is found according to the indicator; the cause comparison tree includes multiple nodes forming a binary tree, wherein the leaf nodes are used to represent the candidate causes of the corresponding business anomaly under the dimension and element; and the branch nodes are used to represent the judgment conditions set for the corresponding business. Based on the actual completion status of the corresponding business under the dimensions and elements and the cause comparison tree, select the cause of the abnormality of the indicator from the candidate causes; It also includes at least one of the following: Update the configuration parameters corresponding to the cause of the abnormality in the indicator; If the abnormality of the target indicator exceeds a preset threshold, a first alarm message is output. Identify the user responsible for maintaining the metrics of the dimensions and elements of the anomaly, and output a second alarm message to the user.

2. The method according to claim 1, characterized in that, Also includes: For the target indicator, based on multiple preset dimensions, and according to the degree of anomaly corresponding to each dimension and the contribution of each element under each dimension, the abnormal dimensions and elements are determined; and / or, Display at least one of the following information for each of the target indicators and related indicators: the degree of anomaly of the indicator, the degree of anomaly of each dimension, and the contribution of each element under each dimension.

3. The method according to claim 1, characterized in that, Regarding the aforementioned abnormal indicators, based on multiple preset dimensions, and according to the degree of abnormality corresponding to each dimension and the contribution of each element under each dimension, the abnormal dimensions and elements are determined, including: For each dimension, based on the predicted and actual values ​​of the abnormal indicators corresponding to each element under the dimension, the degree of abnormality and contribution of each element are calculated, and the degree of abnormality of multiple elements under the dimension is added together to obtain the degree of abnormality of the dimension. The degree of anomaly in each dimension is ranked, and the dimension of anomaly is determined based on the ranking result. Based on the contribution of each element in the anomaly dimension, the anomalous element is determined from multiple elements in the dimension.

4. The method according to any one of claims 1-3, characterized in that, The target indicator and the related indicator are indicators of the same type, wherein the types of indicators include: output, output rate, and accuracy. The actual value corresponding to the output is the number of portions actually produced, and the predicted value is the number of portions that should be produced. The actual value corresponding to the output rate is the ratio of the actual number of portions produced to the upper limit of the number of portions, and the predicted value is the ratio of the number of portions that should be produced to the upper limit of the number of portions. The accuracy rate corresponds to the actual business data and the predicted business data, respectively.

5. A supply chain anomaly handling device, characterized in that, include: The determination module is used to determine at least one related indicator corresponding to the target indicator of the supply chain; wherein the target indicator is the indicator corresponding to the target business in the supply chain, and the related indicator is the indicator corresponding to the business associated with the target business in the supply chain. The diagnostic module is used to determine abnormal indicators based on the degree of abnormality of the target indicator and each related indicator, wherein the abnormal indicator is the indicator among the at least one related indicator that causes the target indicator to be abnormal. The processing module is used to determine the abnormal dimensions and elements in the multiple dimensions based on the preset multiple dimensions, according to the degree of abnormality corresponding to each dimension and the contribution of each element under each dimension. Among them, the degree of anomaly is used to represent the difference between the actual value and the predicted value of the corresponding indicator, and the contribution is used to represent the proportion of the fluctuation of the element in the fluctuation of the abnormal indicator. The determining module is specifically used for: The first service on which the target service is determined; Based on the first service and the second service on which the first service is produced, determine the associated services, and determine the at least one associated indicator based on the associated services; The processing module is also used for: After determining the dimension and element of the anomaly corresponding to any indicator, the corresponding cause comparison tree is found according to the indicator; the cause comparison tree includes multiple nodes forming a binary tree, wherein the leaf nodes are used to represent the candidate causes of the corresponding business anomaly under the dimension and element; and the branch nodes are used to represent the judgment conditions set for the corresponding business. Based on the actual completion status of the corresponding business under the dimensions and elements and the cause comparison tree, select the cause of the abnormality of the indicator from the candidate causes; The processing module is also configured to perform at least one of the following: Update the configuration parameters corresponding to the cause of the abnormality in the indicator; If the abnormality of the target indicator exceeds a preset threshold, a first alarm message is output. Identify the user responsible for maintaining the metrics of the dimensions and elements of the anomaly, and output a second alarm message to the user.

6. An electronic device, characterized in that, include: At least one processor; as well as A memory that is communicatively connected to the at least one processor; The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, cause the electronic device to perform the method according to any one of claims 1-4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, implement the method as described in any one of claims 1-4.

8. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-4.

Images