Range hood spare part multi-level inventory intelligent distribution system and method

Abstract

The invention discloses a range hood spare part multistage inventory intelligent distribution system and method, and belongs to the technical field of range hood spare part inventory management and control, and the system comprises range hood equipment, a plurality of self-service spare part shelves, an integrated spare part warehouse, a spare part management system access terminal, a spare part management system server, and a supplier spare part warehouse. With the spare part management system server as a connection center, the range hood equipment, the self-service spare part goods shelf, the spare part management system access terminal and the supplier spare part warehouse are connected together through a computer communication network. According to the invention, interconnection and intercommunication of internal and external digital systems of an enterprise are realized, data of multi-stage spare part inventory positions are integrated, the spare part demand of the next period of the current time is predicted based on the optimized random forest predictor, and high efficiency, intelligence and refinement of spare part management are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of inventory control of range hood spare parts, and more specifically, to a multi-level inventory intelligent distribution system and method for range hood spare parts. Background Art

[0002] Cigarette machinery equipment refers to the general term for various types of mechanical equipment specifically used for tobacco leaf processing and production, including tobacco-specific equipment, cigarette equipment, tobacco drying equipment, deformation processing equipment, packaging equipment, etc., among which cigarette machinery equipment in particular has a higher level of automation, digitization, and integration. Cigarette machinery equipment is of many categories, with a wide range of manufacturers, complex models, and a wide variety of parts and components with different specifications. In a cigarette workshop, there are often multiple types of cigarette equipment at the same time. On a highly automated assembly line, the normal high-speed operation of cigarette machinery equipment is an important condition to ensure the entire cigarette production.

[0003] Cigarette machine repair and maintenance cannot be separated from spare parts. In order to ensure cigarette production, there are high requirements for the downtime rate and time of cigarette machine, so the inventory decision of cigarette machine spare parts is particularly important. On the one hand, in order to ensure cigarette production, the inventory of cigarette machine spare parts should maintain a high inventory type, inventory rate, and inventory quantity; on the other hand, the refined management of cigarette machine spare parts requires reducing the inventory volume, reducing the inventory occupancy, and improving the spare parts turnover rate. Most of the existing cigarette machine spare parts management is based on the primary goal of ensuring production, and the use of spare parts is mostly based on demand according to the use process. That is, after the front-line mechanics put forward an application for the use of a spare part, the spare parts manager often aims to meet the needs of the front-line mechanics, and conduct spare parts inventory management and spare parts ordering and replenishment operations. To this end, tobacco companies usually adopt optimized management methods to improve the performance and level of cigarette machine spare parts management. For example, common management methods include: establishing a spare parts ledger system, tracking the use, replacement, and scrapping of spare parts to reduce losses in this process.

[0004] With the continuous improvement of the digitalization level of tobacco enterprises, the popularization of tobacco machine spare parts information management system has made the digital footprint of tobacco machine spare parts more accurate and real-time. With the application of big data and artificial intelligence technology, the demand trend of spare parts can be predicted more accurately, so that spare parts inventory managers can make more accurate and efficient inventory management decisions. There are already some commercial inventory management intelligent decision systems, which provide open source time convolutional networks, long short-term memory networks, time series prediction algorithms, etc., which cluster and extract features from the aggregated product historical sales, product attributes and other data, and combine the key technologies such as inventory strategy algorithms, optimal cost calculation models, and inventory allocation models in operations research algorithms to achieve higher prediction accuracy than traditional inventory demand prediction algorithm solutions. The "Site Selection and Inventory Method and System for Maintenance Equipment under Multi-level Supply Chain and Inventory Warehouse" with announcement number CN112288138B discloses the use of hybrid fruit fly optimization algorithms to construct warehouse site selection, inventory assignment, and inventory control decision algorithms for maintenance equipment under multi-level supply chains to reduce site selection costs, ordering costs, and storage costs.

[0005] In current practical work, facing a large number of spare parts types and demand scenarios, as well as the cumbersome process of obtaining, replacing and scrapping, it is an urgent problem to be solved how to improve the efficiency of multi-level intelligent allocation of range hood spare parts inventory, quickly respond to maintenance needs and reduce inventory backlogs. Summary of the invention

[0006] An object of the present invention is to provide a multi-level inventory intelligent distribution system and method for range hood spare parts to solve the aforementioned problems of low efficiency of intelligent inventory distribution, slow response to maintenance needs, serious inventory backlog, and low accuracy of spare parts demand prediction.

[0007] According to a first aspect of the present invention, a multi-level inventory intelligent distribution system for range hood spare parts is provided, comprising a range hood equipment, a plurality of self-service spare parts shelves, an integrated spare parts warehouse, a spare parts management system access terminal, a spare parts management system server and a supplier spare parts warehouse; wherein the self-service spare parts shelves are deployed near the range hood equipment, and the integrated spare parts warehouse is used for unified management of spare parts ordering, replenishment, and distribution of spare parts to a plurality of self-service spare parts shelves; spare parts management personnel access the spare parts management system through the spare parts management system access terminal; the spare parts management system is installed on the spare parts management system server; the supplier spare parts warehouse stores some spare parts for distribution of spare parts to the integrated spare parts warehouse within the company during replenishment and ordering; with the spare parts management system server as the connection center, the range hood equipment, the self-service spare parts shelves, the spare parts management system access terminal and the supplier spare parts warehouse are connected together through a computer communication network.

