Emergency material inventory measuring and calculating method based on gradient boosting decision tree algorithm
Through the emergency material inventory calculation method based on the gradient improvement decision tree algorithm, the emergency material inventory is dynamically monitored and optimized, and the problem of insufficient timeliness and agility of emergency material response is solved, and effective management and multi-dimensional analysis of emergency material inventory is realized.
Patent Information
- Application Number
- CN202411249913.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-06
- Publication Date
- 2024-12-06
AI Technical Summary
It is difficult for the existing technology to effectively utilize big data analysis technology, dynamically monitor and optimize emergency material warehouse storage, and adapt to different levels of emergencies and seasonal changes, resulting in insufficient timeliness and agility in emergency material response.
Using the emergency material inventory calculation method based on the gradient improvement decision tree algorithm, the emergency library storage is dynamically calculated through data grouping, timing analysis, correlation verification and machine learning model construction, seasonal characteristic factors are added, model performance is continuously monitored, and regular updates are updated. , to adapt to business changes.
Dynamic monitoring and optimization of emergency material inventory has been achieved, timely and agility of emergency material response has been improved, and units at all levels have assisted units in multi-dimensional analysis and effective management.
Smart Images

Figure CN119090085A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of material management, and in particular to an emergency material inventory calculation method based on a gradient boosting decision tree algorithm. Background Art
[0002] Emergency supplies refer to power grid repair equipment and materials, emergency repair tools, emergency relief materials and equipment, labor protection supplies, etc. that are needed to prevent emergencies that affect the company's production and operation (including natural disasters, accidents, public health, social security, etc.) and to restore power supply in a short time.
[0003] With the continuous deepening of the construction of the modern green digital supply chain system of State Grid, the Material Department of State Grid has deployed key tasks around the construction of a modern logistics system and the improvement of the quality and efficiency of supply chain management. In terms of improving the quality and efficiency of supply chain management, one of the key tasks is to strengthen inventory and turnover management, requiring "to work with equipment, finance and other professionals, apply big data analysis and other technical means to establish and improve turnover inventory and safety inventory quota calculation models, optimize daily inventory reserve strategies, promote faster inventory turnover, reduce material backlogs, and improve the efficiency and benefits of material utilization."
[0004] Therefore, how to use big data analysis technology, combined with different levels of emergencies, natural disasters and the number, type, regional characteristics, transportation cycle, season and other factors of emergency materials required over the years to conduct big data analysis, assist safety supervision, equipment, logistics and other professional departments to formulate emergency reserve material catalogs, formulate emergency material reserve inventory, realize the dynamic turnover mechanism of emergency reserve materials, and improve the timeliness and agility of emergency material response. We provide a method for calculating emergency material inventory based on the gradient boosting decision tree algorithm. Summary of the invention
[0005] The main purpose of the present invention is to overcome the deficiencies in the prior art and provide an emergency material inventory calculation method based on a gradient boosting decision tree algorithm. The emergency material inventory calculation method based on the gradient boosting decision tree algorithm embeds the model into the existing system, combines the existing management model, dynamically calculates the emergency inventory reserve, automatically updates the original threshold in the system, continuously optimizes the emergency reserve strategy, can add seasonal characteristic factors, dynamically predict emergency material reserves, sustainably monitor the prediction performance of the model, and regularly update the model to adapt to business changes. In addition, it can dynamically monitor the storage quantity of emergency materials and the dynamic fluctuation of thresholds of each unit in provinces, cities, and counties, and assist units at all levels in the effective management and multi-dimensional analysis of emergency materials.
