CN115375216A

CN115375216A - Method and device for determining safety stock

Info

Publication number: CN115375216A
Application number: CN202110559931.5A
Authority: CN
Inventors: 王寅骅; 王子卓; 杨超林; 袁媛
Original assignee: Huawei Technologies Co Ltd
Current assignee: Huawei Technologies Co Ltd
Priority date: 2021-05-21
Filing date: 2021-05-21
Publication date: 2022-11-22

Abstract

The embodiment of the application provides a method and a device for determining safety stock. The method comprises the following steps: determining a plurality of nodes for manufacturing and/or assembling the commodity according to the bill of materials of the commodity, wherein the commodity comprises a finished product or a semi-finished product; determining one or more key nodes needing to plan safety stock and the safety stock of each key node based on the service time of each node and the lead period of the upstream node; the lead period of one node comprises the service time of the node, the manufacturing lead period of the node and the supply lead period. Due to the comprehensive consideration of the relevance of each node in the bill of materials, the service time of each node and the lead period of the upstream node are combined to determine the nodes needing to plan the safety stock and the respective safety stock. Thereby optimizing the safety stock layout of the whole production chain.

Description

Method and device for determining safety stock

Technical Field

The present application relates to the field of inventory management technologies, and more particularly, to a method and apparatus for determining safety inventory.

Background

In the supply chain, inventory management is a very important part, and the reasonable inventory can meet the production requirement and reduce the inventory cost.

At present, the safety stock strategy does not consider the interrelation among a plurality of nodes in a manufacturing network, considers the requirement relationship of each node as independent, and calculates the safety stock of all nodes, and in the actual production process, the safety stock strategy can not carry out global optimization on the safety stock layout of the whole production chain, which can cause the redundancy or shortage of the safety stock.

Therefore, global optimization of the safety stock layout of the whole production chain is a problem which needs to be solved urgently.

Disclosure of Invention

The application provides a method and a device for determining safety stock, which can determine the safety stock based on the service time and the lead time of a node, so that the safety stock layout of the whole production chain is optimized.

In a first aspect, the present application provides a method of determining a safety stock, the method comprising: determining a plurality of nodes for manufacturing and/or assembling a commodity based on a bill of materials of the commodity, wherein the commodity comprises a finished product or a semi-finished product; determining one or more key nodes needing to plan safety stock and the safety stock of each key node based on the service time of each node in the plurality of nodes and the lead period of the upstream node; wherein the lead period of the upstream node comprises a service time of the upstream node, a manufacturing lead period and a supply lead period of the upstream node; the service time represents a response time to a demand of a downstream node.

Based on the technical scheme, after a plurality of nodes for manufacturing and/or assembling commodities are determined, the key nodes needing to be made into safety stock and the safety stock of the key nodes are determined according to the service time of each node and the lead period of the upstream node, so that the relevance of each node of the commodities is comprehensively considered, the key nodes needing to plan the safety stock are determined by comprehensively considering the service time of each node and the lead period of the upstream node in combination with the relation between the upstream node and the downstream node, and the safety stock of each key node is further determined. Due to the fact that the relation between the upstream node and the downstream node, service time, lead period and other factors are combined, the time range of demand fluctuation is covered by the key nodes and the safety stock, and therefore the safety stock layout of the whole production chain is optimized.

With reference to the first aspect, in some possible implementations of the first aspect, determining, based on a service time of each node in the plurality of nodes and a lead period of an upstream node thereof, one or more key nodes that need to plan a safety inventory, and a safety inventory of each key node, includes: calculating the total goods holding cost of the plurality of nodes by taking the service time of each node in the plurality of nodes and the lead period of the upstream node as variables; determining one or more key nodes needing to plan safety stock and the safety stock of each key node by taking the minimum total goods holding cost of the nodes as a target; wherein the node with the safety stock close to zero is not taken as a key node.

In the process, the service time of each node in the plurality of nodes and the lead period of the upstream node of the plurality of nodes are taken as variables, the total goods holding cost of the plurality of nodes is calculated, then one or more key nodes needing to plan safety stock and the safety stock of each key node are determined by taking the minimum total goods holding cost of the plurality of nodes as a target, the method for calculating the safety stock of the key nodes is embodied, and the optimization of the safety stock layout of the whole production chain is realized.

With reference to the first aspect, in some possible implementations of the first aspect, the method further includes: predicting the change trend of the demand of the commodity in a future period based on the historical sales data of the commodity; and determining that safety stock needs to be planned for the commodity according to the change trend of the demand of the commodity in a future period.

In the process, the change trend of the demand of the commodity in the future time period is predicted according to the historical sales data, and whether the safety stock is planned for the commodity is determined according to the change trend, so that some commodities which do not need to be made into the safety stock can be filtered, unnecessary safety stock is avoided on one hand, the calculation amount is reduced for the subsequent safety stock planning step on the other hand, and the efficiency is improved.

With reference to the first aspect, in some possible implementations of the first aspect, predicting a trend of change in demand of the commodity over a future period based on historical sales data of the commodity includes: predicting statistical characteristics of demand of the commodity in a plurality of time windows in the future based on historical sales data of the commodity, wherein the time windows are time windows in a future period; based on the statistical characteristics of the demand in the plurality of time windows, a trend of change in demand of the commodity in the future period is determined.

In the process, according to the historical sales data of the commodities, the statistical characteristics of the demand of the commodities in a plurality of future time windows are predicted, and then according to the statistical characteristics, the variation trend of the demands of the commodities in a future time period is determined, so that the variation trend of the demands is similar to the variation trend of the statistical characteristics of the demand.

Optionally, the statistical characteristics include: the demand quantity is indicated by an expected value and a standard deviation, and the standard deviation is used for indicating the fluctuation rate of the demand quantity.

The demand quantity of the commodity is indicated by the expected value, the fluctuation rate of the demand quantity is indicated by the standard deviation, the larger the standard deviation is, the larger the fluctuation rate can be represented, and the smaller the standard deviation is, the smaller the fluctuation rate can be represented. Therefore, commodities which do not need to be stored in safety can be screened out according to the change of the statistical characteristics, and the calculation amount is reduced.