[0008] Optionally, the spare parts in the supplier spare parts warehouse are transported to the company's internal integrated spare parts warehouse, and then flow to the self-service spare parts shelves, forming a multi-level inventory structure of range hood spare parts, specifically including:

[0009] A primary spare parts inventory consisting of self-service spare parts shelves;

[0010] The secondary spare parts inventory composed of the integrated spare parts warehouse. When the spare parts stored in the self-service spare parts shelf are lower than the safety stock threshold, the mechanic directly initiates a request for use to the integrated spare parts warehouse. The integrated spare parts warehouse, as the secondary spare parts inventory, allocates the corresponding spare parts to the primary spare parts inventory or the mechanic.

[0011] The three-level spare parts inventory consists of the supplier spare parts inventory.

[0012] Optionally, data collection is performed through a multi-level inventory structure of range hood spare parts, including collecting data from an internal digital system of the enterprise and an external digital system of the enterprise; wherein the data collected by the internal digital system of the enterprise include at least one of basic element data of spare parts requisition, basic element data of spare parts application, and basic element data of spare parts allocation; the external digital system of the enterprise is composed of spare parts management systems of various suppliers, and the data collected by the external digital system of the enterprise include basic element data of spare parts allocation.

[0013] According to a second aspect of the present invention, a multi-level inventory intelligent allocation method for range hood spare parts is provided, which is deployed in the above-mentioned spare parts management system, and the method comprises the following steps:

[0014] Step 1: Select a specific time interval before the current time as the data collection time interval;

[0015] Step 2: obtaining a spare parts demand time series data set within the selected data collection time interval, the spare parts demand time series data set including: a spare parts requisition time series data set including spare parts requisition basic element data, a spare parts application time series data set including spare parts application basic element data, a spare parts allocation time series data set including spare parts allocation basic element data, and a spare parts transfer time series data set including spare parts transfer basic element data;

[0016] Step 3: Select a part of the spare parts demand time series data set as the training data set;

[0017] Step 4: Input the training data set into the random forest model to generate a random forest predictor;

[0018] Step 5: Select another part of the spare parts demand time series data set as the test data set;

[0019] Step 6: Adjust the parameters of the random forest model so that the fitting error of the random forest model for the test data set is less than the error threshold, and obtain an optimized random forest predictor;

[0020] Step 7: Based on the optimized random forest predictor, the spare parts demand in the period after the current time is predicted to allocate the types and quantities of spare parts inventory at all levels.

[0021] Optionally, after step 4, the method further includes: uniformly classifying the range hood spare parts according to specific parameters to make different decisions, wherein the classification algorithm process includes:

[0022] According to the frequency of spare parts consumption, the average demand occurrence interval analysis is carried out, and the spare parts are divided into three types: F (fast flow), S (slow flow), and N (non-flow);

[0023] Cluster analysis algorithm is used to further classify the S-class (slow-flow) spare parts into two categories: V-class (important) and D-class (unimportant);

[0024] Cluster analysis algorithm is used to further divide Class F (fast-flowing) spare parts into three categories: Class V (important), Class D (unimportant), and Class E (generally important).

[0025] Optionally, the multi-level inventory intelligent allocation method for smoke machine spare parts also includes: using a spare parts value analysis method to further divide the V-type (important) spare parts of the F-type (fast-flowing) into three categories: A, B, and C;

[0026] The spare parts value analysis method is used to further divide the F (fast-moving) and E (generally important) spare parts into three categories: A, B, and C.

[0027] Optionally, in the process of classifying spare parts into three types, namely, Class F (fast flow), Class S (slow flow), and Class N (non-flow), they are sorted according to the average interval of spare parts demand, and then the average annual consumption amount of spare parts and the annual cumulative consumption amount and proportion of spare parts are calculated. When the cumulative consumption amount reaches 70% of the total spare parts consumption amount, these spare parts are classified into Class F (fast flow); spare parts with a cumulative consumption amount in the range of 70%-90% are classified into Class S (slow flow), and the remaining spare parts are classified into Class N (non-flow).

[0028] Optionally, in step 4, the construction process of the random forest model includes the following steps:

[0029] Step 41: obtaining the data set of the first 80% of the spare parts demand time interval, the data set including the spare parts collection time series data set, the spare parts application time series data set, the spare parts allocation time series data set and the spare parts transfer time series data set;

[0030] Step 42: Using a random selection and replacement sampling technique, randomly select sample data from the first 80% data set of the spare parts demand time interval;

[0031] Step 43: Input the randomly selected sample data into the weak predictor, which uses a classification and regression decision tree (CART);

[0032] Step 44: make the first decision: until the weak predictor outputs an evaluation criterion less than the mean square error threshold; repeat steps 42 to 44: randomly select the nth sample set → input the nth weak predictor → make the nth decision, where n is the number of decision trees in the random forest model;

[0033] Step 45: Construct a basic random forest model, wherein the selected characteristic variables include: historical demand data, historical demand satisfaction rate, historical demand response time, inventory location, and spare parts type, wherein the historical demand data includes historical application data and collection data of the first-level spare parts inventory, historical allocation data of the second-level spare parts inventory, and historical transfer data of the third-level spare parts inventory.

[0034] Optionally, in step 6, the mean absolute percentage error (MAPE) is used as a judgment indicator of the fitting error, and the mean absolute percentage error is defined as the distance between the model predicted value F(t) and the true value A(t) divided by the average value of the true value A(t), that is:

[0035] ;

[0036] Among them, N is the total amount of test data set data, is the subscript of each predicted value-true value pair.