[0006] The technical solution adopted by the present invention to achieve its technical purpose is: a method for calculating emergency material inventory based on a gradient boosting decision tree algorithm, which specifically includes the following steps:
[0007] S1. Group and aggregate the data set, group it by material and warehouse, accumulate the shipment quantity by month, and fill the missing months with "0" value;
[0008] S2. Perform time series analysis on single group data to understand data characteristics through insights into data appearance;
[0009] S3. Use the ADF (Augmented Dickey-Fuller) test on the data to check the stationarity of the data for this particular combination;
[0010] S4. Perform correlation analysis on the data;
[0011] S5. Filter available data attributes, including materials, warehouses, warehouse area, and delivery time;
[0012] S6, using the gradient boosting tree algorithm to split the data set;
[0013] S7. Use machine learning algorithms to learn and fit the internal data features of the training set and complete model construction;
[0014] S8. Use the test set to verify the model, and use industry-standard indicators to quantitatively evaluate the model effect, and compare the predicted values and true values in the test set;
[0015] S9. Calculate, count and manage the above steps by using a management system.
[0016] Preferably, in step S4, the data correlation analysis is to verify the autocorrelation of the time series data through the autocorrelation function (ACF) and the partial autocorrelation function (PACF), so as to verify the influence of the historical periodic data on the predicted data.
[0017] Preferably, in step S6, the data set is split by selecting the first 80% of the given data set as a training set and the last 20% as a test set.
[0018] Preferably, the management system comprises:
[0019] The first acquisition unit is used to acquire the entry and exit records of emergency materials in the emergency warehouse within the preset monitoring area;
[0020] A second acquisition unit is used to periodically or in real time acquire the inventory quantity of emergency supplies in the emergency supply reserve management system;
[0021] A statistical unit, used to count the actual number of emergency supplies in the emergency warehouse scanned by the user using a scanner;
[0022] The updating unit is used to update the inventory quantity of the emergency materials pre-entered into the emergency material reserve management system according to the storage record.
[0023] Preferably, the in-and-out records include out-of-warehouse records and in-warehouse records;
[0024] The updating unit is specifically used for:
[0025] The inventory quantity of emergency materials pre-entered into the emergency material reserve management system is increased according to the warehouse entry record; and the inventory quantity of emergency materials pre-entered into the emergency material reserve management system is reduced according to the warehouse exit record.
[0026] Preferably, the management system further comprises:
[0027] The first alarm unit is used to trigger an alarm message when it is determined that the inventory quantity is less than a preset inventory threshold.
[0028] The second alarm unit is used to trigger an alarm message when it is determined that the actual quantity is different from the current inventory quantity of the emergency materials in the emergency material reserve management system.
[0029] Preferably, the management system further comprises:
[0030] processor;
[0031] a memory for storing instructions executable by the processor;
[0032] Among them, the processor is configured to execute the instructions to implement the emergency material inventory estimation method based on the gradient boosting decision tree algorithm.
[0033] Preferably, the management system further comprises:
[0034] Storage media;
[0035] When the instructions in the storage medium are executed by a processor of the emergency material management system, the emergency material management system can execute the emergency material inventory estimation method based on the gradient boosting decision tree algorithm.
[0036] The working principle and specific usage process of the emergency material inventory estimation method based on the gradient boosting decision tree algorithm:
[0037] First, group and aggregate the data set, group the material and warehouse into a single group, accumulate the number of shipments by month, and fill the missing months with a "0" value. Secondly, perform time series analysis on the single group data, and understand the data characteristics by looking at the data appearance. Then, use the ADF test to check the stationarity of the data for this specific combination, and analyze the test results; perform correlation analysis on the data, and verify the autocorrelation of the time series data through the autocorrelation function and partial autocorrelation function to verify the impact of historical cycle data on the forecast data. Finally, filter the available data attributes, including materials, warehouses, warehouse area, and shipment time.
[0038] Compared with the prior art, the present invention has the following beneficial effects:
[0039] This emergency material inventory calculation method based on the gradient boosting decision tree algorithm embeds the model into the existing system, combines the existing management model, dynamically calculates the emergency inventory reserves, automatically updates the original thresholds in the system, and continuously optimizes the emergency reserve strategy.