With reference to the first aspect, in some possible implementations of the first aspect, predicting a statistical characteristic of demand of the commodity in a plurality of time windows in the future based on historical sales data of the commodity includes: extracting the characteristics of the commodities based on historical sales data of the commodities; constructing a statistical model based on the characteristics of the commodity; and predicting the statistical characteristics of the demand of the commodity in a plurality of time windows in the future through the statistical model.

In the process, the characteristics of the commodity can be extracted according to the historical sales data of the commodity, and then the corresponding statistical model is constructed based on the characteristics of the commodity, so that the constructed statistical model can more accurately predict the statistical characteristics of the demand of the commodity in a plurality of future time windows, and the fitting degree of the predicted result and the actual result is improved.

Optionally, the above features include one or more of: order timing features, commodity attribute features, bill of material features, event features, plan features, business experience features, and external factor features.

Based on one or more of the characteristics, a corresponding statistical model can be constructed, the statistical characteristics of the demand can be predicted, a plurality of extractable characteristics are provided in the process, a plurality of factors influencing the demand are considered, and the method is comprehensive.

With reference to the first aspect, in some possible implementations of the first aspect, the method further includes: the safety inventory of each of the one or more critical nodes is evaluated to determine a demand satisfaction rate for each critical node.

In the process, the safety stock of each key node can be evaluated through simulation to determine the demand satisfaction rate of each key node, so that the effectiveness of the safety stock plan is verified.

With reference to the first aspect, in some possible implementations of the first aspect, the method further includes: and adjusting the safety stock of the key nodes which do not reach the expected index, wherein the key nodes which do not reach the expected index are the nodes of which the demand satisfaction rate does not reach the preset target demand satisfaction rate.

In the process, the safety stock of the key nodes which do not reach the expected index can be adjusted, so that an optimal safety stock plan is obtained, and the safety stock layout of the whole production chain is optimized.

In a second aspect, the present application provides an apparatus for determining a safety stock, comprising means or elements for implementing the method of the first aspect as well as any one of its possible implementations. It should be understood that the respective modules or units may implement the respective functions by executing the computer program.

In a third aspect, the present application provides an apparatus for determining a safety stock, including a processor, configured to execute the method for determining a safety stock described in the first aspect and any one of the possible implementation manners of the first aspect.

The apparatus may also include a memory to store instructions and data. The memory is coupled to the processor, which when executing instructions stored in the memory, may implement the methods described in the above aspects. The apparatus may also include a communication interface for the apparatus to communicate with other devices, which may be, for example, a transceiver, circuit, bus, module, or other type of communication interface.

In a fourth aspect, the present application provides a chip system comprising at least one processor configured to support the implementation of the functionality referred to in the first aspect as well as any one of the possible implementations of the first aspect, for example, to receive or process data and/or information referred to in the method as described above.

In one possible design, the system-on-chip further includes a memory to hold program instructions and data, the memory being located within the processor or external to the processor. The chip system may be formed by a chip, and may also include a chip and other discrete devices.

In a fifth aspect, the present application provides a computer-readable storage medium comprising a computer program which, when run on a computer, causes the computer to carry out the method of the first aspect as well as any one of the possible implementations of the first aspect.

In a sixth aspect, the present application provides a computer program product comprising: a computer program (which may also be referred to as code, or instructions), which when executed, causes a computer to perform the first aspect and the method of any possible implementation of the first aspect.

It should be understood that the second aspect to the sixth aspect of the present application correspond to the technical solutions of the first aspect of the present application, and the beneficial effects achieved by the aspects and the corresponding possible implementations are similar and will not be described again.

Drawings

FIG. 1 is a schematic diagram of a bill of material network;

FIG. 2 is a schematic diagram of a supply chain architecture suitable for use in a method of determining safety inventory provided by an embodiment of the present application;

FIG. 3 is a schematic diagram of a plurality of modules for performing a method of determining a safety inventory as provided by an embodiment of the present application;

FIG. 4 is a schematic flow chart diagram of a method for determining safety stock provided by an embodiment of the application;

FIG. 5 is a schematic diagram of a model used in a regression algorithm provided in an embodiment of the present application;

fig. 6 is a schematic block diagram of an apparatus for determining safety stock provided by an embodiment of the present application.

Detailed Description

The technical solution in the present application will be described below with reference to the accompanying drawings.

Before describing embodiments of the present application, first, a brief description of terms involved in the present application will be given.

1. Safety stock: also known as safe storage, also known as insurance inventory. Refers to the amount of insurance reserve that is expected to prevent uncertainty factors. The uncertainty factor may be, for example, a large number of sudden orders, a sudden delay in delivery date, an increase in temporary usage, a delivery error date, etc. Safety stock is not used under normal conditions, but only when stock is used excessively or delivery is delayed.

2. Bill of materials (BOM): refers to a technical document describing the composition of an enterprise product. For example, a BOM may indicate the structural relationship between the assembly, sub-assembly, component, part, through raw materials of a product, as well as the quantity required.

The BOM may be represented by a BOM network. FIG. 1 is a schematic diagram of a BOM network. A plurality of black dots are shown in fig. 1, and a straight line connecting the plurality of black dots. Each black dot may represent a part, a component, an assembly, or a subassembly, and for ease of description, will be referred to collectively hereinafter as a node. The nodes which are upstream and downstream can be connected through a line, and a plurality of nodes are connected through a line to form a bundle of nodes or a cluster of nodes. This bundle of nodes may correspond to an end product. Such as a, B, C identified in the figure. It should be understood that the commodity described herein may specifically refer to a product sold by an enterprise. The goods may be finished products such as mobile phones, tablet computers, and the like; but also semi-finished products such as batteries, screens, etc.

As can be seen from fig. 1, each commodity includes a plurality of nodes, that is, a large number of special parts and general parts are required for manufacturing each commodity. For example, the node a is a node in the article a, which means that the node a is a node for manufacturing the article a, for example, a special-purpose piece for manufacturing the article a. The nodes B and d are each a node in the products a, B and C, indicating that the nodes B and d are each a node for manufacturing the products a, B and C, for example, a common part. Node c may be used to manufacture or assemble nodes b and d, or may be a common component.