[0037] Optionally, in step 6, the process of adjusting the parameters of the random forest model includes the following steps:

[0038] Step 61: Start a loop to traverse the number of decision trees in the random forest model;

[0039] Step 62: Start a loop to traverse the maximum depth of the decision tree;

[0040] Step 63: Start a loop to traverse the minimum number of samples of leaf nodes;

[0041] Step 64: Execute in the loop: read the 20% data set after the time interval, and calculate the MAPE evaluation index one by one with the predicted value;

[0042] Step 65: Select the set of model parameters with the smallest MAPE as the optimal random forest model.

[0043] The multi-level inventory intelligent distribution system and method for range hood spare parts disclosed in the present invention have the following technical effects:

[0044] The spare parts in the supplier's spare parts warehouse are transported to the company's internal integrated spare parts warehouse, and then flow to the self-service spare parts shelves, forming a multi-level inventory structure for range hood spare parts, which improves the management efficiency of range hood spare parts in the multi-level inventory structure and adapts to rapid response to maintenance needs;

[0045] Based on the data collected from the internal and external digital systems of the enterprise, a more accurate inventory decision of smoke machine spare parts is provided through a data-driven approach, reducing inventory redundancy and improving inventory turnover rate; the interconnection of the internal and external digital systems of the enterprise is realized, and the data of spare parts inventory locations at all levels are integrated for unified management and analysis;

[0046] Use big data and artificial intelligence technologies to improve the accuracy of forecasting spare parts demand trends, thereby optimizing inventory allocation, reducing spare parts inventory costs, and reducing unnecessary inventory backlogs through intelligent forecasting;

[0047] Based on the optimized random forest predictor, the spare parts demand in the next period after the current time is predicted to allocate the types and quantities of spare parts inventory at all levels. Through intelligent allocation methods, the speed of obtaining and delivering spare parts during the repair and maintenance of smoke machine equipment can be accelerated to reduce downtime;

[0048] Uniformly classify the spare parts according to their specific parameters, and ensure the availability of spare parts and optimize the spare parts inventory strategy through detailed classification and value analysis of spare parts;

[0049] The random forest model is used to predict spare parts demand, and the model parameters are optimized to improve the accuracy and reliability of the prediction.

[0050] Further features and advantages of the present invention will become apparent from the following detailed description of exemplary embodiments of the present invention with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention.

[0052] Figure 1 It is a structural diagram of the multi-level inventory intelligent distribution system for range hood spare parts of the present invention;

[0053] Figure 2 for Figure 1 Schematic diagram of the internal spare parts inventory of the enterprise;

[0054] Figure 3 for Figure 1 Schematic diagram of data collection in the multi-level inventory system of smoke machine spare parts;

[0055] Figure 4The overall flow chart of the data-driven multi-level inventory intelligent allocation method for range hood spare parts of the present invention;

[0056] Figure 5 for Figure 4 The overall flow chart of the smoke machine spare parts classification algorithm flow chart;

[0057] Figure 6 for Figure 4 The flowchart for building a random forest model is shown in the overall flowchart of ;

[0058] Figure 7 for Figure 4 Flowchart for optimizing random forest model parameters in the overall flowchart of ;

[0059] Figure 8 for Figure 4 Schematic diagram of time interval selection in the overall flow chart.

[0060] The following are marked in the figure:

[0061] 100. Hood equipment; 101, 102: machine repair personnel; 103, 104: self-service spare parts shelves; 105. Integrated spare parts warehouse; 106. Spare parts management personnel; 107. Spare parts management system access terminal; 108. Spare parts management system server; 109. Supplier spare parts warehouse;

[0062] 201, integrated spare parts warehouse; 202, 203: self-service spare parts shelves; 204, 205: hood equipment; 206, supplier spare parts inventory;

[0063] 300. Mechanic repair personnel; 301. Self-service spare parts shelf; 302. QR code of self-service spare parts shelf; 303. Spare parts management personnel; 304. Spare parts management system access terminal; 305. Supplier spare parts inventory; 306. Supplier spare parts management system; 307. Spare parts management system server; 308. Enterprise internal digital system; 309. Enterprise external digital system; 310. Integrated spare parts warehouse. DETAILED DESCRIPTION

[0064] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement of components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present invention unless otherwise specifically stated.

[0065] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.

[0066] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.

[0067] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0068] The present invention proposes an embodiment of a data-driven multi-level inventory intelligent distribution system for range hood spare parts. Specifically, Figure 1 As shown, it includes the following structures:

[0069] The smoke machine equipment 100, several mechanics 101, 102, several self-service spare parts shelves 103, 104, an integrated spare parts warehouse 105, a spare parts management personnel 106, a spare parts management system access terminal 107, a spare parts management system server 108 and a supplier spare parts warehouse 109; the smoke machine equipment 100 is maintained and repaired by several full-time mechanics 101, 102 in sections to ensure the efficient operation of the entire smoke machine equipment 100; several self-service spare parts shelves 103, 104 are deployed near the smoke machine equipment 100 to facilitate the mechanics 101, 102 to obtain spare parts; the integrated spare parts warehouse 105 is equipped with a plurality of self-service spare parts shelves 103, 104, and a plurality of ... The spare parts warehouse 105 is the spare parts warehouse within the enterprise, responsible for the unified management of spare parts ordering, replenishment, and distribution of spare parts to several self-service spare parts shelves 103 and 104; the spare parts management personnel 106 operates in the integrated spare parts warehouse 105, accesses the spare parts management system through the spare parts management system access terminal 107, and makes corresponding spare parts decisions and operation instructions; the spare parts management system is installed on the spare parts management system server 108; in addition, the supplier spare parts warehouse 109 also stores some spare parts, and when replenishing and ordering, the supplier spare parts warehouse 109 delivers the spare parts to the integrated spare parts warehouse 105 within the company.