[0040] This emergency material inventory estimation method based on the gradient boosting decision tree algorithm can add seasonal characteristic factors during the model application process to dynamically predict emergency material reserves.
[0041] This emergency material inventory estimation method based on the gradient boosting decision tree algorithm continuously monitors the prediction performance of the model during the model application process and regularly updates the model to adapt to business changes.
[0042] This emergency material inventory calculation method based on the gradient boosting decision tree algorithm dynamically monitors the storage quantity and threshold dynamic fluctuations of each unit's emergency materials on a province, city, and county basis, assisting units at all levels in the effective management and multi-dimensional analysis of emergency materials. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] Figure 1 This is a flowchart of the steps of the emergency material inventory estimation method based on the gradient boosting decision tree algorithm.
[0044] Figure 2 This is the system structure diagram of the management system.
[0045] Figure 3 This is a monthly cumulative shipment quantity chart after filling of material 500020749 (wire pulling fittings) and EBAA (Tiaoshan Street Warehouse, Yuncheng).
[0046] Figure 4 This is a monthly cumulative shipment quantity chart after filling for material 500014664 (overhead insulated wire) and EBAA (Tiaoshan Street Warehouse, Yuncheng).
[0047] Figure 5This is the monthly cumulative shipment quantity chart after filling of materials 500134497 (complete set of substation equipment) and EBAA (Tiaoshan Street Warehouse, Yuncheng).
[0048] Figure 6 These are the ACF and PACF diagrams for material 500020749 (wire pulling fittings) and EBAA (Tiaoshan Street Warehouse, Yuncheng).
[0049] Figure 7 These are the ACF and PACF diagrams for material 500014664 (overhead insulated wire) and EBAA (Yuncheng Tiaoshan Street warehouse).
[0050] Figure 8 These are the ACF and PACF diagrams for material 500134497 (complete set of substation equipment) and EBAA (Tiaoshan Street warehouse in Yuncheng).
[0051] Fig. 9 This is the actual monthly shipment quantity and forecasted monthly shipment quantity chart for material 500020749.
[0052] Fig.10 This is the actual monthly shipment quantity and forecasted monthly shipment quantity chart for material 500014664.
[0053] Fig.11 This is the actual monthly shipment quantity and forecasted monthly shipment quantity chart for material 500134497. DETAILED DESCRIPTION
[0054] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, or the positions or positional relationships in which the inventive product is usually placed when in use. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.
[0055] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0056] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention is further described in detail below through the accompanying drawings and embodiments. However, it should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the scope of the present invention. In addition, in the following description, the description of known structures and technologies is omitted to avoid unnecessary confusion of the concept of the present invention.
[0057] Embodiment 1:
[0058] See also Figure 1-11 , a method for calculating emergency material inventory based on gradient boosting decision tree algorithm, specifically including the following steps:
[0059] S1. Group and aggregate the data set, group it by material and warehouse, accumulate the shipment quantity by month, and fill the missing months with "0".
[0060] S2. Perform time series analysis on single group data and understand data characteristics through insights into the data appearance.
[0061] (1) The following three groups with large amounts of data after group aggregation are selected as examples, as shown in the following table:
[0062]
[0063] (2) Analyze the data after filling, such as Figure 3-5 As shown;
[0064] Figure 3 It indicates the monthly cumulative shipment quantity of material 500020749 (wire fittings) and EBAA (Tiaoshan Street Warehouse, Yuncheng) after filling.
[0065] Figure 4 It represents the monthly cumulative shipment quantity after filling of material 500014664 (overhead insulated wire) and EBAA (Tiaoshan Street Warehouse, Yuncheng).
[0066] Figure 5 It indicates the monthly cumulative shipment quantity of material 500134497 (complete set of substation equipment) and EBAA (Tiaoshan Street Warehouse, Yuncheng) after filling.