In the BOM network shown in fig. 1, the connection lines between the nodes show the upstream and downstream relationship between the nodes. For example, node c is an upstream node of nodes b and d; nodes b and d are downstream nodes of node c. It can be seen that there may be one or more upstream nodes, and/or one or more downstream nodes per node. Where an upstream node may be used to produce or assemble a downstream node, the generation and assembly of the downstream node is dependent on the upstream node's manufacturing cycle, supply cycle, and service time. The upstream node may also be referred to as a parent node and the downstream node may also be referred to as a child node. It should be understood that upstream and downstream relationships before a node may also be referred to as hierarchical relationships.

In the BOM network, each node can be embodied as a code, the connection relation among the nodes can represent the upstream and downstream relation among products, the side length among the nodes can represent the lead period, and the characteristic information of the nodes comprises service time. It should be understood that the BOM network is used to describe each node in the BOM used to manufacture the goods and the hierarchical relationships between each node, and thus may also be referred to as a manufacturing network.

3. Decoupling points: means that a non-correlation is established between the supply and the use of the material. That is, a safety stock is established between the preceding and succeeding processes so that fluctuations in supply do not limit the productivity or utilization of the next process. In the embodiment of the present application, the inventory decoupling point may be referred to as a key node, that is, a node where safety inventory needs to be planned.

4. Service time (service time): in a bill of material network, the service time of a node refers to the response time of the node to the demand of its downstream nodes.

5. The advance period: the lead period for a node may include the service time for the node, and the time required for manufacturing and shipping of downstream replenishment from the node. The time required for manufacturing may be referred to as a manufacturing lead period, and the time required for transportation may be referred to as a supply lead period. For example, if the service time of a node is S, the time required for manufacturing and transportation of a downstream node for replenishment from the node is T, and the lead time of the node is S + T. That is, the time to be advanced for replenishing from the node by the downstream node is S + T, i.e., the replenishment advance period of the downstream node of the node is S + T.

It should be understood that the longer the service time of a node is, the longer the replenishment lead period of the downstream node is, and the greater the risk that the downstream node needs to bear on the node due to uncertain demand. That is, the longer a node is served, the less security inventory is required for that node and the more security inventory is required downstream.

Currently, the security inventory policy does not consider the interrelationship between nodes in the BOM network, that is, each node is regarded as an independent node, and the requirement of each node is regarded as an independent requirement. In the specific implementation, the future demand of each node is predicted according to a traditional demand prediction model; then, performing safety stock calculation of a single node, for example, calculating the safety stock of each node based on normal distribution and a traditional safety stock calculation model; finally, the security inventory policy is periodically updated. For example, the security inventory policy is updated on a monthly basis, i.e., the security inventory is calculated once a month.

However, the above method for determining safety stock lacks flexibility, and is difficult to adapt to complex scenes such as multi-level stocks, for example, when a certain final product has multi-level nodes, it is difficult to simply calculate safety stock for each node.

Therefore, the application provides a method for determining safety stock, which considers the BOM network of each commodity as a whole, takes the BOM network as a decision basis, comprehensively considers the relation among nodes of each level, and takes the service time, the manufacturing lead time and the supply lead time of each node as decision variables to calculate the safety stock of each node in the whole production chain, so that the safety stock layout in the whole production chain is optimized.

The method for determining safety stock provided by the embodiment of the application will be described in detail below with reference to the accompanying drawings.

For the understanding of the embodiment of the present application, first, a brief description will be made of a supply chain architecture suitable for the method of the embodiment of the present application with reference to fig. 2. Fig. 2 is a schematic diagram of a supply chain architecture suitable for use in the method for determining safety stock provided by the embodiment of the present application.

As shown in FIG. 2, the supply chain architecture includes a sales department, a main planning department, a production department, and a procurement department. The sales department can carry out sales prediction on the codes, the main plan department converts the sales prediction of each code into main plan demand prediction on all the codes based on factors such as sales prediction, capacity and the like, the production department further considers various constraints such as materials, substitution, priority and the like, optimizes a future production plan, and outputs a future purchasing plan to the purchasing department according to the plan. The master planning department may also calculate the safety stock for each code based on sales forecasts for each code in the future to achieve safety stock optimization. The master planning department may provide each coded safety stock to the production department as a reference level. The production department can compare the reference level of the safety stock with the production plan, thereby ensuring that the future end-of-term stock reaches the reference level of the safety stock.

It should be appreciated that the above-described safety inventory optimization may be implemented by performing the method of determining safety inventory provided by embodiments of the present application. Fig. 3 schematically illustrates a plurality of modules for performing the method for determining a safety stock provided by the embodiment of the present application. The multiple modules may be deployed, for example, in a computing device to implement different functions. The modules may be integrated in the same computing device or may be distributed across multiple computing devices. The computing device may be, for example, a server. The embodiments of the present application do not limit this. For convenience of description, the device for executing the method for determining the safety stock provided by the embodiment of the present application is simply referred to as the device for determining the safety stock hereinafter.

As shown in fig. 3, the computing device includes: the system comprises a demand forecasting module, an inventory strategy module, an inventory planning module, an inventory simulation module and an inventory tuning module.

The demand forecasting module is used for forecasting the statistical characteristics of the demand of various commodities. The statistical characteristics of the demand may include a weighted average, an upper bound, and a standard deviation. The weighted average is also a desired value of the demand for the future period. Illustratively, in the future month, if the probability that the demand of a certain commodity is a is m%, the probability that the demand is b is n%, the probability that the demand is c is p%, and the probability that the demand is d is q%, the expected value w = a × m% + b × n% + c × p% + d × q% of the demand of the commodity in the future month. The maximum of the demand amounts a, b, c, d may serve as an upper bound for the demand of the good in the next month. The standard deviation can be determined according to the above-mentioned demand amounts a, b, c, d and the average value, and specifically, the calculation formula of the standard deviation in the prior art can be referred to, and is not described in detail here.

The inventory policy module may combine product characteristics, inventory characteristics, supply characteristics, business requirements, and the like to pre-process the nodes required by each commodity, for example, filter some nodes; as another example, the nodes are classified. The product characteristics comprise the demand of the product, the manufacturing period, the supply period of raw materials, the easiness in loss, the easiness in expiration and the like. The inventory characteristics may include fluctuating characteristics of the inventory, such as, for example, historical data, where a node is likely to be redundant, and less safety inventory may be made. The business requirements may include, for example, but are not limited to, less safety inventories for nodes with high cargo values and high fluctuations, more safety inventories for nodes with low cargo values and low fluctuations, and so forth.