[0070] It should be noted that, from the perspective of spare parts flow, in the multi-level inventory intelligent distribution system for smoke machine spare parts, the spare parts in the supplier spare parts warehouse 109 flow into the integrated spare parts warehouse 105 within the company, and then flow to the self-service spare parts shelves 103 and 104, forming a multi-level inventory structure for smoke machine spare parts; from the perspective of data flow, in the multi-level inventory intelligent distribution system for smoke machine spare parts, the spare parts management system server 108 is used as the connection center to connect the smoke machine equipment 100, the mechanics 101 and 102, the self-service spare parts shelves 103 and 104, and the spare parts management system access terminal 107, and the supplier spare parts warehouse 109 through a computer communication network. The embodiment of the present invention realizes multi-level inventory management of smoke machine spare parts through a data-driven intelligent distribution system, improves overall efficiency, can dynamically adjust the inventory structure according to actual needs, reduce unnecessary inventory backlogs, and reduce inventory costs.

[0071] Figure 2 Further explanation Figure 1 Schematic diagram of the multi-level inventory system for range hood spare parts: the self-service spare parts shelves 202 and 203 constitute the first-level spare parts inventory, which is the spare parts inventory closest to the range hood equipment 204 and 205; the range hood equipment 204 and 205 often occupy a large space and are often far away. Therefore, in order to adapt to the needs of rapid response to maintenance, the self-service spare parts shelves 202 and 203 are flexibly located and are optimized according to the walking distance of the mechanics; the integrated spare parts warehouse 201 constitutes the second-level spare parts inventory. When the spare parts stored in the self-service spare parts shelves 202 and 203 are lower than the safety inventory threshold, or the self-service spare parts shelves 202 and 203 do not have the corresponding spare parts reserves, the mechanics directly initiate a request for use to the integrated spare parts warehouse 201, and the integrated spare parts warehouse 201 serves as the second-level The spare parts inventory distributes corresponding spare parts to the primary spare parts inventory or mechanics; the primary spare parts inventory and the secondary spare parts inventory within the dotted box in the figure constitute the internal spare parts inventory 200 of the enterprise; in addition, not all spare parts must be purchased and stored in the internal spare parts inventory 200 of the enterprise. For some special spare parts with high value and slow circulation, the supply chain management inventory (VMI) model is often adopted, that is, the corresponding spare parts are not purchased in advance, but the data connection is maintained with the supplier spare parts inventory 206. When these spare parts are needed, the supplier spare parts inventory 206 identifies the spare parts demand and then ships them. Therefore, the supplier spare parts inventory 206 constitutes a three-level spare parts inventory; the above three-level spare parts inventory system is an embodiment of the multi-level inventory system for range hood spare parts of the present invention.

[0072] Figure 3 Further explained Figure 1 The schematic diagram of data collection in the multi-level inventory system of smoke machine spare parts in the figure includes collecting data from the internal digital system 308 of the enterprise in the dotted box and the external digital system 309 of the enterprise in the solid box; the data collected by the internal digital system 308 of the enterprise includes: the data of the mechanic 300 picking up spare parts from the automatic spare parts shelf 301 in the first-level spare parts inventory, and the data containing six basic elements of spare parts picking (mechanic number, self-service spare parts shelf number, spare parts material) generated by the mechanic 300 by scanning the QR code 302 of the self-service spare parts shelf Number, quantity, collection time, and delivery time); if the spare parts needed by the mechanic 300 are not in stock on the self-service spare parts shelf 301 of the first-level spare parts inventory, the mechanic 300 initiates a collection application to the integrated spare parts warehouse 310 and waits for delivery. During this process, six basic elements of spare parts application data are generated (mechanic number, self-service spare parts shelf number, spare parts material number, quantity, application time, and delivery time); therefore, if the self-service spare parts shelf 301 already has reserves, the collection time and the delivery time are the same time.

[0073] Similarly, the enterprise's internal digital system 308 also includes secondary spare parts inventory collection data. The spare parts manager 303 is responsible for managing and scheduling the integrated spare parts warehouse 310, accessing the spare parts management system server 307 through the spare parts management system access terminal 304, collecting spare parts inventory data, and making ordering, replenishment, and allocation decisions; the operation of the spare parts manager 303 generates data containing six basic elements of spare parts allocation (spare parts manager number, spare parts material number, quantity, allocation warehouse number, ordering time, delivery time), where the allocation warehouse refers to the various warehouse locations of the primary spare parts inventory and the secondary spare parts inventory.

[0074] In addition, the enterprise's external digital system 309 is composed of various supplier spare parts management systems 306. The inventory status of the supplier spare parts inventory 305 is connected to the spare parts management system server 307 in the enterprise's internal digital system 308 through the supplier spare parts management system 306. The supplier spare parts inventory 305 is composed of a three-level spare parts inventory. The collected data includes six basic elements of spare parts allocation data (supplier number, spare parts material number, quantity, allocation time, limited time, and delivery time). The allocation time refers to the time when the allocation request is initiated by the internal spare parts manager 303 of the enterprise, the limited time is the latest time required for delivery by the spare parts manager 303, and the delivery time is the actual delivery time of the spare parts. A three-level inventory consisting of self-service spare parts shelves and supplier spare parts inventory is introduced into the multi-level inventory system of smoke machine spare parts to achieve interconnection between the internal and external digital systems of the enterprise, facilitate the collection of the collection data, application data, allocation data, and transfer data of spare parts inventory locations at all levels, and use the historical demand satisfaction rate and historical demand response time of spare parts inventory locations at all levels, as well as a set of spare parts classification methods provided, to train a random forest model, analyze and predict the demand for spare parts inventory locations at all levels, so as to perform intelligent spare parts allocation at all levels of spare parts inventory.

[0075] Figure 4 The overall flow chart of a data-driven multi-level inventory intelligent allocation method for range hood spare parts of the present invention includes the following steps:

[0076] Step 1: Start 400: The allocation process is generally divided into three stages;

[0077] The first stage: model training, including the following steps:

[0078] Step 2: Select the time interval 401 of the model training data set: generally select a time interval before the current time, such as 12 months or 24 months;

[0079] Step 3: Generate spare parts demand time series data set 402: In the spare parts intelligent prediction system of the present invention, four time series data sets are collected, namely, a spare parts requisition time series data set including six spare parts requisition basic element data, a spare parts application time series data set including six spare parts application basic element data, a spare parts allocation time series data set including six spare parts allocation basic element data, and a spare parts transfer time series data set including six spare parts transfer basic element data;

[0080] The four time series data sets correspond to spare parts inventory at each level. Therefore, for the first-level spare parts inventory, each record of the spare parts demand time series data set contains (self-service spare parts shelf number, spare parts material number, quantity, collection time (or application time), delivery time); for the second-level spare parts inventory, each record of the spare parts demand time series data set contains (allocation warehouse number, spare parts material number, quantity, ordering time, delivery time); and for the third-level spare parts inventory, each record of the spare parts demand time series data set contains (supplier number, spare parts material number, quantity, transfer time, delivery time); these three types of spare parts demand time series data sets can be unified into a format of time series data;

[0081] Step 4: Select the first 80% of the data set in the time interval as the training data set 403: The data used for training the model is the first 80% of the data set in the time interval, rather than randomly selecting 80% of the data set, so that after training a basic model parameter, the later 20% of the data set is used to optimize the model parameter;

[0082] Step 5: Spare parts classification 404: Different types of spare parts have different inventory strategies and quantities. Different decisions can be made after uniformly classifying them according to the frequency, importance, and value of spare parts.

[0083] Step 6: Input random forest model 405: Random forest model is an algorithm that combines the ideas of bagging and random subspace. It is an integrated predictor model formed by multiple decision trees. The random forest algorithm uses multiple decision trees as base learners and uses bagging to generate different training sets. Random attribute selection is introduced in the training process of a single tree. After multiple trainings, multiple different decision trees are obtained, and these trees are combined into a random forest predictor. The random forest model and historical data are used to improve the prediction accuracy of spare parts demand and reduce prediction errors.

[0084] The second stage: model optimization, including the following steps:

[0085] Step 7: Use the last 20% of the data set in the time interval as the test data set 406: The 20% of the data set in the time interval closer to the current time contains more accurate current information and can be better used to optimize the random forest model;

[0086] Step 8: Optimizing the parameters of the random forest model 407: Continuously adjust the parameters of the random forest model to minimize the fitting error of the model for 20% of the test data set; preferably, the mean absolute percentage error ( ) as the criterion for fitting error; the mean absolute percentage error is defined as the model prediction value With the true value The distance between the two is divided by the true value The average value of is: ;in is the total amount of test data set data, is the subscript for each pair of predicted value and true value;

[0087] Phase 3: Spare Parts Forecasting and Allocation:

[0088] Step 9: Predicting spare parts demand in the next period 408: Using the optimized random forest predictor to predict spare parts demand in the next period of the current time, so as to allocate the types and quantities of spare parts inventory at all levels;

[0089] Step 10: End 409: The overall process ends.

[0090] The embodiment of the present invention effectively reduces the downtime of the range hood equipment due to the lack of spare parts through a quick response spare parts collection and distribution process. The spare parts collection process is optimized, so that the maintenance personnel can obtain the required spare parts more quickly and speed up the maintenance work.

[0091] Figure 5 Further explanation Figure 4 The flow chart of the spare parts classification algorithm of China Tobacco Machinery Co., Ltd. includes the following steps:

[0092] Step 1: Start with 500;

[0093] Step 2: FSN classification 501: Based on the spare parts consumption frequency, average demand occurrence interval analysis 517 is performed to classify spare parts into three types: F (fast flow) 503, S (slow flow) 502, and N (non-flow) 504. Among them, the N (non-flow) 504 type is not further classified and directly jumps to step 7;

[0094] As a preferred embodiment, this embodiment sorts the spare parts according to the average interval of spare parts demand, and then calculates the average annual consumption amount of spare parts and the annual cumulative consumption amount and proportion of spare parts. When the cumulative consumption amount reaches 70% of the total spare parts consumption amount, these spare parts are classified as Class F (fast flow); spare parts with cumulative consumption amounts in the range of 70%-90% are classified as Class S (slow flow), and the rest of the spare parts are classified as Class N (non-flow), that is, the 70%-20%-10% principle;

[0095] Step 3: Classify 505 the importance of Class S (slow flow) 502: Using the cluster analysis 518 algorithm, the Class S (slow flow) 502 spare parts are further classified into two categories: Class V (important) 507 and Class D (unimportant) 508, and then jump to Step 7;

[0096] Step 4: Classify 506 the importance of Class F (fast-flowing) 503: Using the cluster analysis 518 algorithm, the Class F (fast-flowing) 503 spare parts are further divided into three categories: Class V (important) 509, Class D (unimportant) 510, and Class E (generally important) 511;