[0067] S3. Use the ADF (Augmented Dickey-Fuller) test on the data to check the stationarity of the data for this particular combination.
[0068] Test result analysis:
[0069]
[0070] S4. Perform correlation analysis on the data. Correlation analysis of data verifies the autocorrelation of time series data through the autocorrelation function (ACF) and partial autocorrelation function (PACF), which is used to verify the impact of historical periodic data on forecast data. Figure 6-8 As shown:
[0071] Figure 6 Represents the ACF and PACF of material 500020749 (wire fittings) and EBAA (Tiaoshan Street warehouse, Yuncheng).
[0072] Figure 7 Represents the ACF and PACF of material 500014664 (overhead insulated wire) and EBAA (Yuncheng Tiaoshan Street Warehouse).
[0073] Figure 8 Represents the ACF and PACF of material 500134497 (complete set of substation equipment) and EBAA (Tiaoshan Street warehouse in Yuncheng).
[0074] The following conclusions can be drawn from the chart: the values in the three groups of time series have a certain correlation with their previous lagged values, indicating that the time series have a certain degree of autoregression.
[0075] S5. Filter available data attributes, including materials, warehouses, warehouse area, and delivery time.
[0076] (1) The following dimensional data are constructed through feature engineering. Considering the possible influence of seasonal factors, more time dimensions are added, as shown in the following table:
[0077] name Property name Remark materials marerial storehouse warehouse Warehouse area area Year of shipment year Shipping month month Monthly outbound quantity quantity Quantity shipped out before 1-12 months lag_1-lag_12 A total of 12 dimensions Maximum outbound quantity within 2-12 months rollingmax_2-rollingmax_12 A total of 11 dimensions Minimum outbound quantity within 2-12 months rollingmin_2-rollingmin_12 A total of 11 dimensions Average outbound quantity within 2-12 months rollingmean_2-rollingmean_12 A total of 11 dimensions
[0078] Through the above data engineering, the three sets of data obtained 846, 691 and 299 valid data respectively, covering 51 data dimensions.
[0079] (2) For commonly used time series prediction algorithms (ARIMA, LSTM, etc.), the amount of data is relatively small, which will have a significant impact on the prediction accuracy. Therefore, the gradient boosting tree algorithm (XGBoost) is considered.
[0080] S6. Use the gradient boosting tree algorithm to split the data set; the data set is split by selecting the first 80% as the training set and the last 20% as the test set in a given data set.
[0081] S7. Use machine learning algorithms to learn and fit the internal data features of the training set and complete model construction.
[0082] S8. Use the test set to verify the model, and use industry-standard indicators to quantitatively evaluate the model effect, and compare the predicted values and true values in the test set.
[0083] The red dotted line in the figure is the predicted value, and the blue solid line is the true value, which can intuitively show the accuracy of the prediction. Figure 9-11 As shown:
[0084] Fig. 9 Indicates the actual monthly shipment quantity and forecasted monthly shipment quantity of material 500020749.
[0085] Fig.10 Indicates the actual monthly shipment quantity and forecasted monthly shipment quantity of material 500014664.
[0086] Fig.11 Indicates the actual monthly shipment quantity and forecasted monthly shipment quantity of material 500134497.
[0087] The test results of the three sets of data are analyzed as follows:
[0088]
[0089] The average absolute errors of the three groups of data were all less than 17%, indicating that the model results were well fitted with the true values; the R2 was all above 0.75, indicating that the model had good explanatory power (>0.6).
[0090] To sum up the above results, the gradient boosting tree model can effectively predict the emergency material reserve quota of each warehouse in the next cycle, and provide auxiliary support for the emergency material reserve. By combining the gradient boosting tree model with the traditional emergency management model, a deeper prediction of the power emergency material inventory reserve can be made, and finally the power emergency material inventory reserve can be built into a comprehensive system integrating "monitoring-early warning-procurement", which can improve the timeliness and agility of emergency material response.
