CN116029421A - Data-driven after-sales service electronic product part stock quantity calculating method - Google Patents
Data-driven after-sales service electronic product part stock quantity calculating method Download PDFInfo
- Publication number
- CN116029421A CN116029421A CN202211550735.2A CN202211550735A CN116029421A CN 116029421 A CN116029421 A CN 116029421A CN 202211550735 A CN202211550735 A CN 202211550735A CN 116029421 A CN116029421 A CN 116029421A
- Authority
- CN
- China
- Prior art keywords
- electronic product
- month
- stock quantity
- data
- prediction model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a data-driven after-sales service electronic product part stock quantity calculating method, which is characterized in that a simple and efficient optimizing method is provided for the problems of electronic product stock quantity prediction and inventory cost optimization, historical data is fully utilized, a prediction model is established by combining an opportunity constraint method, a robust optimizing solution is obtained by efficiently solving the prediction model, a stock quantity purchasing decision scheme in a subsequent time period can be given out by adopting the prediction model obtained by solving, an actual after-sales service manufacturer is guided to make a corresponding decision, and under the condition of meeting the user stock quantity replacing requirement, the related cost is optimized.
Description
Technical Field
The invention belongs to the technical field of spare part demand quantity calculation, and particularly relates to a data-driven after-sales service electronic product part stock quantity calculation method.
Background
Along with the continuous development of information technology, the dependence degree of daily work and life of people on electronic products is higher and higher at present, and electronic products such as notebook computers, mobile phones and printers are very popular, and especially mobile phones are an indispensable important component of daily life as a communication tool. The high-strength use accelerates the ageing of the parts of the electronic product, and the after-sales maintenance service demand of consumers on the electronic product is obviously improved. When the user needs to repair the electronic product and replace the spare parts in the warranty period but cannot meet the requirement, the user can generate dissatisfaction on related brands of products, so that the brand reputation is influenced, and the sales condition of subsequent products is negatively influenced. Thus, electronic product after-market service providers need to manufacture the required repair parts during warranty periods to meet the customer's demand for after-market service. On the other hand, the rapid updating of the electronic product model leads to the failure of the part suppliers to maintain the specific spare part production line for a long time, and the warranty period of the electronic product is generally about one to three years, so that the related part manufacturers cannot maintain the spare part production line of the specific model continuously, and the continuous supply of the spare parts in the warranty period cannot be ensured.
In order to ensure the reliability of the consumer to the brand and the quality of the after-sales service, the after-sales service provider needs to prepare enough spare parts before the production line of the related products goes off line so as to ensure that the maintenance and replacement requirements of the consumer are met. Thus, the after-market service provider needs to make decisions about the purchase amount of the relevant spare parts according to the needs of the consumer, thereby supplementing the inventory. Meanwhile, the purchase, inventory and destruction costs of spare parts are brought about by the order of spare parts, and the cost of after-sales service is greatly increased when the purchase amount of spare parts is far greater than the required amount. Therefore, it is necessary to accurately predict the required amount of spare parts. Traditional spare part prediction methods often consider large equipment, such as airplanes, manufacturing equipment and other expensive equipment (caps) with long warranty periods, which are confirmed by the purchasing population and have long maintenance periods, and the spare parts must be maintained or replaced once the product is damaged. Electronic products such as mobile phones, notebook computers and printers are user-type products (consumer products) sold in large quantities, the sales volume is large, whether maintenance is affected by factors such as cost, user experience, use time, whether new products are marketed or not, uncertainty is high, and related researches are deficient. Therefore, the problems of forecasting the spare part demand of the electronic product and optimizing the inventory management have academic research value and application prospect.
In view of the above-mentioned objectives, there have been related studies including Kim t.y. et al predicting the number of spare parts required at < Spare part demand forecasting for consumer goods using Installed Base information > using the Installed Base (IB) of spare parts of an electronic product, which classifies the Installed Base as: calculating the number of devices (Lifetime IB, IBL) existing in the market according to the expected service life; calculating the number of devices (IBW) existing in the market according to the Warranty period; considering that maintenance is not economical and products are discarded, the number of devices (IBE) existing in the market; the number of devices currently on the market (Mixed IB, IBM) is considered in terms of consumer maintenance costs and profits. Based on the concept, the work utilizes the survival probability distribution to simulate the damage condition of the electronic product, fits the functional relation between the demand number D and IB, and predicts the demand number of the mobile phone in the subsequent time period.