The inventory planning module may determine key nodes and their safety inventories based on the service times of each node and may combine inventory parameters, cost parameters, generation constraints, inventory constraints, target service levels, and the like. Wherein the cost parameter is used for describing the cargo value of the node. The inventory parameter refers to how much safety inventory is normally maintained. For example, a usage for one week may be maintained as a safety stock. It should be understood that the cost parameters and inventory parameters may be obtained based on historical data. The target service level is a probability describing the actual demand that can be covered. For example, the expected value of the future demand is w, the upper bound of the demand is w1, and the actual demand is k% which indicates that the possibility that the demand is less than w1 is greater than or equal to k%.

The inventory simulation module may evaluate the safety inventory of each key node output by the inventory planning module in conjunction with business objectives and constraints. The inventory tuning module may evaluate the simulation results to determine whether expected metrics are achieved. And in the case that the expected index is not reached, the safety stock of the key node output by the stock planning module can be further optimized.

The above-mentioned safety-inventory optimization module includes, but is not limited to, the following application scenarios, and several application scenarios of the safety-inventory optimization module are described in detail below.

Scene one: and comprehensively considering the ex-warehouse prediction.

The error is amplified by decomposing downwards according to the upper node, and the upper bound of the safety stock of the raw material has larger deviation because the lead period is not a definite time. In order to make the overall security inventory strategy error smaller, an upper bound can be set for the requirement of each node, and the security inventory is made under the condition that the requirement does not exceed the upper bound.

Scene two: the cost and the fluctuation ratio are considered.

In the business, the experience and habit of setting the safety stock according to the historical ex-warehouse quantity, and in order to control the cost, different upper bounds can be set according to the material cost and the fluctuation rate, namely, the upper bounds of the safety stock of the nodes are manually set. For example, a safety stock can be established less for nodes with high material cost and high fluctuation, a lower point can be set at the upper boundary of the safety stock, and a higher point can be set at the upper boundary of the safety stock for nodes with low material cost and low fluctuation.

Scene three: consider a coded Service Level Agreement (SLA).

The raw materials and the middleware need to meet the requirements of the external orders and the parent items, but the SLA commitment is only effective to the external orders, and if the constraint that S is less than or equal to SLA is directly added to the items, the commitment service time of the items to the parent items can be too short. It should be understood that an SLA commitment that is valid only for an external order means that for the finished product there is a SLA commitment that is customer-oriented, and for the middleware there is no SLA commitment.

If a code can be directly sold as an external order or used as middleware to manufacture other finished products, a virtual node can be set to represent the requirement of the external order, an SLA commitment is added to the virtual node, the prediction of the requirement of the external order is transferred to the virtual node, and the virtual node and the original node have the same fluctuation rate. It should be appreciated that the SLA commitment can be a constraint in computing the security inventory.

Scene four: the code shelf life is considered.

The manufacturing lead period and the provisioning lead period for all nodes are considered. For points with shorter shelf life and longer guaranteed service time, security inventory is not set, and for nodes with longer shelf life and shorter guaranteed service time, security inventory levels are increased.

It should be understood that the specific functions of the modules in fig. 3 will be described in detail below in conjunction with the method embodiment in fig. 4, and are not described in detail here for brevity.

The method for determining the safety stock provided by the embodiment of the application will be described in detail with reference to fig. 4. The embodiment shown in fig. 4 describes the steps by taking the modules shown in fig. 3 as an example. However, this should not be construed as limiting the embodiments of the present application in any way. The various modules shown in fig. 3 are merely partitioned based on different functions. These modules may or may not be physically separate.

Fig. 4 is a schematic flow chart of a method 400 for determining a safety stock according to an embodiment of the present application. As shown in fig. 4, the method 400 may include steps 410 through 460.

In step 410, the demand prediction module predicts a statistical characterization of demand for a plurality of items over a plurality of time windows in the future based on historical sales data for the plurality of items.

Specifically, the prediction of the demand for the commodity may be implemented using a regression algorithm. Illustratively, the characteristics of each commodity are extracted from historical sales data, rules in the characteristics are summarized, modeling is further performed through a regression algorithm, a prediction model is obtained, and then the demand of the commodity in any future period is predicted through the prediction model. The historical sales data may include, but is not limited to, historical order data, historical ex-warehouse data, and the like.

In the embodiment of the present application, features of multiple dimensions of various commodities can be extracted based on historical sales data, and for example, the following dimensions can be included but not limited to:

1. an order timing feature. The order time sequence feature refers to a feature extracted from a historical demand sequence, wherein the historical demand sequence can be a sequence obtained by arranging historical sales data of the goods according to a time sequence. The historical demand sequences may be categorized according to different characteristics. For example, the feature types may be classified into a hysteresis feature, a time feature, a scroll window feature, and the like. Wherein the lag characteristic indicates that there is a lag in time to sell relative to time to order. For example, if a certain commodity is expected to be sold in the spring festival, it is required to order in advance. The time feature may include a feature plotted in terms of an absolute position and a relative position of a quarter, a month, and a week, and the rolling window feature is similar to the hysteresis feature, in which, considering that there is a time difference between the order time and the arrival time, after the arrival time is needed for sales, there is a period needed for ordering goods to sell after the arrival time, for example, the period is one month, a certain goods is ordered in 3 months, sold in 4 months, ordered in 4 months, and sold in 5 months. In addition, according to different characteristic sources, the time sequence characteristics of the commodities and the time sequence characteristics of the related commodities can be further divided. The article's own timing characteristics and the associated article timing characteristics may each include the above-described after characteristics, time characteristics, rolling window characteristics, and the like.

2. And (4) commodity attribute characteristics. The commodity attribute feature can be obtained from the attribute of the commodity. For example, the commodity attribute characteristics may include, lifecycle status, material cost, initial inventory, product family, commodity description. The product family refers to the kind of the commodity, and for example, the product family may be food, living goods, and the like. The commodity description may be performed by extracting a keyword using Natural Language Processing (NLP) to describe the commodity. It should be understood that the above features are only a few examples of the product attribute features provided in the embodiments of the present application, and that the product attribute features may also include other features derived from the product's own attributes, including but not limited to the present application.