[0091] Embodiment 2:
[0092] See also Figure 2 Based on the above embodiment, the method for calculating the emergency material inventory based on the gradient boosting decision tree algorithm calculates, counts and manages the above steps by using a management system. The management system includes:
[0093] The first acquisition unit is used to acquire the in-and-out records of emergency materials in the emergency warehouse within a preset monitoring area; the in-and-out records include out-of-warehouse records and in-warehouse records.
[0094] The second acquisition unit is used to periodically or in real time acquire the inventory quantity of emergency supplies in the emergency supply reserve management system.
[0095] The statistical unit is used to count the actual quantity of emergency supplies in the emergency warehouse scanned by the user using a scanner.
[0096] An updating unit, used to update the inventory quantity of emergency materials pre-entered into the emergency material reserve management system according to the storage record;
[0097] The updating unit is specifically used to: increase the inventory quantity of emergency materials pre-entered into the emergency material reserve management system according to the storage record; and reduce the inventory quantity of emergency materials pre-entered into the emergency material reserve management system according to the storage record.
[0098] Furthermore, in the embodiment of the present invention, the management system further includes:
[0099] The first alarm unit is used to trigger an alarm message when it is determined that the inventory quantity is less than a preset inventory threshold.
[0100] The second alarm unit is used to trigger an alarm message when it is determined that the actual quantity is different from the current inventory quantity of the emergency materials in the emergency material reserve management system.
[0101] Furthermore, in the embodiment of the present invention, the management system further includes:
[0102] a processor, a memory for storing instructions executable by the processor;
[0103] Among them, the processor is configured to execute the instructions to implement the emergency material inventory estimation method based on the gradient boosting decision tree algorithm as described.
[0104] Furthermore, in the embodiment of the present invention, the management system further includes:
[0105] Storage medium, when the instructions in the storage medium are executed by the processor of the emergency material management system, the emergency material management system can execute the emergency material inventory estimation method based on the gradient boosting decision tree algorithm.
[0106] The storage medium includes a memory of instructions, and the above instructions can be executed by a processor of the management system to complete the above method. The storage medium can be a non-temporary computer-readable storage medium. For example, the non-temporary computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a tape, a floppy disk, and an optical data storage device, etc.
[0107] Specifically, when in use, the first acquisition unit acquires the in-and-out records of emergency materials in the emergency warehouse in the preset monitoring area, wherein the in-and-out records include out-and-out records and in-and-out records. The second acquisition unit acquires the inventory quantity of emergency materials in the emergency material reserve management system periodically or in real time.
[0108] The actual number of emergency supplies in the emergency warehouse scanned by the user using a scanner is counted by the statistical unit. The inventory quantity of emergency supplies pre-entered into the emergency supplies reserve management system is updated by the updating unit according to the warehouse entry and exit records; the updating unit is specifically used to: increase the inventory quantity of emergency supplies pre-entered into the emergency supplies reserve management system according to the warehouse entry records, and reduce the inventory quantity of emergency supplies pre-entered into the emergency supplies reserve management system according to the warehouse exit records.
[0109] When the first alarm unit determines that the inventory quantity is less than the preset inventory threshold, an alarm message is triggered. When the second alarm unit determines that the actual quantity is different from the current inventory quantity of emergency materials in the emergency material reserve management system, an alarm message is triggered.
[0110] Among them, the processor is used to store a memory of the processor executable instructions, and the processor is configured to execute the instructions to implement the emergency material inventory estimation method based on the gradient boosting decision tree algorithm as described.
[0111] Furthermore, when the instructions in the storage medium are executed by a processor of the emergency material management system, the emergency material management system is enabled to execute the emergency material inventory estimation method based on the gradient boosting decision tree algorithm.
[0112] The solution in this embodiment can be selectively combined with the solutions in other embodiments for use.