According to the prediction method, the historical data is used for linear fitting to obtain the corresponding coefficient, but under the condition of simple assumption and practical application, only the trend of the spare part demand can be predicted generally, but accurate spare part purchasing decision suggestion cannot be given, and the factors such as purchasing, storing and destroying costs are not considered, so that the obtained prediction result is not accurate enough.
Disclosure of Invention
In view of the above, the invention provides a data-driven after-sales service electronic product part stock quantity calculating method, which realizes the quantitative calculation of the electronic product part stock quantity based on historical data.
The invention provides a data-driven after-sales service electronic product part stock quantity calculating method, which comprises the following steps:
step 2, constructing a mixed integer programming model based on a data-driven opportunity constraint method to establish an electronic product part stock quantity prediction model, wherein the method is as follows:
wherein Q (i) is the number of spare parts purchased from suppliers in the ith month, r is a decision variable, X (i) is input data of the ith month, c p Cost of acquisition for a single spare part c b Cost per unit time for backlog of individual spare parts c h For the cost per unit time of a single spare part out-of-stock, b (i) is the spare part inventory backlog for month i, h (i) is the spare part out-of-stock for month i;
step 3, inputting the electronic product input sample set into the electronic product part stock quantity prediction model, and solving the electronic product part stock quantity prediction model by using a solver to obtain a determined electronic product part stock quantity prediction model;
and 4, in actual use, according to the input variable preprocessing mode in the step 1, acquiring actual historical data occurring before the predicted time period as an actual input variable, and inputting the actual input variable into the electronic product part stock quantity prediction model obtained in the step 3 to obtain the electronic product part stock quantity required to be purchased in the predicted time period.
Further, the input variable preprocessing method in step 1 establishes a w+3-dimensional input variable X (i) for acquiring the sales of W months before the predicted time period:
X(i)＝{B(i-W),B(i-W+1),...,B(i-1),1,D(i-1),i}
wherein B (i-W) is the selling amount of the electronic product in the (i-W) month, B (i-W+1) is the selling amount of the electronic product in the (i-W+1) month, B (i-1) is the selling amount of the electronic product in the (i-1) month, and D (i-1) is the demand of the specific spare part in the (i-1) month.
Further, the data-driven opportunity constraint inequality constructed based on the sample averaging method in the electronic product part stock quantity prediction model is as follows:
wherein M is a very large integer, S (i) is the stock quantity of spare parts in month i; gamma (i) is an integer variable, and when the value is 1, the requirement of the spare part in the ith month is not met, and when the value is 0, the requirement of the spare part in the ith month is met; alpha is the opportunity constraint threshold.
Further, the opportunity constraint threshold has a value of 0.01.
The beneficial effects are that:
the invention provides a simple and efficient optimization method aiming at the problems of electronic product spare part prediction and inventory cost optimization, fully utilizes historical data, combines an opportunity constraint method to establish a prediction model, obtains a robust optimization solution by efficiently solving the prediction model, adopts the prediction model obtained by solving to give a spare part purchase decision scheme in a subsequent time period, guides an actual after-sales service manufacturer to make a corresponding decision, optimizes related cost under the condition of meeting the spare part replacement requirement of a user, and has more sufficient utilization of data and more excellent optimization performance compared with the existing spare part requirement prediction method.
Drawings
Fig. 1 is a schematic diagram of an after-market electronic product spare part inventory optimization problem.
Fig. 2 is a schematic diagram of actual sales data of a certain model of mobile phone.
FIG. 3 is a schematic diagram of the monthly spare part purchase amount and the inventory result under the actual demand by using the data-driven after-sales electronic product part stock amount calculation method provided by the invention.
FIG. 4 is a schematic diagram of monthly spare part purchases and inventory results thereof at actual demand using the Kim T.Y. et al method.
Detailed Description
The invention will now be described in detail by way of example with reference to the accompanying drawings.
The invention aims at solving the technical problem of stock optimization of electronic product spare parts (such as batteries, cameras, mainboards and the like) in the after-sale stage, and a schematic diagram of the problem is shown in the attached figure 1. Specifically, taking a mobile phone as an example, as shown in the upper half of fig. 1, the important time nodes after the mobile phone is marketed include: the method comprises the steps of selling, daily purchase period of spare parts, stopping selling and finally purchasing from manufacturers; for a user, after buying the mobile phone, the user enters a mobile phone warranty period (set as W), and the mobile phone can seek assistance of an after-sales service provider to repair and replace spare parts when the problem that the mobile phone cannot be solved occurs in the warranty period; after a given mobile phone is sold, a user can search for a replacement part from the rear of the mobile phone and repair the mobile phone due to the failure of the mobile phone, and the like, so that the corresponding part is required. The invention helps the after-sales service side to decide the purchase amount of the given part of the electronic product after sales per month, thereby minimizing the absolute value of the difference between the inventory and the actual demand in the given time period. The electronic product selling amount per month is B (i) for a given spare part, and the demand of the specific spare part to be calculated is D (i).