3. And (4) BOM characteristics. BOM features may refer to information about a good in a BOM network. For example, a hierarchical relationship between nodes, that is, an upstream-downstream relationship between nodes may be included. The BOM characteristics may also include structural characteristics of a hierarchy in which each node is located, that is, structural characteristics of nodes included in the commodity in the BOM network, for example, structures of a node b and a node c required for producing a certain commodity, a node d required for generating a node b, and nodes downstream of the node d.

4. An event feature. Event characterization may refer to separating historical repetitive events and analyzing the summarized corresponding features.

5. A planning feature. The plan features may refer to features extracted from a sales plan or a procurement plan.

6. And (4) business experience characteristics. The business experience characteristics may refer to characteristics of a rule formula formed by converting business experiences related to the demand.

7. Characteristics of external factors. The external factor feature refers to a feature extracted from an external factor. For example, features extracted from weather. It should be understood that the need to extract the externality features may be contingent on the circumstances.

As can be seen from the above listed features, the feature dimensions extracted for various commodities in the embodiment of the present application are high, and a feature selection algorithm may be used to perform dimension reduction. On the other hand, in order to reduce the risk of overfitting of the prediction model, the commodities can be grouped by adopting a clustering algorithm, and the prediction models of the same group of commodities can adopt the same analytic expression. Specifically, the characteristics of the commodities can be gathered in the middle by adopting a clustering algorithm, but are not scattered, so that the typical characteristics of the commodities are conveniently found, the commodities with similar characteristics are divided into a group, the characteristics of the same group of commodities can be used as a shared training sample, the risk of overfitting of a prediction model is reduced, and meanwhile, the loss of rules can be avoided.

By way of example and not limitation, the models employed by the regression algorithm may include time series models, machine learning models, and deep learning models.

Fig. 5 is a schematic diagram of a model used in a regression algorithm provided in an embodiment of the present application.

As shown in fig. 5, the time series model may include, but is not limited to, the following categories: cubic exponential smoothing (e.g., holt-hots), autoregressive integrated smoothing averaging (ARIMA), and exponential smoothing (exponential smoothing).

Machine learning models may include, but are not limited to, the following classes: gradient boosting decision tree (SVM), extreme gradient boosting (Xgboost), random forest (random forest), support Vector Machine (SVM).

Deep learning models may include, but are not limited to, the following classes: recurrent Neural Network (RNN), gated Recurrent Unit (GRU), antagonistic adaptive network (GAN), long-short term memory (LSTM).

It should be understood that the demand of a certain product in a certain period of time in the future is predicted through a regression algorithm, the regression prediction can be performed through any one of the models, and a weighted model with multiple models fused can also be used for performing the regression prediction. The embodiments of the present application do not limit this.

Taking a random forest as an example, the demand in the future period of time for solving the commodity is simply explained.

The random forest algorithm is a parallel decision tree algorithm, and a final predicted value is obtained by constructing m trees and averaging the m results. The prediction formula is: final predicted value = (first tree predicted value + second tree predicted value + \8230; + mth tree predicted value)/m. For example: the predicted value of the first tree is 100, the second tree is 80, the third tree is 60, and the final predicted value = (100 +80+ 60)/3 =80.

It should be understood that reference is made to the prior art for relevant descriptions of random forests and other various models. For the sake of brevity, no detailed description is provided herein.

After the prediction model is established, historical order data and historical ex-warehouse data of each commodity can be input into the prediction model, and the prediction model can output the statistical characteristics of the demand of each commodity in the future period. The statistical features may include a weighted average, an upper bound, and a standard deviation. The description of the statistical characteristics has been detailed above, and for brevity, will not be described again here.

Based on the above described method, the predictive model may predict statistical characteristics of demand for each commodity over multiple time windows in the future. For example, the statistical characteristics of the demand of each commodity in N time windows from time window 1 to time window N (N > 1 and is an integer) are predicted. The multiple time windows may be multiple continuous time windows, or multiple time windows divided according to a time sequence. For example, the plurality of time windows may be a plurality of consecutive months, such as a plurality of months, e.g., 1 month, 2 months, 3 months, 4 months, etc., in units of months; or a plurality of discontinuous months, for example, a plurality of months such as 1 month, 3 months, 5 months, 7 months, etc.; there may also be multiple time windows that slide, such as 1 month 1 day to 1 month 31 days, 1 month 16 days to 2 months 15 days, 2 months 1 day to 2 months 28 days, 2 months 16 days to 3 months 15 days, 3 months 1 day to 3 months 31 days, and so on. For the sake of brevity, this is not further enumerated here. Since the plurality of time windows are arranged in time series, they may also be referred to as a time window sequence.

It should be understood that the plurality of time windows may be a plurality of time windows belonging to a predetermined period of time in the future. The preset time period can be used for observing the change trend of the demands of various commodities.

In step 420, the items for which safety stock planning is required are determined based on the statistical characteristics of the demand of each item over a plurality of time windows.

According to the statistical characteristics of the demand of each commodity in a plurality of time windows, the variation trend of the demand of each commodity in a future time period can be obtained, and whether the safety stock needs to be planned for the commodity can be determined based on the variation trend of the demand of each commodity in the future time period.

For example, based on the prediction of the demand amount of each product in a plurality of time windows, the trend of change in the demand amount of each product can be obtained. For example, if the weighted average of a certain commodity is higher and higher in a plurality of time windows, the demand of the commodity shows an ascending trend; if the demand of the commodity is in a descending trend, the average value is lower and lower. It should also be understood that the standard deviation may represent the fluctuation rate of the demand. The larger the standard deviation, the larger the fluctuation rate can be represented, and the smaller the standard deviation, the smaller the fluctuation rate can be represented.

Based on the statistical characteristics of the demand of each commodity predicted in a plurality of time windows, it can be determined which commodities have an ascending demand, which commodities have a descending demand, which commodities have an unchanged demand, which commodities have a larger demand fluctuation, and which commodities have a smaller demand fluctuation. For the commodities with the increasing demand, safety stock planning is needed, and especially for the commodities with larger demand fluctuation, more safety stock can be made; for the commodity with smaller demand fluctuation quantity, less safety stock can be made. The safety stock may not be made for the commodity whose demand is in a descending trend or unchanged, or the safety stock may be made for a small amount of commodity whose demand is in a descending trend and whose demand fluctuation is large, and the safety stock may not be made for commodity whose demand is in an ascending trend and whose demand fluctuation is small.