[0113] It should be noted that although the above embodiments have been described in this article, the scope of patent protection of the present invention is not limited thereby. Therefore, based on the innovative concept of the present invention, changes and modifications made to the embodiments described herein, or equivalent structures, equivalent processes or equivalent functional transformations made using the contents of the present specification and drawings, directly or indirectly applying the above technical solutions to other related technical fields are all included in the scope of protection of the patent of the present invention.
Claims
1. A method for calculating emergency material inventory based on a gradient boosting decision tree algorithm, characterized in that: The specific steps are as follows: S1. Group and aggregate the data set, group it by material and warehouse, accumulate the shipment quantity by month, and fill the missing months with "0" value; S2. Perform time series analysis on single group data to understand data characteristics through insights into data appearance; S3. Use the ADF test on the data to check the stationarity of the data for this particular combination; S4. Perform correlation analysis on the data; S5. Filter available data attributes, including materials, warehouses, warehouse area, and delivery time; S6, using the gradient boosting tree algorithm to split the data set; S7. Use machine learning algorithms to learn and fit the internal data features of the training set and complete model construction; S8. Use the test set to verify the model, and use industry-standard indicators to quantitatively evaluate the model effect, and compare the predicted values and true values in the test set; S9. Calculate, count and manage the above steps by using a management system.
2. The method for calculating emergency material inventory based on a gradient boosting decision tree algorithm according to claim 1 is characterized in that: In step S4, the data correlation analysis is to verify the autocorrelation of the time series data through the autocorrelation function and the partial autocorrelation function, so as to verify the influence of the historical periodic data on the predicted data.
3. The method for calculating emergency material inventory based on a gradient boosting decision tree algorithm according to claim 1 is characterized in that: In step S6, the data set is split by selecting the first 80% of the given data set as a training set and the last 20% as a test set.
4. A method for calculating emergency material inventory based on a gradient boosting decision tree algorithm according to any one of claims 1 to 3, characterized in that: The management system comprises: The first acquisition unit is used to acquire the entry and exit records of emergency materials in the emergency warehouse within the preset monitoring area; A second acquisition unit is used to periodically or in real time acquire the inventory quantity of emergency supplies in the emergency supply reserve management system; A statistical unit, used to count the actual number of emergency supplies in the emergency warehouse scanned by the user using a scanner; The updating unit is used to update the inventory quantity of the emergency materials pre-entered into the emergency material reserve management system according to the storage record.
5. The method for calculating emergency material inventory based on a gradient boosting decision tree algorithm according to claim 4 is characterized in that: The in-and-out records in the management system include out-of-stock records and in-stock records; The updating unit is specifically used for: The inventory quantity of emergency materials pre-entered into the emergency material reserve management system is increased according to the warehouse entry record; and the inventory quantity of emergency materials pre-entered into the emergency material reserve management system is reduced according to the warehouse exit record.
6. The method for calculating emergency material inventory based on a gradient boosting decision tree algorithm according to claim 5 is characterized in that: The management system also includes: The first alarm unit is used to trigger an alarm message when it is determined that the inventory quantity is less than a preset inventory threshold. The second alarm unit is used to trigger an alarm message when it is determined that the actual quantity is different from the current inventory quantity of the emergency materials in the emergency material reserve management system.
7. The method for calculating emergency material inventory based on a gradient boosting decision tree algorithm according to claim 6 is characterized in that: The management system also includes: processor; a memory for storing instructions executable by the processor; Wherein, the processor is configured to execute the instructions to implement the emergency material inventory estimation method based on the gradient boosting decision tree algorithm as described in any one of claims 1-3.
8. The method for calculating emergency material inventory based on a gradient boosting decision tree algorithm according to claim 7 is characterized in that: The management system also includes: Storage media; When the instructions in the storage medium are executed by a processor of the emergency material management system, the emergency material management system is enabled to execute the emergency material inventory estimation method based on the gradient boosting decision tree algorithm as described in any one of claims 1 to 3.