It can be seen that the key technical problems to be solved by the invention include: how to select appropriate data as input data for predicting future spare part demands; how to build a data driving opportunity constraint optimization mixed integer programming model, history data is fully utilized, and decision coefficients can be effectively obtained by solving the history data; for a given N, how to control constraint intensity by adjusting the opportunity constraint threshold, the smaller the opportunity constraint threshold, the less tolerant the model is, and the inventory value in the historical data does not meet the requirement of the demand, otherwise, the optimization model can be properly relaxed, and in the application process, how to balance the calculation complexity and constraint satisfaction degree of the optimization model through the opportunity constraint threshold.
The invention provides a data-driven after-sales service electronic product part stock quantity calculating method, which specifically comprises the following steps:
In the invention, the input sample data X is W+3 dimension data, and the format is as follows:
X(i)＝{B(i-W),B(i-W+1),...,B(i-1),1,D(i-1),i}
the front W dimension data represents the sales of W months before the time point i, namely the sales data of the mobile phone in the warranty period at present; the w+1st dimension data is constant, since the purchase amount is represented by Q (i) =r T X (i) determines that dimension W+1 represents a constant in the purchase amount that is independent of other data, Q (i) is the number of spare parts purchased from the supplier in month i, and r is a decision variable; dimension w+2 is the demand of the previous month, and if i=1, the position is 0; dimension w+3 represents the number of months currently from the start of the sale.
And 2, constructing a mixed integer programming model based on a data-driven opportunity constraint method to establish a stock quantity prediction model of the electronic product part.
In order to enable business solving software to solve in a limited time, the invention constructs a mixed integer programming model of the problem based on a data-driven opportunity constraint method, and the specific construction process comprises the following steps:
step 2.1, defining intermediate variables, decision variables and other variables as shown in the following table:
and 2.2, according to the description of the problems, defining a problem optimization target as shown in a formula (1), wherein the N groups of historical data nodes are utilized to minimize the sum of backlog, backlog and spare part acquisition cost on all the data nodes, and the minimization of the value can realize fitting of spare part requirements and optimization of cost.
And 2.3, defining an inventory iteration relation on adjacent time periods, wherein the inventory of the current month is determined by accumulating the current month purchase quantity and subtracting the current month demand quantity from the previous month inventory, and the inventory is specifically shown in a formula (2), wherein S (-1) =0 is a critical condition, and the inventory quantity at the first month is only related to the current month demand and purchase quantity.
Step 2.4, based on the sample averaging method (Sample Average Approach, SAA), constructing a data-driven opportunity constraint inequality as shown in equations (4) and (5). The formula (4) defines whether each data node meets the user requirement, if the month stock quantity and the order quantity meet the requirement, the formula is satisfied when r (i) =0, and if the month stock quantity and the order quantity do not meet, the formula is satisfied when r (i) =1, therefore, the value of the intermediate variable r (i) can represent whether the month requirement is met, and r (i) =1 represents that the spare part requirement of the ith month is not met. In the formula (5), r (i) is counted up, the sum of the r (i) and the number of time periods which are not satisfied with the current month demand in the N data nodes is defined according to the opportunity constraint optimization, the frequency of the r (i) is required to be smaller than a set threshold alpha, and the formula (5) is formed.
Steps 2.5, b (i) represent the number of spare parts backlogged for the ith month period, b (i) 0 in equations (6-9) and (15) is a linear representation of b (i) =max (S (i), 0, equations (6) and (15) ensure b (i) is equal to or greater than S (i) and 0, equation (9) represents u 1 (i) And u 2 (i) At least one variable is 1, i.e., b (i) satisfies b (i). Ltoreq.S (i) or b (i). Ltoreq.0 in formulas (7) and (8), in combination with the above constraint, constitutes b (i) =S (i) or b (i) =0.