Here, the safety stock of the commodity means that each node for manufacturing and/or assembling the commodity is subjected to safety stock. For example, in the BOM network shown in fig. 1, if the product a is to be safely stocked, the nodes (e.g., the nodes a, b, c, and d shown in the figure) for forming the product a may be considered to be safely stocked.

The method for determining the safety stock provided by the embodiment of the present application will be described below by taking a node as an example. The node may be a node for manufacturing or assembling one or more goods. For convenience of description, the method of determining the security inventory of each node is described below in terms of one of the above-mentioned commodities. The article is designated, for example, as a first article, which may be manufactured and/or assembled from a plurality of nodes. The first product may be, for example, a product that needs to be put into safety stock determined in

steps

410 and 420, or a product that needs to be put into safety stock that is set manually. The embodiment of the present application does not limit this.

Step 430, the inventory policy module preprocesses the plurality of nodes that make up the first commodity.

And the pretreatment is carried out on a plurality of nodes, so that the calculation amount of subsequent steps can be reduced. Thus, preprocessing multiple nodes may be an optional step. The means for determining a safety stock may also perform the subsequent steps directly without performing step 430. In other words, the inventory policy module is an optional module. The demand volume and statistical characteristics of the various items output by the demand forecasting module may also be directly input to the inventory planning module.

One possible implementation is that the inventory policy module may pre-process the plurality of nodes based on product characteristics, inventory characteristics, supply characteristics, business requirements, and the like.

Yet another possible implementation manner is that the entry and exit degree of each node can be obtained by analyzing the BOM network structure of the first commodity. The entry and exit degree of a node may specifically refer to the number of upstream nodes, the number of downstream nodes, and the pairing relationship of the node. And filtering out nodes which do not need to be subjected to safety stock according to the access degree of each node, the historical ex-warehouse quantity and the demand of downstream nodes of each node, the historical stock of each node and the lead period of each node.

Optionally, preprocessing the plurality of nodes may include: the plurality of nodes are filtered to filter out nodes that do not need to be kept in safety stock. For example, there may be some nodes in the tail-out phase that do not need to do safety stock; some nodes are decided not to be used any more, and safety stock does not need to be made; there are also nodes where scheduling inventory may be low, and therefore, no safety inventory is required. Nodes that do not need to do a security inventory are not limited to the few cases listed above, and are not listed here for brevity.

Additionally, as previously described, in a BOM network, each commodity includes, i.e., is manufactured and/or assembled from, a plurality of nodes. However, some of the nodes have high cargo value and high fluctuation, and can be stored less because of high cost and large fluctuation; some nodes are low in goods value and fluctuation, and can be stored more due to low cost and small fluctuation; some nodes are high in goods value and low in fluctuation, and can be stored in a small amount due to high cost and small fluctuation; still others are low in freight value and highly fluctuating, and can be stocked in small quantities due to large fluctuations but at low cost. Thus, preprocessing the plurality of nodes may also include classifying the plurality of nodes.

Optionally, the preprocessing the plurality of nodes may further include: the nodes are classified, and the upper and/or lower bounds of the security inventory are set for the nodes of different classifications. Illustratively, the nodes may be classified for different characteristics, and different upper bounds of the security inventory may be set for each class of node. For example, high freight value and high fluctuation nodes are classified as one class, and can be kept in stock as little as possible, for example, the upper bound is a1; low freight value and low fluctuation are classified into one category, and more stocks can be made, for example, the upper bound is a2; high cargo values and low fluctuations are one category, and a small amount of inventory can be made, for example, the upper bound is a3; low cargo value high fluctuation is a category, and a small amount of stock can be made, for example, with an upper bound of a4. Wherein a1, a2, a3 and a4 are natural numbers, a1 is the minimum, a2 is the maximum, and a3 and a4 are in the middle.

In step 440, the key nodes and the safety stock of each key node are determined based on the service time of each node and the lead period of its upstream node.

The key node may specifically be a node that needs to plan a safety stock. The service time, manufacturing lead period, and provisioning lead period for a node may be referred to as the lead period for the node. Thus, the advance period of the upstream node includes: a service time, a manufacturing lead period, and a provisioning lead period for the upstream node. For any node, based on the manufacturing lead time and the supply lead time of replenishment from an upstream node, the service time of the upstream node and the service time of the node, the time range of fluctuation of the demand covered by the node can be obtained, and further the safety stock required by the node can be obtained by combining the fluctuation of the demand.

It can be understood that the longer the lead period of the upstream node is, the more the safety stock of the node is; the longer a node is served, the less security inventory the node may need. The longer the lead period of the upstream node of any one node is, the more safety stock the node needs.

It can be seen that the security inventory of each node may be a function of its own service time, the lead period of the upstream node. If with each nodeThe service time and the lead period of the upstream node are decision variables, and a function f (x) can be constructed _i ,y _i ). Wherein x is _i Represents the service time of a node I (I is more than or equal to 1 and less than or equal to I, I is a natural number) in the I (I is a natural number) nodes, y _i Indicating the lead period of node i.

If the total stock cost of the safety stock is minimized, the key nodes and the safety stock of each key node can be solved. For example, assume that node i has a cost p _i ，p _i The total cost of shipment for > 0,I nodes can be expressed as:

the total cost of goods holding is minimized, so that the decoupling point of the safety stock, namely the key node, and the safety stock of each key node can be solved.

It should be understood that the safety stock of each node is not a simple function of service time, lead time, etc., and other factors may affect the safety stock. Such as the connection relationships in the BOM structure, the supply of nodes, the raw materials of the product, etc.

Further, some constraints may be introduced in solving the safety stock of each key node. For example, the constraint may include setting the target service level such that the target service level of each node reaches k%. For another example, for some nodes, an upper bound and/or a lower bound are introduced, so that the total goods-holding cost is minimized under the condition that the security library of each key node has constraints.

In step 450, the inventory simulation module evaluates the safety inventory of each critical node.

In the inventory simulation module, the same time window as described in step 410 may be considered. The inventory simulation module can be used for restoring as many business rules as possible in the original production flow to evaluate the safety inventory plan.