[0097] As a preferred embodiment, this embodiment adopts the Fuzzy C-Means Algorithm (FCM), which is the most widely used and relatively successful algorithm among many fuzzy clustering algorithms. The algorithm idea is to obtain the membership of each sample point to all class centers by optimizing the objective function, so as to determine the category of the sample point to achieve the purpose of automatically classifying the sample data. Each sample is assigned a membership function belonging to each cluster, and the samples are classified according to the size of the membership value; the spare parts classification index selects five types of index factors, namely: economic characteristics, the selected indicators include spare parts unit price, and the data source is historical order data; demand characteristics, the selected indicators include monthly average demand and demand fluctuation coefficient, and the data source is historical monthly collection volume; supply characteristics, the selected indicators include average delivery time and delivery time variation rate, and the data source is historical spare parts delivery time data; importance characteristics, the selected indicators include criticality and replaceable rows, and the data source is expert comprehensive evaluation quantitative data;

[0098] In addition, as a preference, the classification index determination criteria of Class V (Important) 509, Class D (Unimportant) 510, and Class E (Generally Important) 511 in this embodiment are: Class V (Important) 509 meets the determination criteria of large demand fluctuation coefficient, long average delivery time, large delivery time variation rate, high criticality, and low replaceability; Class E (Generally Important) 511 meets the determination criteria of medium demand fluctuation coefficient, medium average delivery time, medium delivery time variation rate, medium criticality, and medium replaceability; Class D (Unimportant) 510 meets the determination criteria of small demand fluctuation coefficient, short average delivery time, small delivery time variation rate, low criticality, and high replaceability;

[0099] Step 5: Value classification 512 of Class V (Important) 509 spare parts of Class F (Fast Flow) 503: Use the spare parts value analysis 519 method to further classify such spare parts into three categories: Class A, B, and Class C 514, and jump to step 7;

[0100] Step 6: Value classification 513 of Class E (generally important) 511 spare parts of Class F (fast-moving) 503: Use the spare parts value analysis 519 method to further classify such spare parts into three categories: Class A, B, and Class C 515, and jump to step 7;

[0101] As a preference, this embodiment performs spare parts price analysis 519 on spare parts based on the unit price of spare parts, and uses ABC to represent the value level of spare parts, and the rules are: spare parts with a unit price lower than 100 yuan are classified as Class C spare parts, i.e., spare parts with a unit price of 100-10,000 yuan are classified as Class B spare parts, i.e., spare parts with a general value; spare parts with a value greater than 10,000 yuan are classified as spare parts with a high value;

[0102] Step 7: End 516.

[0103] The embodiment of the present invention ensures high availability and reasonable inventory control of spare parts through detailed classification and value analysis of spare parts.

[0104] Figure 6 Further explanation Figure 4 Flowchart for building a random forest model in the overall flowchart;

[0105] Step 1: The first 80% data set 600 of the spare parts demand time interval is a spare parts demand time series data set after the four time series data sets, namely, spare parts collection time series data set, spare parts application time series data set, spare parts allocation time series data set, and spare parts transfer time series data set, are merged;

[0106] Step 2: randomly extracting the first sample set 601: using a sampling technique of random selection and replacement, randomly extracting sample data from the first 80% data set of the spare parts demand time interval;

[0107] Step 3: input the first weak predictor 604: the first weak predictor adopts classification and regression decision tree (CART), and the initial parameters include: the maximum depth of the decision tree, the evaluation standard, the minimum number of samples for splitting nodes when classifying, the minimum number of samples for leaf nodes, the minimum standard for the sum of weights of all leaf nodes, the judgment threshold of the error reduction after splitting, the judgment threshold of whether the node is split, and other parameters; preferably, the initial parameters are set as follows: the maximum depth of the decision tree is 10, the mean square error is selected as the evaluation standard, the minimum number of samples for splitting nodes when classifying is 2, the minimum number of samples for leaf nodes is 1, the minimum standard for the sum of weights of all leaf nodes is 0, the judgment threshold of the error reduction after splitting is 0, and the judgment threshold of whether the node is split is 0.0000001;

[0108] Step 4: Make a first decision 607: until the first weak predictor outputs an evaluation criterion less than a mean square error threshold;

[0109] Similarly, repeat steps 2 to 4: randomly select a second sample set 602 → input a second weak predictor 605 → make a second decision 608;

[0110] Similarly, repeat steps 2 to 4;

[0111] Repeat steps 2 to 4: randomly extract the nth sample set 603 → input the nth weak predictor 606 → make the nth decision 609, where n is the number of decision trees in the random forest model. Preferably, the parameter is initially set to 200;

[0112] Step 5: Construct a basic random forest model 610: Select five feature variables, including: historical demand data, historical demand satisfaction rate, historical demand response time, inventory location, and spare parts type; the historical demand data refers to the historical application data and collection data of the first-level spare parts inventory, the historical allocation data of the second-level spare parts inventory, and the historical transfer data of the third-level spare parts inventory; the historical demand satisfaction rate refers to the proportion of spare parts in the spare parts inventory when the spare parts inventory operators at all levels submit spare parts applications. For the first-level spare parts inventory, the proportion of spare parts that can be directly collected by mechanics on the self-service spare parts shelves; for the second-level spare parts inventory, the proportion of spare parts that can be directly collected by mechanics on the self-service spare parts shelves when the spare parts management personnel receive applications from mechanics The historical demand response time is the time interval from when the spare parts inventory operators at all levels submit spare parts applications to when the spare parts are delivered. For the first-level spare parts inventory, if the mechanic directly collects the spare parts from the self-service spare parts shelf, the demand response time is zero. If the spare parts cannot be collected directly from the self-service spare parts shelf, the waiting time for the integrated spare parts warehouse to deliver the spare parts is the demand response time. For the second-level spare parts inventory, the waiting time for delivery after the spare parts manager sends an application to the supplier is the demand response time. The inventory location is the specific location of the spare parts inventory at all levels. Finally, the spare parts type is the result of classification based on the spare parts classification algorithm.

[0113] Figure 7 Further explanation Figure 4 The flowchart for optimizing the parameters of the random forest model in the overall flowchart includes the following steps:

[0114] Step 1: Start the loop and traverse the number of decision trees in the random forest model, which is 700;

[0115] Step 2: Start the loop and traverse the decision tree to a maximum depth of 701;

[0116] Step 3: Start the loop and traverse the minimum number of samples of leaf nodes, which is 702;

[0117] Step 4: Execute in the loop: read the 20% data set after the time interval 703, and calculate the MAPE evaluation index 704 one by one with the predicted value;

[0118] Step 5: Select the set of model parameters 705 with the smallest MAPE as the optimal random forest model.

[0119] like Figure 8 Shown Figure 4Schematic diagram of time interval selection in the overall process, specifically, a time interval before the current time includes the first 80% data set 800 and the last 20% data set 801, the first 80% data set 800 is used as the training data set, the last 20% data set 801 is used as the test data set, and the spare parts demand 802 of the next period is predicted at the current time; in the next period, the selected time interval is shifted backward, including the first 80% data set 803 of the next period and the last 20% data set 804 of the next period, the first 80% data set 803 of the next period is used as the training data set, and the last 20% data set 804 of the next period is used as the test data set, and the spare parts demand 805 of the next period is predicted to allocate the types and quantities of spare parts inventories at all levels.

[0120] The multi-level inventory intelligent distribution system and method of the smoke machine spare parts of the present invention realizes the intelligentization of spare parts demand prediction, inventory allocation and optimization by constructing a multi-level inventory structure and using big data and artificial intelligence technology. Among them, the multi-level inventory structure not only improves the management efficiency of spare parts, but also adapts to the requirements of rapid response to maintenance needs and reduces the downtime caused by spare parts shortage. The system collects multi-dimensional data such as spare parts usage, maintenance records, market demand, etc. through the internal digital system of the enterprise (such as ERP, CRM, etc.) and the external digital system (such as supplier management system, etc.). Based on the collected data, the system uses data analysis technology to provide more accurate spare parts inventory decisions, reduce inventory redundancy, and improve inventory turnover. The interconnection and interoperability of the internal and external digital systems of the enterprise is realized, and the data of spare parts inventory locations at all levels are integrated to facilitate unified management and analysis, and improve the transparency and efficiency of inventory management. Based on the optimized random forest predictor, the spare parts demand in the period after the current time can be predicted, and the types and quantities of spare parts inventory at all levels can be allocated accordingly. Spare parts classification and value analysis ensure the availability and timeliness of spare parts and optimize inventory strategies; through intelligent allocation and rapid response mechanisms, downtime caused by spare parts shortages is significantly reduced. In summary, the multi-level inventory intelligent allocation system and method for smoke machine spare parts integrates modern logistics and information technology to achieve efficient, intelligent and refined spare parts management, providing strong support for the repair and maintenance work of enterprises.

[0121] The above describes in detail the structure, features and effects of the present invention based on the embodiments shown in the drawings, but the above is only a preferred embodiment of the present invention. It should be noted that the technical features involved in the above embodiments and their preferred methods can be reasonably combined and matched into a variety of equivalent schemes by those skilled in the art without departing from or changing the design ideas and technical effects of the present invention; therefore, the present invention is not limited to the scope of implementation shown in the drawings, and all changes made in accordance with the concept of the present invention, or modifications to equivalent embodiments with equivalent changes, which still do not exceed the spirit covered by the specification and drawings, should be within the protection scope of the present invention.

Claims

1. A multi-level inventory intelligent distribution system for range hood spare parts, characterized in that: It includes range hood equipment, several self-service spare parts shelves, an integrated spare parts warehouse, a spare parts management system access terminal, a spare parts management system server and a supplier spare parts warehouse; wherein the self-service spare parts shelves are deployed near the range hood equipment, and the integrated spare parts warehouse is used for unified management of spare parts ordering, replenishment, and distribution of spare parts to several self-service spare parts shelves; spare parts management personnel access the spare parts management system through the spare parts management system access terminal; the spare parts management system is installed on the spare parts management system server; the supplier spare parts warehouse stores some spare parts for distribution of spare parts to the company's internal integrated spare parts warehouse during replenishment and ordering; with the spare parts management system server as the connection center, the range hood equipment, the self-service spare parts shelves, the spare parts management system access terminal and the supplier spare parts warehouse are connected together through a computer communication network.

2. The multi-level inventory intelligent distribution system for range hood spare parts according to claim 1 is characterized in that: The spare parts in the supplier's spare parts warehouse are transported to the company's internal integrated spare parts warehouse, and then flow to the self-service spare parts shelves, forming a multi-level inventory structure for range hood spare parts, specifically including: A primary spare parts inventory consisting of self-service spare parts shelves; The secondary spare parts inventory composed of the integrated spare parts warehouse. When the spare parts stored in the self-service spare parts shelf are lower than the safety stock threshold, the mechanic directly initiates a request for use to the integrated spare parts warehouse. The integrated spare parts warehouse, as the secondary spare parts inventory, allocates the corresponding spare parts to the primary spare parts inventory or the mechanic. The three-level spare parts inventory consists of the supplier spare parts inventory.

3. The multi-level inventory intelligent distribution system for range hood spare parts according to claim 2 is characterized in that: Data collection is performed through the multi-level inventory structure of smoke machine spare parts, including collecting data from the enterprise's internal digital system and the enterprise's external digital system; wherein the data collected by the enterprise's internal digital system includes at least one of the basic element data of spare parts requisition, the basic element data of spare parts application, and the basic element data of spare parts allocation; the enterprise's external digital system is composed of the spare parts management systems of various suppliers, and the data collected by the enterprise's external digital system includes the basic element data of spare parts allocation.

4. A multi-level inventory intelligent allocation method for range hood spare parts, characterized in that: Deployed in the spare parts management system in claims 1 to 3, the method comprises the following steps: Step 1: Select a specific time interval before the current time as the data collection time interval; Step 2: obtaining a spare parts demand time series data set within the selected data collection time interval, the spare parts demand time series data set including: a spare parts requisition time series data set including spare parts requisition basic element data, a spare parts application time series data set including spare parts application basic element data, a spare parts allocation time series data set including spare parts allocation basic element data, and a spare parts transfer time series data set including spare parts transfer basic element data; Step 3: Select a part of the spare parts demand time series data set as the training data set; Step 4: Input the training data set into the random forest model to generate a random forest predictor; Step 5: Select another part of the spare parts demand time series data set as the test data set; Step 6: Adjust the parameters of the random forest model so that the fitting error of the random forest model for the test data set is less than the error threshold, and obtain an optimized random forest predictor; Step 7: Based on the optimized random forest predictor, the spare parts demand in the period after the current time is predicted to allocate the types and quantities of spare parts inventory at all levels.

5. The multi-level inventory intelligent allocation method for range hood spare parts according to claim 4 is characterized in that: After step 4, the method further includes: uniformly classifying the range hood spare parts according to the specific parameters to make different decisions, wherein the classification algorithm process includes: According to the frequency of spare parts consumption, the average demand occurrence interval analysis is carried out, and the spare parts are divided into three types: F (fast flow), S (slow flow), and N (non-flow); Cluster analysis algorithm is used to further classify the S-class (slow-flow) spare parts into two categories: V-class (important) and D-class (unimportant); Cluster analysis algorithm is used to further divide Class F (fast-flowing) spare parts into three categories: Class V (important), Class D (unimportant), and Class E (generally important).

6. The multi-level inventory intelligent allocation method for range hood spare parts according to claim 5 is characterized in that: Also includes: Using the spare parts value analysis method, the V-type (important) spare parts of the F-type (fast-moving) are further divided into three categories: A, B, and C; The spare parts value analysis method is used to further divide the F (fast-moving) and E (generally important) spare parts into three categories: A, B, and C.

7. The multi-level inventory intelligent allocation method for range hood spare parts according to claim 5 is characterized in that: In the process of classifying spare parts into three types, namely Class F (fast flow), Class S (slow flow), and Class N (non-flow), they are sorted according to the average interval of spare parts demand, and then the average annual consumption amount of spare parts and the annual cumulative consumption amount and proportion of spare parts are calculated. When the cumulative consumption amount reaches 70% of the total spare parts consumption amount, these spare parts are classified as Class F (fast flow); spare parts with cumulative consumption amounts in the range of 70%-90% are classified as Class S (slow flow), and the rest of the spare parts are classified as Class N (non-flow).

8. The multi-level inventory intelligent allocation method for range hood spare parts according to claim 4 is characterized in that: In step 4, the construction process of the random forest model includes the following steps: Step 41: obtaining the data set of the first 80% of the spare parts demand time interval, the data set including the spare parts collection time series data set, the spare parts application time series data set, the spare parts allocation time series data set and the spare parts transfer time series data set; Step 42: Using a random selection and replacement sampling technique, randomly select sample data from the first 80% data set of the spare parts demand time interval; Step 43: Input the randomly selected sample data into the weak predictor, which uses a classification and regression decision tree (CART); Step 44: make the first decision: until the weak predictor outputs an evaluation criterion less than the mean square error threshold; repeat steps 42 to 44: randomly select the nth sample set → input the nth weak predictor → make the nth decision, where n is the number of decision trees in the random forest model; Step 45: Construct a basic random forest model, wherein the selected characteristic variables include: historical demand data, historical demand satisfaction rate, historical demand response time, inventory location, and spare parts type, wherein the historical demand data includes historical application data and collection data of the first-level spare parts inventory, historical allocation data of the second-level spare parts inventory, and historical transfer data of the third-level spare parts inventory.

9. The multi-level inventory intelligent allocation method for range hood spare parts according to claim 4 is characterized in that: In step 6, the mean absolute percentage error (MAPE) is used as the judgment indicator of the fitting error. The mean absolute percentage error is defined as the distance between the model predicted value F(t) and the true value A(t) divided by the average value of the true value A(t), that is: ； Among them, N is the total amount of test data set data, is the subscript of each predicted value-true value pair.

10. The multi-level inventory intelligent allocation method for range hood spare parts according to claim 9, characterized in that: In step 6, the process of adjusting the parameters of the random forest model includes the following steps: Step 61: Start a loop to traverse the number of decision trees in the random forest model; Step 62: Start a loop to traverse the maximum depth of the decision tree; Step 63: Start a loop to traverse the minimum number of samples of leaf nodes; Step 64: Execute in the loop: read the 20% data set after the time interval, and calculate the MAPE evaluation index one by one with the predicted value; Step 65: Select the set of model parameters with the smallest MAPE as the optimal random forest model.

Images