Steps 2.6, h (i) represent the number of spare parts missing for the ith month period. Similar to step 2.5, h (i). Gtoreq.0 in equation (10-13) and equation (15) is a linear representation of h (i) =max (-S (i), 0).
Step 2.7, formula (14) shows that the purchase amount of the ith month is in a linear relation with an input variable X, wherein the input variable is obtained after the data preprocessing in step 1, and r is a decision variable.
And 3, inputting the electronic product input sample set constructed in the step 1 into the electronic product part stock quantity prediction model constructed in the step 2, and solving the mixed integer programming model by utilizing a solver to obtain decision variables in the electronic product part stock quantity prediction model, so as to obtain an opportunity constraint threshold value and finish the solving of the electronic product part stock quantity prediction model.
And 4, in actual use, according to the input variable preprocessing mode in the step 1, acquiring actual historical data occurring before the predicted time period as an actual input variable, and inputting the actual input variable into the electronic product part stock quantity prediction model obtained by solving in the step 3 to obtain the electronic product part stock quantity required to be purchased in the predicted time period.
To verify the performance of the present invention, the algorithm is verified using a brand of actual production sales data. The invention adopts Python language programming, calls Gurobi 9.5.1 educational version solver, the computing environment is Intel (R) core (TM) i7-4790 [email protected]/16GB RAM, and the operating system is windows 10. Some mobile phone sales data are shown in fig. 2, wherein a blue trend line indicates the sales number of mobile phones in the month, and an orange trend line indicates the number of mobile phones in the warranty period in the market under the condition of keeping the quantity for 1 year (12 months). Based on the data, the first half data of n=30 is taken as historical data, and the method of the invention and the prediction method in the literature are respectively utilized to predict and optimize the subsequent data, wherein the opportunity constraint threshold value alpha=0.01 of the invention. The results are shown in fig. 3-4, fig. 3 is a predicted optimizing result of the invention, fig. 4 is a document predicted result, blue data nodes in the figure are actual mobile phone spare part demand, orange data nodes are predicted results for demand, and green nodes are stock conditions under actual demand by using the predicting method to make stock decisions. As can be seen from fig. 3, the stock level is kept positive under the spare part purchasing decision of the present invention, no stock shortage condition occurs, and the stock water line is not maintained at a higher level for a long time, so that the stock cost is lower. As can be seen from fig. 4, the method predicts the spare part demand trend well, but under the spare part purchase decision under the prediction method in the document, the stock is basically in a negative value, the stock is continuously out of stock, the demand of the user cannot be satisfied, and the after-sales service public praise of the brand is difficult to guarantee.
Therefore, the invention can obtain spare part purchasing strategy by establishing the opportunity constraint mixed integer programming model driven by data and solving the existing data to the maximum extent, and is suitable for the practical engineering application problem.
In summary, the above embodiments are only preferred embodiments of the present invention, and are not intended to limit the scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (4)
1. A method for calculating the inventory of a data-driven after-market electronic product component, comprising the steps of:
step 1, determining an input variable preprocessing mode, and constructing an input sample set of the electronic product;
step 2, constructing a mixed integer programming model based on a data-driven opportunity constraint method to establish an electronic product part stock quantity prediction model, wherein the method is as follows:
wherein Q (i) is the number of spare parts purchased from suppliers in the ith month, r is a decision variable, X (i) is input data of the ith month, c p Cost of acquisition for a single spare part c b Cost per unit time for backlog of individual spare parts c h For the cost per unit time of a single spare part out-of-stock, b (i) is the spare part inventory backlog for month i, h (i) is the spare part out-of-stock for month i;
step 3, inputting the electronic product input sample set into the electronic product part stock quantity prediction model, and solving the electronic product part stock quantity prediction model by using a solver to obtain a determined electronic product part stock quantity prediction model;
and 4, in actual use, according to the input variable preprocessing mode in the step 1, acquiring actual historical data occurring before the predicted time period as an actual input variable, and inputting the actual input variable into the electronic product part stock quantity prediction model obtained in the step 3 to obtain the electronic product part stock quantity required to be purchased in the predicted time period.
2. The after-market electronic product component stock amount calculating method according to claim 1, wherein the input variable preprocessing method in step 1 establishes a w+3-dimensional input variable X (i) for acquiring a sales amount of W months before a predicted period:
X(i)＝{B(i-W),B(i-W+1),...,B(i-1),1,D(i-1),i}
wherein B (i-W) is the selling amount of the electronic product in the (i-W) month, B (i-W+1) is the selling amount of the electronic product in the (i-W+1) month, B (i-1) is the selling amount of the electronic product in the (i-1) month, and D (i-1) is the demand of the specific spare part in the (i-1) month.