One possible implementation is to input the demand of the goods output by the demand forecasting module in a plurality of time windows in the future into the inventory simulation module, and the inventory simulation module randomly generates the actual demand of each node. By way of example and not limitation, the inventory simulation module may randomly generate the actual demand for each node using the Monte Carlo method. The demand fulfillment status is then evaluated based on the safety inventory policy and the actual demand.

It should be understood that, in the prediction process of the demand prediction module, there may be situations where the consideration is not comprehensive, and the demand of the output commodity in the future time period may be inaccurate, so that the monte carlo method may be applied to simulate the actual demand.

Optionally, the parameters input by the inventory simulation module may also include, but are not limited to, the following: safety stock of each node, stock of existing raw materials or semi-finished products, sales plans, business rules in the production flow, and the like.

Optionally, the inventory simulation module may further randomly generate an actual lead period by using a monte carlo method according to the input lead period of the high-risk node.

It should be understood that if the number of simulations is large enough, it will have a comprehensive evaluation meaning. For example, multiple sets of demand data may be generated based on the multiple time windows, the simulation process may be repeated to obtain multiple sets of inventory and backorder quantities, and the average inventory and backorder quantities may be calculated to determine the demand satisfaction rate. For example, the actual demand is simulated by using a monte carlo method, and the demand satisfaction rate is calculated according to the ratio of the inventory to the actual demand.

It should also be appreciated that the specific process of simulating actual demand or actual lead time using the Monte Carlo method and evaluating safety inventory plans via the inventory simulation module is referred to in the art and will not be described in detail herein for the sake of brevity.

In step 460, the inventory tuning module tunes the safety inventory of the key nodes that do not meet the expectation index.

According to the evaluation result output by the simulation module, the safety stock of the product with the demand satisfaction rate or other key indexes not meeting the expected indexes is further optimized, namely the adjustment of the stock strategy.

The desired target may be set manually. For example, a target demand satisfaction rate may be preset for each key node. And aiming at the nodes with the demand satisfaction rate lower than the preset target demand satisfaction rate, adjusting the safety stock of the key nodes based on the quantity of the shortage and the demand output by the stock simulation module and the target demand satisfaction rate. For example, if a code is in an out-of-stock state, the safety stock can be increased, if the stock of a node is redundant, the safety stock can be reduced, and if the demand satisfaction rate does not meet the expected target after adjustment, the algorithm can be iterated for multiple times until the demand satisfaction rates of all nodes reach the respective corresponding target demand satisfaction rates, and the lowest safety stock level is output, that is, the lowest safety stock level meeting the expected targets of all nodes.

Based on the technical scheme, a plurality of nodes for manufacturing and/or assembling the commodity are determined according to the BOM network of the commodity, and then the key nodes needing to plan safety stock and the safety stock of each key node are determined based on the lead period of each node and the service time of the upstream node of the node.

Fig. 6 is a schematic block diagram of an apparatus 600 for determining a safety stock provided by an embodiment of the present application. The apparatus 600 may be a chip system, or may also be an apparatus configured with a chip system, so as to implement the function of determining the security inventory in the foregoing method embodiment. In the embodiment of the present application, the chip system may be composed of a chip, and may also include a chip and other discrete devices.

As shown in fig. 6, the apparatus 600 may include a processor 610 and a communication interface 620. Communication interface 620 may be used, among other things, to communicate with other devices via a transmission medium such that the apparatus used in apparatus 600 may communicate with other devices. The communication interface 620 may be, for example, a transceiver, an interface, a bus, a circuit, or a device capable of performing a transceiving function. The processor 610 may utilize the communication interface 620 to input and output data and to implement the method of determining safety stock described in the corresponding embodiment of fig. 4.

Illustratively, the processor 610 may be configured to determine a plurality of nodes for manufacturing and/or assembling a commodity based on a bill of materials of the commodity, the commodity including a finished product or a semi-finished product; determining one or more key nodes needing to plan safety stock and the safety stock of each key node based on the lead period of each node in the plurality of nodes and the service time of the upstream node; the node lead period comprises a service time of the node, a manufacturing lead period and a supply lead period of the node. For details, reference is made to the detailed description of the method embodiments, which is not repeated herein.

Optionally, the apparatus 600 further comprises at least one memory 630 for storing program instructions and/or data. A memory 630 is coupled to the processor 610. The coupling in the embodiments of the present application is an indirect coupling or communication connection between devices, units or modules, and may be in an electrical, mechanical or other form, which is used for information interaction between the devices, units or modules. The processor 610 may operate in conjunction with the memory 630. The processor 610 may execute program instructions stored in the memory 630. At least one of the at least one memory may be included in the processor.

The specific connection medium between the processor 610, the communication interface 620 and the memory 630 is not limited in the embodiments of the present application. In fig. 6, the processor 610, the communication interface 620, and the memory 630 are connected by a bus 640. The bus 640 is represented by a thick line in fig. 6, and the connection between other components is merely illustrative and not intended to be limiting. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown in FIG. 6, but that does not indicate only one bus or one type of bus.

The present application provides a computer program product, the computer program product comprising: a computer program (also referred to as code, or instructions), which when executed, causes a computer to perform the method of the embodiment shown in fig. 4.

The present application also provides a computer-readable storage medium having stored thereon a computer program (also referred to as code, or instructions). Which when executed, causes a computer to perform the method of the embodiment shown in fig. 4.

It should be understood that the processor in the embodiments of the present application may be an integrated circuit chip having signal processing capability. In implementation, the steps of the above method embodiments may be performed by integrated logic circuits of hardware in a processor or by instructions in the form of software. The processor may be a general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic device, or discrete hardware components. The various methods, steps, and logic blocks disclosed in the embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present application may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in ram, flash memory, rom, prom, or eprom, registers, etc. storage media as is well known in the art. The storage medium is located in a memory, and a processor reads information in the memory and completes the steps of the method in combination with hardware of the processor.

It will also be appreciated that the memory in the embodiments of the subject application can be either volatile memory or nonvolatile memory, or can include both volatile and nonvolatile memory. The non-volatile memory may be a read-only memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an electrically Erasable EPROM (EEPROM), or a flash memory. Volatile memory can be Random Access Memory (RAM), which acts as external cache memory. By way of example, and not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), dynamic random access memory (dynamic RAM, DRAM), synchronous Dynamic Random Access Memory (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), SLDRAM (synchronous DRAM), and direct rambus RAM (DR RAM). It should be noted that the memory of the systems and methods described herein is intended to comprise, without being limited to, these and any other suitable types of memory.

As used in this specification, the terms "unit," "module," and the like are intended to refer to a computer-related entity, either hardware, firmware, a combination of hardware and software, or software in execution.

Those of ordinary skill in the art will appreciate that the various illustrative logical blocks and steps (step) described in connection with the embodiments disclosed herein may be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the technical solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application. In the several embodiments provided in the present application, it should be understood that the disclosed apparatus, device and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one position, or may be distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit.

In the above embodiments, the functions of the functional units may be fully or partially implemented by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions (programs). The procedures or functions described in accordance with the embodiments of the present application are generated in whole or in part when the computer program instructions (programs) are loaded and executed on a computer. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website site, computer, server, or data center to another website site, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that includes one or more available media. The usable medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a magnetic tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a Solid State Disk (SSD)), among others.

The functions may be stored in a computer-readable storage medium if they are implemented in the form of software functional units and sold or used as separate products. Based on such understanding, the technical solution of the present application or portions thereof that substantially contribute to the prior art may be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a read-only memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.

The above description is only for the specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present application, and shall be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A method of determining a safety stock, comprising:

determining a plurality of nodes for manufacturing and/or assembling a commodity based on a bill of materials of the commodity, wherein the commodity comprises a finished product or a semi-finished product;

determining one or more key nodes needing to plan safety stock and the safety stock of each key node based on the service time of each node in the plurality of nodes and the lead period of the upstream node; the upstream node's lead period comprises a service time of the upstream node, the service time representing a response time to a demand of a downstream node, a manufacturing lead period of the upstream node, and a provisioning lead period.

2. The method of claim 1, wherein determining one or more critical nodes that need to plan safety inventory, and the safety inventory amount of each critical node, based on the service time of each of the plurality of nodes and the lead period of its upstream nodes, comprises:

calculating the total goods-holding cost of the plurality of nodes by taking the service time of each node in the plurality of nodes and the lead period of the upstream node as variables;

determining one or more key nodes needing to plan safety stock and the safety stock of each key node by taking the minimum total goods holding cost of the nodes as a target; wherein the node with the safe inventory amount close to zero is not taken as the key node.

3. The method of claim 1 or 2, wherein the method further comprises:

predicting a trend of change in demand of the commodity over a future period of time based on historical sales data for the commodity;

and determining that safety stock needs to be planned for the commodity according to the change trend of the demand of the commodity in the future time period.

4. The method of claim 3, wherein predicting a trend of change in demand for the good over a future period of time based on historical sales data for the good comprises:

predicting a statistical characteristic of demand of the commodity in a plurality of time windows in the future based on historical sales data of the commodity, wherein the time windows are time windows in the future period;

determining a trend of change in demand of the good over the future time period based on the statistical features of the demand over the plurality of time windows.

5. The method of claim 4, wherein the statistical characteristic comprises an expectation value and a standard deviation, the expectation value being indicative of the demand quantity and the standard deviation being indicative of a fluctuation rate of the demand quantity.

6. The method of claim 4 or 5, wherein predicting a statistical characterization of demand for the good over a plurality of time windows in the future based on historical sales data for the good comprises:

extracting features of the commodity based on historical sales data of the commodity;

constructing a statistical model based on the characteristics of the commodity;

and predicting the statistical characteristics of the demand of the commodity in a plurality of time windows in the future through the statistical model.

7. The method of claim 6, wherein the features include one or more of: order timing characteristics, commodity attribute characteristics, bill of material characteristics, event characteristics, plan characteristics, business experience characteristics, and externality characteristics.

8. The method of any one of claims 1 to 7, further comprising:

evaluating a safety inventory of each of the one or more critical nodes to determine a demand satisfaction rate for each critical node.

9. The method of claim 8, wherein the method further comprises:

and adjusting the safety stock of the key nodes which do not reach the expected indexes, wherein the key nodes which do not reach the expected indexes are the nodes of which the demand satisfaction rate does not reach the preset target demand satisfaction rate.

10. An apparatus for determining a safety inventory, comprising at least one processor configured to execute a computer program to cause the apparatus to:

determining one or more key nodes needing to plan safety stock and the safety stock of each key node based on the service time of each node in the plurality of nodes and the lead period of the upstream node; the upstream node's lead period comprises a service time of the upstream node, which represents a response time to a demand of a downstream node, a manufacturing lead period of the upstream node, and a provisioning lead period.

11. The apparatus of claim 10, wherein the at least one processor is configured to execute the computer program to cause the apparatus to:

calculating the total goods cost of the plurality of nodes by taking the service time of each node in the plurality of nodes and the lead period of the upstream node thereof as variables;

12. The apparatus of claim 10 or 11, wherein the at least one processor is configured to execute the computer program to cause the apparatus to:

13. The apparatus of claim 12, wherein the at least one processor is configured to execute the computer program to cause the apparatus to:

determining a trend of change in demand for the commodity over the future period of time based on the statistical features of the demand in the plurality of time windows.

14. The apparatus of claim 13, wherein the statistical characteristic comprises an expected value indicative of the demand and a standard deviation indicative of a fluctuation rate of the demand.

15. The apparatus of claim 13 or 14, wherein the at least one processor is configured to execute the computer program to cause the apparatus to:

constructing a statistical model based on the characteristics of the commodity;

16. The apparatus of claim 15, wherein the features comprise one or more of: order timing characteristics, commodity attribute characteristics, bill of material characteristics, event characteristics, plan characteristics, business experience characteristics, and externality characteristics.

17. An apparatus according to any of claims 10 to 16, wherein the at least one processor is configured to execute the computer program to cause the apparatus to:

evaluating a safety inventory of each of the one or more key nodes to determine a demand satisfaction rate for each key node.

18. The apparatus of claim 17, wherein the at least one processor is configured to execute the computer program to cause the apparatus to:

19. A computer-readable storage medium, on which a computer program is stored, which, when run on a computer, causes the computer to carry out the method according to any one of claims 1 to 9.

20. A computer program product, comprising a computer program which, when executed, causes a computer to perform the method of any one of claims 1 to 9.