3. The after-market service electronic product part stock quantity calculation method of claim 1, wherein the data-driven opportunity constraint inequality constructed based on a sample averaging method in the electronic product part stock quantity prediction model is:
wherein M is a very large integer, S (i) is the stock quantity of spare parts in month i; gamma (i) is an integer variable, and when the value is 1, the requirement of the spare part in the ith month is not met, and when the value is 0, the requirement of the spare part in the ith month is met; alpha is the opportunity constraint threshold.
4. The after-market electronic product component inventory amount calculation method of claim 3, in which said opportunity constraint threshold has a value of 0.01.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211550735.2A CN116029421A (en) | 2022-12-05 | 2022-12-05 | Data-driven after-sales service electronic product part stock quantity calculating method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211550735.2A CN116029421A (en) | 2022-12-05 | 2022-12-05 | Data-driven after-sales service electronic product part stock quantity calculating method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116029421A true CN116029421A (en) | 2023-04-28 |
Family
ID=86076776
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211550735.2A Pending CN116029421A (en) | 2022-12-05 | 2022-12-05 | Data-driven after-sales service electronic product part stock quantity calculating method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116029421A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116933937A (en) * | 2023-07-27 | 2023-10-24 | 北京理工大学 | Model transfer learning-based electronic product part stock quantity prediction method |
-
2022
- 2022-12-05 CN CN202211550735.2A patent/CN116029421A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116933937A (en) * | 2023-07-27 | 2023-10-24 | 北京理工大学 | Model transfer learning-based electronic product part stock quantity prediction method |
| CN116933937B (en) * | 2023-07-27 | 2024-01-26 | 北京理工大学 | Model transfer learning-based electronic product part stock quantity prediction method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kadambala et al. | Closed loop supply chain networks: Designs for energy and time value efficiency | |
| US10872309B2 (en) | System and method for root cause analysis and early warning of inventory problems | |
| Ayvaz et al. | Stochastic reverse logistics network design for waste of electrical and electronic equipment | |
| Ramezani et al. | A new multi-objective stochastic model for a forward/reverse logistic network design with responsiveness and quality level | |
| Kim et al. | Optimal inventory control in a multi-period newsvendor problem with non-stationary demand | |
| Han et al. | Collection channel and production decisions in a closed-loop supply chain with remanufacturing cost disruption | |
| KR20210116435A (en) | Systems and methods for inventory management and optimization | |
| US20070185867A1 (en) | Statistical modeling methods for determining customer distribution by churn probability within a customer population | |
| US20140278779A1 (en) | Churn prediction and management system | |
| WO2008039457A2 (en) | Price markdown apparatus | |
| Sillanpää et al. | Forecasting replenishment orders in retail: value of modelling low and intermittent consumer demand with distributions | |
| KR101642216B1 (en) | Method and apparatus for analyzing missing not at random data and recommendation system using the same | |
| US20210390498A1 (en) | Systems and methods for inventory management and optimization | |
| Guo et al. | Designing a sustainable-remanufacturing closed-loop supply chain under hybrid uncertainty: Cross-efficiency sorting multi-objective optimization | |
| CN111292149A (en) | Method and device for generating return processing information | |
| US11687875B2 (en) | Distribution-independent inventory approach under multiple service level targets | |
| CN116029421A (en) | Data-driven after-sales service electronic product part stock quantity calculating method | |
| Baghizadeh et al. | Closed‐loop supply chain design with sustainability aspects and network resilience under uncertainty: Modelling and application | |
| CN113469595A (en) | Intelligent supply chain system and server platform | |
| Zou et al. | Pricing‐Decision and Coordination Contract considering Product Design and Quality of Recovery Product in a Closed‐Loop Supply Chain | |
| Ghezavati et al. | Development of an optimization model for product returns using genetic algorithms and simulated annealing | |
| EP1811445A1 (en) | Churn prediction and management system | |
| Ding et al. | Revealing the hidden potentials of Internet of Things (IoT)-An integrated approach using agent-based modelling and system dynamics to assess sustainable supply chain performance | |
| US20160283897A1 (en) | Days of inventory determination based on constraints | |
| Farshbaf-Geranmayeh et al. | Closed loop supply chain network design under uncertain price-sensitive demand and return |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |