Multi-level inventory control methods and devices

Abstract

A method and apparatus for controlling multi-level inventory are provided. The multi-level inventory structure includes multiple central warehouses storing spare parts allocated from suppliers and multiple regional warehouses storing spare parts allocated from the multiple central warehouses. Each of the multiple central warehouses radiates to at least one of the multiple regional warehouses. The multi-level inventory control method includes: checking the inventory level, stockout level, and stockout level of each central warehouse, regional warehouse, and regional warehouse according to an inspection cycle; determining a steady-state probability of the inventory state based on the inventory level, stockout level, and regional warehouse stockout level, whereby the steady-state probability characterizes the inventory state when the multi-level inventory reaches a steady state; and determining an inventory strategy corresponding to the multiple central warehouses and multiple regional warehouses based on the steady-state probability of the inventory state.

Description

Technical Field

[0001] This disclosure generally relates to the field of multi-level inventory control, and more specifically, to a method and apparatus for controlling multi-level inventory. Background Technology

[0002] In existing multi-level inventory optimization methods, lateral transfers within a two-level service spare parts inventory system are typically considered. An accurate cost calculation formula is derived using the limiting distribution of a Markov process on a baseline inventory strategy. While accurate, this approach is time-consuming and unsuitable for large-scale systems. Other existing techniques employ approximation methods to calculate minimum costs, proposing two-level spare parts inventory models that include lateral transfers and emergency transportation. While using approximations improves solution speed, the assumptions underlying these approximation methods limit their application scenarios. Summary of the Invention

[0003] One of the objectives of the exemplary embodiments disclosed herein is to provide a method for controlling multi-level inventory.

[0004] According to one aspect of this disclosure, a method for controlling multi-level inventory is provided, wherein the structure of the multi-level inventory includes multiple central warehouses storing spare parts allocated from suppliers and multiple regional warehouses storing spare parts allocated from the multiple central warehouses, each of the multiple central warehouses radiating to at least one of the multiple regional warehouses, the control method comprising: checking the inventory level of each central warehouse, the stockout level of each central warehouse, the inventory level of each regional warehouse, and the stockout level of each regional warehouse according to an inspection cycle; determining a steady-state probability of inventory status based on the inventory level of each central warehouse, the stockout level of each central warehouse, the inventory level of each regional warehouse, and the stockout level of each regional warehouse, the steady-state probability of inventory status being used to characterize the inventory status when the multi-level inventory reaches a steady state; and determining an inventory strategy corresponding to the multiple central warehouses and the multiple regional warehouses based on the steady-state probability of inventory status.

[0005] According to embodiments of this disclosure, the inventory strategy may include checkpoints and maximum inventory levels for each central warehouse and each regional warehouse, wherein when the inventory level of any central warehouse or any regional warehouse is lower than the corresponding checkpoint, the inventory level of any central warehouse or any regional warehouse is increased to the corresponding maximum inventory level by adding the spare parts.

[0006] According to embodiments of this disclosure, the step of determining the steady-state probability of inventory status may include: determining the inventory status transition probability based on the inventory quantity of each central warehouse, the stockout quantity of each central warehouse, the inventory quantity of each regional warehouse, and the stockout quantity of each regional warehouse, wherein the inventory status transition probability is used to characterize the inventory changes of spare parts in a multi-level inventory; and determining the steady-state probability of inventory status based on the inventory status transition probability.

[0007] According to embodiments of this disclosure, the inspection cycle is divided into a pre-arrival stage and a post-arrival stage based on a pre-set lead time. The stockout quantity of each regional warehouse includes the stockout quantity before arrival and the stockout quantity after arrival of each regional warehouse. The inventory status transition probability includes the period-end inventory status transition probability of each regional warehouse, the period-end inventory status transition probability of each central warehouse, and the pre-arrival-post-arrival inventory status transition probability of each regional warehouse.

[0008] According to embodiments of this disclosure, the step of determining the inventory status transition probability includes: obtaining the period-end inventory status transition probability of each regional warehouse based on the inventory level of each regional warehouse, the stockout level before arrival at each regional warehouse, and the stockout level after arrival at each regional warehouse; for each central warehouse, obtaining the period-end inventory status transition probability of the central warehouse based on the inventory level of the central warehouse and the stockout levels before arrival and after arrival at the regional warehouses radiated by the central warehouse; obtaining the pre-arrival-post-arrival inventory status transition probability of each regional warehouse based on the inventory level of each regional warehouse, the stockout level before arrival at each regional warehouse, and the stockout level after arrival at each regional warehouse; and determining the inventory status transition probability based on the period-end inventory status transition probability of each regional warehouse, the period-end inventory status transition probability of each central warehouse, and the pre-arrival-post-arrival inventory status transition probability of each regional warehouse.

[0009] According to embodiments of this disclosure, the step of determining an inventory strategy corresponding to multiple central warehouses and multiple regional warehouses based on the steady-state probability of inventory status includes: obtaining supply chain reliability and cost based on at least one of inventory status transition probability and inventory status steady-state probability; and constructing an objective function based on cost, with supply chain reliability as a constraint, and determining the inventory strategy corresponding to multiple central warehouses and multiple regional warehouses by minimizing the objective function.

[0010] According to embodiments of this disclosure, during each inspection cycle, the inventory level of each regional warehouse or each central warehouse includes the initial inventory level corresponding to the start time of the inspection cycle. The inspection cycle is divided into a pre-arrival stage and a post-arrival stage based on a pre-set lead time. Supply chain reliability is obtained through the following steps: determining the quantity to be satisfied within a predetermined time for each regional warehouse based on the demand before arrival, the demand after arrival, the initial inventory level, and the maximum inventory level; for each central warehouse, obtaining the quantity to be satisfied within a predetermined time for that central warehouse based on its initial inventory level, its maximum inventory level, and the quantity to be satisfied within a predetermined time for the regional warehouses it covers; and obtaining the supply chain reliability based on the steady-state probability of inventory status, the quantity to be satisfied within a predetermined time for each regional warehouse, and the quantity to be satisfied within a predetermined time for each central warehouse.

[0011] According to embodiments of this disclosure, the costs include the manufacturer's total cost and the owner's total cost of the spare parts, wherein the manufacturer's total cost includes at least one of the following: supplier allocation costs, allocation costs for each central warehouse, allocation costs for each regional warehouse, ordering costs, transportation costs, and holding costs.

[0012] According to another aspect of this disclosure, a computer-readable storage medium is provided, characterized in that the computer-readable storage medium stores instructions or programs that, when executed by a processor, implement the control method described above.

[0013] According to another aspect of this disclosure, a control device for multi-level inventory is provided, wherein the structure of the multi-level inventory includes multiple central warehouses storing spare parts allocated from suppliers and multiple regional warehouses storing spare parts allocated from the multiple central warehouses, each of the multiple central warehouses radiating to at least one of the multiple regional warehouses, the control device comprising: a parameter acquisition unit for checking the inventory quantity of each central warehouse, the stockout quantity of each central warehouse, the inventory quantity of each regional warehouse, and the stockout quantity of each regional warehouse according to an inspection cycle; a probability determination unit for determining a steady-state probability of the inventory state based on the inventory quantity of each central warehouse, the stockout quantity of each central warehouse, the inventory quantity of each regional warehouse, and the stockout quantity of each regional warehouse, the steady-state probability of the inventory state being used to characterize the inventory state when the multi-level inventory reaches a steady state; and a strategy determination unit for determining an inventory strategy corresponding to the multiple central warehouses and the multiple regional warehouses based on the steady-state probability of the inventory state.

[0014] According to one or more aspects of this disclosure, the provided multi-level inventory control method and control apparatus calculate costs and supply chain reliability based on steady-state probabilities, and implement a non-dominated multi-level inventory strategy with the goal of minimizing manufacturer and owner costs and supply chain reliability as a constraint.

[0015] Furthermore, according to one or more aspects of this disclosure, the multi-level inventory control method and control device decompose the state transition probability into: the state transition probability of the ending inventory of the regional warehouse, the state transition probability of the ending inventory of the central warehouse, and the state transition probability of the regional warehouse before and after arrival. After calculating the three types of probabilities separately, the transition probability of spare part k in all warehouses within subsystem i is obtained, and then the steady-state probability of spare part k in all warehouses within subsystem i is obtained. By decomposing the conditional probability, the computational complexity of the state transition probability is simplified, and the curse of dimensionality is alleviated.

[0016] Furthermore, with advancements in manufacturing and technology, lead times have been significantly reduced, rendering conventional assumptions about lead times in the prior art (e.g., lead time is greater than or equal to the inspection cycle, or the lead time is assumed to be zero) out of touch with current realities. According to one or more aspects of this disclosure, the multi-level inventory control method and apparatus perform steady-state analysis of the multi-level inventory system when the lead time is less than the inspection cycle, resulting in a multi-level inventory strategy that better reflects actual conditions. Attached Figure Description

[0017] The above and other objects and features of exemplary embodiments of this disclosure will become clearer from the following description taken in conjunction with the accompanying drawings, which exemplarily illustrate the embodiments, wherein:

[0018] Figure 1 This is a scene diagram illustrating a multi-level inventory control method according to an embodiment of the present disclosure;

[0019] Figure 2 This is a flowchart illustrating a multi-level inventory control method according to an embodiment of the present disclosure;

[0020] Figure 3 This is a flowchart illustrating the determination of the steady-state probability of inventory status and the determination of inventory strategy according to embodiments of the present disclosure;

[0021] Figure 4 These are example diagrams illustrating different schemes according to embodiments of the present disclosure;

[0022] Figure 5 This is a cost example diagram showing different lead times; and

[0023] Figure 6 This is a block diagram illustrating a control device for a multi-level inventory according to an embodiment of the present disclosure. Detailed Implementation

[0024] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings, examples of which are illustrated in the drawings, wherein the same reference numerals always refer to the same parts. The embodiments will now be described with reference to the accompanying drawings in order to explain this disclosure.

[0025] Figure 1 This is a scene diagram illustrating a multi-level inventory control method according to an embodiment of the present disclosure.

[0026] by Figure 1 Taking the wind turbine manufacturer's inventory system in the supply chain structure shown in the figure as an example, the top layer is the supplier of service spare parts; the second layer is the wind turbine manufacturer's service spare parts central warehouse, with N warehouses. C The third layer is the service spare parts warehouse for wind turbine manufacturers, and the number of regional warehouses radiating from the central warehouse i is N. L，iThe fourth layer represents the wind farm owned by the wind turbine operator. Normally, wind farms do not store spare parts. When a regional warehouse j, which is radiated by central warehouse i, requires spare part k, the first consideration is whether regional warehouse j can meet the demand. If regional warehouse j cannot meet the demand, then it is sequentially allocated from central warehouse i and urgently ordered from the supplier.

[0027] In existing technologies, lateral transfers in a two-tier spare parts inventory system are typically considered. The limiting distribution of a Markov process is used to derive an accurate cost calculation formula based on a baseline inventory strategy. Furthermore, existing technologies derive an accurate recursive process under a baseline inventory strategy at a central warehouse to determine the expected inventory holding cost and stockout cost of the inventory system under the Poisson demand assumption, and use this result to study the allocation policy problem of a single retailer. Alternatively, in the context of a single-warehouse, multi-retailer inventory system, an (R, nQ) strategy is used to procure spare parts from suppliers, deriving an accurate probability distribution of retailer inventory levels, and using these distributions to obtain an accurate expression for the system cost. As briefly described above, while existing technologies are accurate, they are time-consuming and unsuitable for large-scale system calculations. Other existing technologies use approximate methods to calculate minimum costs, proposing a two-tier spare parts inventory model that includes lateral and emergency transportation. Alternatively, heuristic methods applicable to practice can approximate solutions to real-world single-warehouse, multi-retailer problems. While the aforementioned existing technologies use approximation to build models and improve solution speed, their assumptions limit the application scenarios of these approximation methods.

[0028] To simplify the description and avoid unnecessarily obscuring the concept of this disclosure, the following assumptions can be made based on embodiments of this disclosure: no lateral transfers occur between central warehouses or regional warehouses; emergency orders from suppliers have shorter lead times but higher ordering costs compared to normal orders; and because regional warehouses are located close to wind farms, it is assumed that the time T for all spare parts to travel from the regional warehouse to the wind farm is... L Equal to the duration specified for regional warehouse supply chain reliability Unit transfer cost c L It also remains unchanged. However, the response time T from the central warehouse i and the supplier to the wind farm remains unchanged. C，i，k T S，i，k Unit transfer costs c D，i，k c P，i，k They will be differentiated based on the type of spare parts and the central warehouse.

[0029] In this embodiment, each warehouse can employ an (s, S) strategy when ordering spare parts from a spare parts supplier, with an inventory check cycle of T. The check cycle refers to the period during which the remaining inventory is counted and it is determined whether materials need to be ordered. The spare parts inventory check cycle can be a fixed value or a variable value. Hereinafter, the check cycle is described as a fixed value, but the present disclosure is not limited thereto. For example, the inventory strategy includes checkpoints and maximum inventory levels for each central warehouse and each regional warehouse, wherein when the inventory level of any central warehouse or any regional warehouse falls below the corresponding checkpoint, the inventory level of that central warehouse or regional warehouse is increased to the corresponding maximum inventory level by adding spare parts. Specifically, at a certain order checkpoint, for central warehouse i and regional warehouse j, the initial inventory level U of spare parts k satisfies... i，k U ij，k (Or initial inventory) less than or equal to reorder point s i，k s i，j，k The warehouse with the specified conditions will order the inventory of spare part k up to the maximum inventory level S. i，k or S i，j，k Otherwise, spare part k will not be ordered. After one inspection cycle T, the ending inventory of spare part k in the central warehouse i and the regional warehouse j is U′. i，k 、U′ i，j，k Spare parts inventory plan Specifically, N S The ordering points s and maximum inventory S of a type of spare part in all warehouses, i.e. {s} i，k S i，k}.in For spare parts k, there are ordering points in the central warehouse i and all the regional warehouses radiating from it; Let k be the maximum inventory of spare parts in central warehouse i and all its surrounding regional warehouses.

[0030] Therefore, cost and reliability can be further considered to find the optimal service spare parts inventory plan. This minimizes the total cost for both the manufacturer and the owner, while meeting supply chain reliability constraints. The question concerns finding the optimal service spare parts inventory plan. Reference Figure 2 and Figure 3 To provide a more detailed description.

[0031] Figure 2 This is a flowchart illustrating a multi-level inventory control method according to an embodiment of the present disclosure. Figure 3 This is a flowchart illustrating the determination of the steady-state probability of inventory status and the determination of inventory strategy according to embodiments of the present disclosure. As described above, the multi-level inventory structure includes multiple central warehouses storing spare parts allocated from suppliers and multiple regional warehouses storing spare parts allocated from the multiple central warehouses, each of the multiple central warehouses radiating to at least one of the multiple regional warehouses.

[0032] Reference Figure 2 In step S10, the inventory level of each central warehouse, the stockout level of each central warehouse, the inventory level of each regional warehouse, and the stockout level of each regional warehouse are checked according to the inspection cycle.

[0033] In this embodiment, during each inspection cycle, the inventory level of each regional warehouse or each central warehouse includes the initial inventory level corresponding to the start time of the inspection cycle. For example, according to the inspection cycle, for each central warehouse i and regional warehouse j, at a certain order inspection point, the initial inventory level U of spare part k can be checked. i，k U i，j，k (Or initial inventory). After one inspection cycle T, the ending inventory of spare part k in central warehouse i and regional warehouse j is U′. i，k 、U′ i，j，k .

[0034] In this embodiment, lead time, or component lead time, refers to the procurement cycle of the component. Since the supplier's preparation time is much longer than the transportation time from the central warehouse to the regional warehouse, this disclosure assumes that the lead time of component k arriving at the regional warehouse and the central warehouse is the same, both being T. P，k When the lead time is shorter than the inspection period, the inspection period is divided into a pre-arrival stage and a post-arrival stage based on a pre-set lead time. For example, the lead time T can be used... P，k The inspection cycle T is divided into a pre-delivery stage and a post-delivery stage. The average demand for spare parts k in regional warehouses j, which are radiated by central warehouse i, is λ during the inspection cycle T. i，j，k The average demand quantities in the pre-delivery and post-delivery stages are respectively in, Therefore, the shortage quantity of spare parts k in regional warehouse j before and after delivery can be checked. And the shortage quantity of spare parts k before and after the arrival of central warehouse i.

[0035] In step S20, based on the inventory level of each central warehouse (e.g., the initial inventory level U of the central warehouse)... i，k and ending inventory U′ i，k The stockout quantity for each central warehouse (e.g., the stockout quantity of spare parts k before and after the arrival of goods in central warehouse i is...). ), the inventory level of each regional warehouse (e.g., the initial inventory level U of the regional warehouse). i，j，k and ending inventory U′ i，j，k ) and the stockout quantity for each regional warehouse (e.g., the stockout quantity of spare parts k before and after the arrival of goods in regional warehouse j is ). Determine the steady-state probability of the inventory state, which is used to characterize the inventory state when the multi-level inventory reaches a steady state.

[0036] More specifically, the step of determining the steady-state probability of the inventory status may include: determining the inventory status transition probability based on the inventory quantity of each central warehouse, the stockout quantity of each central warehouse, the inventory quantity of each regional warehouse, and the stockout quantity of each regional warehouse, wherein the inventory status transition probability is used to characterize the inventory change of the spare parts in the multi-level inventory; and then determining the steady-state probability of the inventory status based on the inventory status transition probability.

[0037] The following will combine Figure 1 The example shown illustrates how supply chain structure descriptions yield the probability of inventory state transitions.

[0038] According to the principle of total probability, given the initial inventory state of spare part k in subsystem i, the conditional probability of its ending inventory state can be written as:

[0039]

[0040] in, Let be the vector of the ending inventory of spare part k in all warehouses within subsystem i. The vector of initial inventory. Let be the vector of the number of spare parts k out of stock before and after their arrival in all regional warehouses within subsystem i.

[0041] Since the ending inventory status of regional warehouses is independent of each other, the ending inventory status of the central warehouse is only related to its beginning inventory status and the amount of stockouts in the regional warehouses during the inspection period, and is unrelated to the ending inventory status of the regional warehouses. The first term on the right side of the equation can be decomposed into:

[0042]

[0043] The second term on the right side of the equation can be factored into:

[0044]

[0045] in Let be the shortage quantity vector of spare part k before and after its arrival in regional warehouse j under central warehouse i. The inventory state transition probability can be obtained from equations (1), (2), and (3).

[0046]

[0047]

[0048] In the above formula, Pr(U′) i，k |U i，k ) Need to calculate |U i，k |×|U i，k There are three probabilities, and the calculation of each probability is very complex, requiring consideration of |X|. i，k | Different types of out-of-stock situations. (Here|Ui，k |and|X i，k |These are U i，k and X i，k After decomposition (size of state space) Pr(U′) i，k |U i，k X i，k ) is a function that returns either 0 or 1, because U′ i，k It can be made by U i，k and X i，k Calculated. Furthermore, Pr(U′) i，j，k |U i，j，k X i，j，k ) and Pr(X i，j，k |U i，j，k The results can be saved for reuse.

[0049] When the inspection cycle is divided into pre-arrival and post-arrival stages based on a pre-set lead time, the stockout quantity X for each regional warehouse is... i，j，k Including stockouts before arrival at each regional warehouse And the amount of stockouts after arrival at each regional warehouse The inventory state transition probability Pr(U′) i，k |U i，k This includes the period-end inventory state transition probability Pr(U′) for each regional warehouse. i，j，k |U i，j，k X i，j，k The probability of state transition of the ending inventory of each central warehouse, Pr(U′). i，k |U i，k X i，k ) and the inventory status transition probability Pr(X) before and after arrival for each regional warehouse i，j，k |U i，j，k ).

[0050] The following will list the end-of-period inventory state transition probability Pr(U′) for the specific calculation area warehouse. i，j，k |U i，j，k X i，j，k The probability of state transition of ending inventory in the central warehouse, Pr(U′). i，k |U i，k X i，k ) and the inventory status transition probability Pr(X) before and after arrival for each regional warehouse i，j，k |U i，j，k ).

[0051] First, based on the inventory level of each regional warehouse, the stockout level before arrival at each regional warehouse, and the stockout level after arrival at each regional warehouse, obtain the ending inventory state transition probability Pr(U′) for each regional warehouse. i，j，k |U i，j，k Xi，j，k ).

[0052] For a regional warehouse, based on the principles of conditional probability and total probability, given the initial inventory level and the stockout levels before and after arrival, the conditional transition probability of its ending inventory state is Pr(U′). i，j，k |U i，j，k X i，j，k The calculation formula is as follows:

[0053]

[0054] The probability of inventory status transition can be calculated by converting it into the probability of spare parts demand. Considering regional warehouses, the demand for a spare part before and after delivery is independent. Therefore, the transition probability can be decomposed into the product of the probabilities of demand before and after delivery. Furthermore, it is necessary to discuss whether spare parts were ordered at the beginning of the period separately.

[0055] (1)When U i，j，k >s i，j，k At that time, no restocking is needed. Furthermore, the discussion will proceed depending on whether a stockout occurs:

[0056] (I)If If there are no stock shortages before or after delivery, then

[0057]

[0058]

[0059] Where poss(Y, λ) is the probability density function of the Poisson distribution, Y is the demand quantity, and λ is the average demand quantity.

[0060] (II) If This means that at least one stockout occurs before or after the arrival of goods, and the ending inventory level is 0. The probability of this inventory level is:

[0061] Pr(U′ i，j，k |U i，j，k X i，j，k )=I(U′ i，j，k =0) (7)

[0062] (2)When U i，j，k ≤s i，j，k When this happens, restocking is necessary. The discussion will also consider whether a stockout occurs:

[0063] (I)If If there is a shortage of goods before the goods arrive, then

[0064] Pr(U′ i，j，k |U i，j，k X i，j，k )=I(U′ i，j，k=0) (8)

[0065] (II) If but

[0066]

[0067]

[0068] (III) If and If there was no shortage of stock before delivery, but stock shortages occurred after delivery due to consumption during transportation, then...

[0069]

[0070] Secondly, for each central warehouse, the ending inventory transition probability Pr(U′) is obtained based on the central warehouse's inventory level and the stockout levels of the regional warehouses it serves before and after the arrival of goods. i，k |U i，k X i，k ).

[0071] For the central warehouse, its transition probability is related to whether other regional warehouses are out of stock. Given the initial inventory level and the corresponding regional warehouse stockouts before and after arrival, its ending inventory conditional transition probability Pr(U′) is... i，k |U i，k X i，k )for

[0072] (1)When U i，k >s i，k When no replenishment is needed, the ending inventory of the central warehouse is the remaining quantity after satisfying the stockout requirements of its corresponding regional warehouses; otherwise, it is 0. Therefore, the inventory state probability can also be represented by an indicator function:

[0073]

[0074] (2)When U i，k ≤s i，k Restocking is required at this time:

[0075] like That is, if the initial inventory of the central warehouse is insufficient to meet the stockout requirements of all its corresponding regional warehouses, then...

[0076]

[0077] In other cases

[0078]

[0079] Next, based on the inventory level of each regional warehouse, the stockout level before arrival at each regional warehouse, and the stockout level after arrival at each regional warehouse, the inventory state transition probability Pr(X) before arrival and after arrival for each regional warehouse is obtained. i，j，k |U i，j，k ).

[0080] For a regional warehouse, according to the principle of conditional probability, given the initial inventory level, the conditional transition probabilities of stockout quantities before and after delivery are as follows:

[0081] (1)When U i，j，k >s i，j，k No restocking is needed at this time:

[0082] (I)If but

[0083]

[0084] (II) If and but

[0085]

[0086] (III) If but

[0087]

[0088] (2)When U i，j，k ≤s i，j，k When restocking is needed, we need to consider another scenario: a stock shortage occurs before the goods arrive, but no stock shortage occurs after the goods arrive.

[0089] (I)If and but

[0090]

[0091] (II) If and but

[0092]

[0093] (III) If and but

[0094]

[0095] (IV) If but

[0096]

[0097] Based on the period-end inventory state transition probability Pr(U′) of each regional warehouse i，j，k |U i，j，k X i，j，k The probability of state transition of the ending inventory of each central warehouse, Pr(U′). i，k |U i，k X i，k ) and the pre-arrival-post-arrival inventory state transition probability Pr(X) for each regional warehouse. i，jk |U i，j，k Determine the inventory state transition probability Pr(U′) i，k |U i，k (See equation (4)).

[0098] Reference Figure 2 and Figure 3 The steady-state probability of an inventory state can be determined based on the inventory state transition probabilities. The steps for determining the steady-state probability from the inventory state transition probabilities can be obtained using the limiting distribution of a Markov process. Redundant descriptions are omitted here. This reflects the likelihood (probability) of the system being in a certain state after reaching steady state; that is, the steady-state probability of the inventory state.

[0099] Reference Figure 2 and Figure 3 In step S30, inventory strategies corresponding to multiple central warehouses and multiple regional warehouses are determined based on the steady-state probability of inventory status.

[0100] The manufacturing supply chain consists of multiple upstream and downstream enterprises, and changes at any node in the chain affect the entire supply chain. To mitigate the impact of unforeseen events on the supply chain system, it is necessary to develop optimal inventory strategies for suppliers and ensure that the supply chain reliability under these strategies reaches predetermined values ​​to guarantee the stability and reliability of the entire supply chain. In real-world business, suppliers build multi-level inventory systems to quickly meet the needs of remotely located customers. The goal of optimizing the inventory system is to maximize its reliability by allocating resources accordingly. For example, adjusting inventory strategies can improve system reliability and reduce corresponding costs.

[0101] Specifically, supply chain reliability and cost are obtained based on at least one of the inventory state transition probability and the inventory state steady-state probability; and an objective function is constructed based on cost, with supply chain reliability as a constraint, and the inventory strategy corresponding to multiple central warehouses and multiple regional warehouses is determined by minimizing the objective function.

[0102] The following will refer to Figure 3 , combined Figure 1 The example shown illustrates how supply chain structure descriptions are based on steady-state probabilities of inventory status to obtain supply chain reliability.

[0103] First, it can be based on the demand before the arrival of goods in each regional warehouse. Demand after goods arrive at each regional warehouse Initial inventory U of each regional warehouse i，j，k and the maximum inventory S of each regional warehouse i，j，k Determine the quantity to be satisfied within the predetermined time for each regional warehouse. and In this embodiment, the scheduled time can be a scheduled delivery time, such as the time agreed upon in the contract.

[0104] For example, Let k be the quantity vector of spare part k before and after arrival in all regional warehouses within subsystem i, and let k be the quantity to be satisfied within a predetermined time in each regional warehouse. and The details are as follows:

[0105]

[0106]

[0107] Then, for each central warehouse, based on the initial inventory U of that central warehouse... i，k and the maximum inventory S of the central warehouse i，k And the quantity to be satisfied within a predetermined time period for the regional warehouses covered by the central warehouse (e.g., the quantity to be satisfied within a predetermined time period for the regional warehouses). and The quantity required to be met within the predetermined timeframe of the central warehouse is obtained. and in, and The calculation formula is as follows:

[0108] when

[0109]

[0110] when hour

[0111]

[0112]

[0113] Based on the steady-state probability of the inventory status (e.g., the steady-state probability is related to the initial inventory U of the central warehouse). i，k The corresponding steady-state probability Pr(U) i，k ), steady-state probability and demand before regional warehouse arrival Corresponding steady-state probability And the demand after the arrival of goods in the regional warehouse in the steady-state probability. Corresponding steady-state probability The quantity to be satisfied within the predetermined time for each regional warehouse. and And the quantity to be met within the scheduled time for each central warehouse and To obtain the reliability of the supply chain.

[0114] In a specific embodiment, the supply chain reliability R of a single spare part k across all regional warehouses within subsystem i is... i，L，k It can be obtained through the following steps:

[0115]

[0116] Supply chain reliability of a single spare part k across all warehouses within subsystem i:

[0117]

[0118] At this point, this disclosure utilizes the Poisson distribution parameters of different spare parts in different warehouses to weight the supply chain reliability of spare parts in all regional warehouses within subsystem i, thereby calculating the supply chain reliability of all spare parts in all regional warehouses across the entire system, i.e.:

[0119]

[0120] Similarly, the supply chain reliability of all spare parts across all warehouses in the entire system can be calculated, i.e.:

[0121]

[0122] The following will refer to Figure 3 , combined Figure 1 The example shown illustrates a supply chain structure that describes the cost of obtaining costs.

[0123] In this embodiment, the cost includes the manufacturer’s total cost and the owner’s total cost of the spare parts, wherein the manufacturer’s total cost includes at least one of the following: supplier allocation cost, allocation cost of each central warehouse, allocation cost of each regional warehouse, ordering cost, transportation cost, and holding cost.

[0124] In the following text, taking spare part k within subsystem i as an example, we will consider the probabilities of all inventory states, as well as the quantity and unit price under different states, to finally obtain the expected values ​​of various costs. The total manufacturer cost and total owner cost listed below are merely examples and should not be construed as limiting this disclosure.

[0125] Supplier allocation costs

[0126] In subsystem i, given the initial inventory levels of all warehouses, and considering the stockout quantities of all regional warehouses in subsystem i before and after the arrival of goods, the conditional transition probability of the stockout quantity of the central warehouse before the arrival of goods can be expanded as follows:

[0127]

[0128] Among them, the stockout quantity in the central warehouse is related to the initial inventory quantity and the stockout quantity in the corresponding regional warehouses, as shown in the first term on the right side of equation (30). It can be written as an indicator function, i.e., I(a=0), which has a value of 1 when a=0 and 0 otherwise.

[0129]

[0130] The second term Pr(X) on the right-hand side of equation (30) i，k |U i，k This can be expanded as follows:

[0131]

[0132] Similarly, the conditional transition probability of the shortage quantity in the central warehouse of subsystem i after the arrival of goods can be expanded as follows:

[0133]

[0134] Among them, the second term Pr(X) on the right side of equation (33) i，k |U i，k As in equation (32), the first term There are two types of indicator functions depending on whether they are ordered:

[0135]

[0136] Furthermore, when the central warehouse of subsystem i lacks spare part k, the supplier allocation cost includes emergency ordering costs and emergency transportation costs to the wind farm. Due to the difficulty in coordinating spare parts, these costs are related to the quantity of spare part k and its unit cost across different subsystems. Therefore, this disclosure sums up all unit costs and refers to them collectively as the supplier allocation unit cost c. P，i，k Therefore, in subsystem i, its supplier allocation cost is as follows:

[0137]

[0138] Central warehouse transfer costs

[0139] In subsystem i, given the initial inventory levels of all warehouses, and considering the stockout quantities of all regional warehouses in subsystem i before and after the arrival of goods, the conditional transition probability of the central warehouse's transfer quantity before the arrival of goods can be expanded as follows:

[0140]

[0141] Among them, the second term Pr(X) on the right side of equation (36) i，k |Ui，k As shown in equation (32), the quantity transferred from the central warehouse is related to the shortage quantity of its corresponding regional warehouse. If the initial inventory of the central warehouse is insufficient to meet the shortage quantity, all inventory needs to be transferred to the regional warehouse. Therefore, the first term... It can be written as an indicator function:

[0142]

[0143] Similarly, the conditional transition probability of the quantity to be allocated in the central warehouse of subsystem i after the goods arrive can be expanded as follows:

[0144]

[0145] Among them, the first term on the right side of equation (38) There are two types of indicator functions depending on whether they are ordered:

[0146]

[0147] When the regional warehouse of subsystem i lacks spare part k, the central warehouse allocation cost refers to the cost of emergency transportation to the wind farm, with the central warehouse allocation unit cost c. D，i，k Therefore, in subsystem i, the central warehouse allocation cost is as follows:

[0148]

[0149] Regional warehouse transfer costs

[0150] In subsystem i, given the initial inventory of regional warehouse j, and considering the demand quantity of regional warehouse j before delivery, the conditional transition probability that regional warehouse j meets the demand quantity before delivery can be expanded as follows:

[0151]

[0152] in It can be written as an indicator function:

[0153]

[0154] Introducing the demand quantity of region j before and after delivery, the conditional transition probability of satisfying the quantity after delivery can be expanded as follows:

[0155]

[0156] in Let be the vector of wind farm demand for spare part k from regional warehouse j under central warehouse i before and after delivery. The first term on the right-hand side of the equation has two characteristic functions:

[0157]

[0158] There are N subsystems i L，i There are regional warehouses, and the unit cost of supplying wind farms from these regional warehouses is c. L Therefore, in subsystem i, the allocation costs for all its regional libraries are as follows:

[0159]

[0160] in,

[0161]

[0162]

[0163] Ordering costs and shipping costs

[0164] At the beginning of each inspection cycle, if the inventory U of spare part k in the warehouse of subsystem i is less than or equal to the reorder point s, spare part k needs to be ordered from the supplier. This incurs ordering costs and transportation costs. The ordering cost is only related to the number of orders and not the order quantity. The ordering cost for each subsystem i is as follows:

[0165]

[0166]

[0167] Furthermore, when ordering spare parts, this disclosure ignores the cost differences in transportation from the supplier to different warehouses within subsystem i. Therefore, the transportation cost for each subsystem i is as follows:

[0168]

[0169] Holding costs

[0170] Unit holding costs (e.g., inventory holding costs) are related to spare parts unit prices and annual interest rate V.

[0171]

[0172] The calculation of holding costs for a single regional warehouse is based on U i，j，k Whether restocking is needed depends on two situations:

[0173] (1)When U i，j，k >s i，j，k No restocking is needed at this time:

[0174]

[0175] (2)When U i，j，k ≤s i，j，k When restocking is required, the holding time for spare parts varies depending on the restocking situation, and therefore needs to be calculated separately:

[0176]

[0177] in

[0178]

[0179] To calculate the holding costs of the central warehouse, we need to use the following transfer quantities. The average transfer quantity before replenishment arrives at the central warehouse is...

[0180]

[0181] The average transfer quantity after replenishment at the central warehouse is

[0182]

[0183] According to U i，k Whether restocking is needed depends on two situations.

[0184] WhenU i，k >s i，k No restocking is needed at this time:

[0185]

[0186] (2)When U i，k ≤s i，k Restocking is required at this time:

[0187]

[0188] Therefore, the total holding cost of a subsystem is

[0189]

[0190] Total Ownership Cost

[0191] When a wind turbine malfunctions, operation can only be restored if spare parts are replaced. However, due to the unavailability of spare parts, the resulting downtime losses are borne by the wind turbine owner, and the cost is as follows:

[0192]

[0193] W in equation (57) S，i，k This includes downtime costs incurred due to supplier allocation to the wind farm. When subsystem i is out of stock, it needs to be replenished from the supplier, and the transportation time for spare part k is T. S，i，k The downtime loss per hour is E i At this point, the downtime cost within subsystem i is:

[0194]

[0195] W in equation (57) C，i，k This includes downtime costs incurred when spare parts are transferred from the central warehouse to the wind farm. When a regional warehouse in subsystem i is out of stock, replenishment from the central warehouse is required, with a transportation time of T for spare part k. C，i，k At this point, the downtime cost within subsystem i is:

[0196]

[0197] W in equation (57) L，i，k The downtime costs incurred due to emergency supplies from the regional warehouse to the wind farm. The transportation time for spare part k is T. L At this point, the downtime cost within subsystem i is:

[0198]

[0199] The following will refer to Figure 3 , combined Figure 1 The example shown illustrates a supply chain structure description based on a cost-based objective function, constrained by supply chain reliability, and an example of determining an inventory strategy by minimizing the objective function.

[0200] In this example embodiment, since we need to minimize two objective functions simultaneously, namely the manufacturer cost TC and the owner cost W, both of which are about The function. Furthermore, the regional warehouse supply chain reliability R L Need to be greater than Central warehouse supply chain reliability R L，C Need to be greater than

[0201] This disclosure optimizes inventory strategies from the manufacturer's perspective. During the warranty period, the manufacturer fulfills its maintenance obligations for the wind turbines and bears the costs of spare parts inventory. Simultaneously, to maintain a good relationship with the owner, it's necessary to consider the owner's downtime losses. Therefore, spare parts inventory strategy optimization is a bi-objective optimization problem. By adjusting the weights of manufacturer and owner costs, different non-dominated solutions can be obtained. For example, refer to... Figure 3 The PSO (Proportional Strategy for Non-Dominated Inventory) method can be used to solve for non-dominated solutions. When selecting a non-dominated inventory strategy, downtime costs are given greater weight if the turbine owner has a long-term strategic partnership, and vice versa. Since the two types of costs have different compositions and differ in magnitude, directly using cost values ​​would render the objective function weights uninterpretable. Therefore, this disclosure uses a minimum-maximum method to normalize the two objective functions. In this case, the maximum and minimum values ​​of the two types of costs can be pre-calculated, and under the condition of satisfying reliability constraints, the maximum and minimum values ​​of each individual objective function can be directly optimized. In contrast, methods such as Z-score standardization cannot be used because the required cost mean and variance are difficult to calculate in advance.

[0202] In this embodiment, the normalized manufacturer cost objective function Owner cost objective function as follows:

[0203]

[0204]

[0205] The final equation that needs to be solved in this disclosure is as follows:

[0206]

[0207] Where stw1+w2=1,

[0208]

[0209]

[0210] To obtain inventory plans with different cost combinations for wind turbine manufacturers and owners, it is only necessary to continuously change the weights w1 and w2 in equation (63). This disclosure uses a particle swarm optimization algorithm to solve the inventory strategy. In this embodiment, the particle swarm optimization algorithm exhibits good performance and has the advantages of having a small number of parameters and being easy to implement.

[0211] Below, we will use a partial inventory system of a wind turbine manufacturer as an example to specifically describe how it solves the optimal spare parts inventory plan.

[0212] In this embodiment, for simplicity, it is assumed that the top layer of the warehouse system contains a spare parts supplier, and the second layer contains two central warehouses, A and B, belonging to two wind turbine manufacturers. Central warehouse A has inconvenient transportation, while central warehouse B has relatively convenient transportation. The third layer consists of two regional warehouses below each central warehouse, namely A1, A2, B1, and B2. Regional warehouses A1 and A2 also have less convenient transportation than regional warehouses B1 and B2. The fourth layer comprises the wind farms owned by the wind turbine owners. Central warehouse A and regional warehouses A1 and A2 require spare parts for service to the wind farms. For management purposes, central warehouse B and regional warehouses B1 and B2 need to provide service spare parts for wind farms. Manage. Spare parts. Its weight and volume are much larger than those of spare parts. All other things being equal, spare parts The delivery time is longer than that of spare parts. The transportation time is long.

[0213] The wind turbine manufacturer's procurement strategy (s, S) has an inspection cycle T of 90 days and a unit holding cost c. H，i，kThe annual interest rate V is 20%. When spare parts are supplied directly from the regional warehouse to the wind farm, the unit freight cost c... L The cost is 40 yuan, and the downtime is T. L The downtime is 4 hours. Based on the unit's full-load operation and the electricity price in each central storage area, estimate the hourly downtime cost E of the wind turbine in central storage A. i The hourly downtime cost E in central warehouse B is 124 yuan. i The price is 90 yuan. The contract stipulates that the regional warehouse supply chain reliability R... L minimum value It is 0.40, time For 4 hours; Central warehouse supply chain reliability R L，C minimum value It is 0.70, time The lead time is 24 hours. The unit price, lead time, and Poisson distribution parameters of each spare part in different regional warehouses are shown in Table 1.

[0214] Table 1: Parameters of Four Spare Parts

[0215]

[0216]

[0217] Figure 4 These are example diagrams illustrating different schemes according to embodiments of the present disclosure.

[0218] In this embodiment, the model built using this disclosure is used to obtain five inventory schemes. The cost and supply chain reliability of each scheme are shown in Table 2.

[0219] Table 2: Results of the Five Inventory Strategies

[0220] parameter Option 1 Option 2 Option 3 Option 4 Option 5 <![CDATA[w1]]> 0.00 0.25 0.50 0.75 1.00 <![CDATA[w2]]> 1.00 0.75 0.50 0.25 0.00 TC 22467 14146 12346 11262 10419 W 6159.9 7474.9 8381.9 9842.4 14499 <![CDATA[R L ]]> 0.7509 0.6954 0.6598 0.5803 0.5304 <![CDATA[R L，c ]]> 0.9871 0.9662 0.9606 0.9523 0.8767

[0221] The non-dominated solutions of the 5 schemes are as follows Figure 4 As shown, the minimum cost for the wind turbine owner is 6159.9 yuan, and the cost for the wind turbine manufacturer is 22467 yuan. The minimum cost for the wind turbine manufacturer is 10419 yuan, and the cost for the wind turbine owner is 14499 yuan. Different inventory plans can be generated under different time and cost conditions depending on the weights of the two objective functions. For ease of explanation, only five plans are shown again, and these five plans should not be construed as limiting this disclosure.

[0222] Table 3: Inventory Strategies for Wind Turbine Manufacturers to Minimize Costs

[0223]

[0224] Table 3 lists the inventory strategy that minimizes the cost for wind turbine manufacturers while ensuring supply chain reliability to meet contractual requirements. This strategy is suitable for general customers. For customers with long-term strategic partnerships, an inventory strategy with a higher weighting for wind turbine owner costs can be chosen.

[0225] Figure 5 This is a cost example diagram showing different lead times.

[0226] As mentioned above, existing inventory planning model studies typically only consider lead times as integer multiples of the inspection cycle. This disclosure will verify the practical application effect of the model when the lead time is less than the inspection cycle. An integer multiple of the inspection cycle (e.g., one time) will be used as the lead time. In this case, the approximate lead time will be the same as the inspection cycle. Using the parameters listed in Tables 1 to 3, five optimal inventory plans will be used to calculate two types of costs under the actual lead time parameter. The difference between these costs and the minimum cost calculated using the actual lead time parameter will be compared.

[0227] Because the inspection cycle is longer than the actual lead time, all other things being equal, when the approximate lead time equals the inspection cycle, wind turbine manufacturers need to stock more spare parts; and wind turbine owners need to wait longer after shutdown. Therefore, Figure 5 The actual lead time cost curve shown is located to the left of the approximate lead time cost curve (i.e., the control cost curve). Under the five different weights, the total cost, considering the actual lead time, is lower than the cost of the control group using the approximate lead time when the total cost is minimized.

[0228] It should be noted that the specific embodiments of this disclosure have been described above with reference to the wind turbine manufacturer's inventory system. However, the concept of this disclosure is not limited to the wind turbine manufacturer's inventory system, nor is it limited to a two-level inventory structure such as a central warehouse and regional warehouses. For example, modifications to more-level inventory structures and other exemplary embodiments are intended to be included within the scope of this disclosure.

[0229] According to one or more aspects of this disclosure, the provided multi-level inventory control method and control apparatus calculate costs and supply chain reliability based on steady-state probabilities, and implement a non-dominated multi-level inventory strategy with the goal of minimizing manufacturer and owner costs and supply chain reliability as a constraint.

[0230] Furthermore, according to one or more aspects of this disclosure, the multi-level inventory control method and control device decompose the state transition probability into: the state transition probability of the ending inventory of the regional warehouse, the state transition probability of the ending inventory of the central warehouse, and the state transition probability of the regional warehouse before and after arrival. After calculating the three types of probabilities separately, the transition probability of spare part k in all warehouses within subsystem i is obtained, and then the steady-state probability of spare part k in all warehouses within subsystem i is obtained. By decomposing the conditional probability, the computational complexity of the state transition probability is simplified, and the curse of dimensionality is alleviated.

[0231] Furthermore, unlike the conventional assumptions about lead time in the prior art (e.g., lead time is greater than or equal to the inspection period, or lead time is assumed to be 0), according to one or more aspects of this disclosure, the multi-level inventory control method and control device perform steady-state analysis of the multi-level inventory system when the lead time is less than the inspection period, so that the resulting multi-level inventory strategy is more in line with the actual situation.

[0232] Figure 6 This is a block diagram illustrating a control device for a multi-level inventory according to an embodiment of the present disclosure.

[0233] Reference Figure 6 The multi-level inventory control device 100 according to embodiments of the present disclosure includes a parameter acquisition unit 110, a probability determination unit 120, and a strategy determination unit 130. The structure and reference of the multi-level inventory... Figure 1 The described inventory structures are the same or similar, so redundant descriptions are omitted here.

[0234] The parameter acquisition unit 110 is configured to check the inventory level, stockout level, and inventory level of each central warehouse, as well as the stockout level of each regional warehouse, according to the inspection cycle. The parameter acquisition unit 110 is configured to execute reference... Figure 2 Step S10 is described, so redundant descriptions are omitted here.

[0235] The probability determination unit 120 is configured to determine the steady-state probability of the inventory state based on the inventory level of each central warehouse, the stockout level of each central warehouse, the inventory level of each regional warehouse, and the stockout level of each regional warehouse. The steady-state probability characterizes the inventory state when the multi-level inventory reaches a steady state. The probability determination unit 120 is configured to execute a reference... Figure 2 and Figure 3 Step S20 is described, so redundant descriptions are omitted here.

[0236] The strategy determination unit 130 is configured to determine the inventory strategy corresponding to the plurality of central warehouses and the plurality of regional warehouses based on the steady-state probability of the inventory status. The strategy determination unit 130 is configured to execute a reference... Figure 2 and Figure 3Step S30 is described, so redundant descriptions are omitted here.

[0237] According to various embodiments of this disclosure, apparatus (e.g., modules or their functions) or methods can be implemented by programs or instructions stored in a computer-readable storage medium. When such instructions are executed by a processor, the processor can perform a function corresponding to the instruction or a method corresponding to the instruction (e.g., a multi-level inventory control method). At least a portion of a module can be implemented (e.g., executed) by a processor. At least a portion of a programmed module can include modules, programs, routines, instruction sets, and procedures for performing at least one function.

[0238] The multi-level inventory control method according to embodiments of this disclosure can be implemented by software, hardware, or a combination thereof. The hardware device can be implemented by one or more software modules for performing the operations of the various embodiments of this disclosure.

[0239] The modules or programming modules disclosed herein may include at least one of the aforementioned components, with some components omitted or others added. The operations of the modules, programming modules, or other components may be executed sequentially, in parallel, cyclically, or probingly. Furthermore, some operations may be executed in a different order, may be omitted, or may be extended with other operations.

[0240] An exemplary embodiment of the present disclosure may provide a controller for a multi-level inventory, which may include a processor (not shown) and a memory (not shown), wherein the memory stores a computer program that, when executed by the processor, implements the multi-level inventory control method as described in the exemplary embodiment above.

[0241] The multi-level inventory control method and multi-level inventory control device according to embodiments of the present disclosure can optimize inventory configuration and effectively reduce costs.

[0242] While some exemplary embodiments of this disclosure have been described, those skilled in the art will understand that modifications may be made to these embodiments without departing from the principles and spirit of this disclosure, which are defined by the claims and their equivalents. For example, technical features of different embodiments may be combined.

[0243] index:

[0244] i: Central repository number

[0245] j: Regional library number

[0246] k: Spare part number

[0247] Parameter list:

[0248] N CNumber of central warehouses

[0249] N L，i : Number of regional libraries corresponding to central library i

[0250] N M，i : Number of spare parts types in the regional warehouse corresponding to central warehouse i

[0251] N S Total number of spare parts for the system

[0252] T: Inspection cycle

[0253] T P，k Lead time of spare part k

[0254] s: Ordering point

[0255] S: Maximum inventory level

[0256] U: Beginning inventory for one inspection interval

[0257] U′: Ending inventory for one inspection interval

[0258] X: Quantity out of stock

[0259] λ: Average quantity of spare parts required during the corresponding period

[0260] Y: Quantity demanded

[0261] Q: The quantity must be met within the time stipulated in the contract.

[0262] m: Quantity allocated

[0263] w k Unit price of spare part k

[0264] P i，k : Inventory status transition probability of central warehouse i and its regional warehouse spare parts k

[0265] π i，k Steady-state probability of inventory status of central warehouse i and its regional warehouse spare parts k

[0266] Average transfer quantity of spare parts k before arrival at the central warehouse

[0267] Average transfer quantity of spare parts k after arrival in the central warehouse

[0268] n i，j，k The average inventory quantity of spare parts k in the regional warehouse corresponding to central warehouse i before delivery.

[0269] · i，j，k : Parameters of spare parts k in regional library j corresponding to central library i.

[0270] ·i，k Parameters of spare parts k in the central warehouse i

[0271] Parameters of central warehouse i before the arrival of regional warehouse j and spare parts k.

[0272] Parameters after the arrival of central warehouse i and corresponding regional warehouse j spare parts k.

[0273] Parameters of spare parts k before arrival at the central warehouse

[0274] Parameters after the central warehouse i spare parts k arrive

[0275] ·TC: Total Cost for Wind Turbine Manufacturers

[0276] TC min Minimum total cost for wind turbine manufacturers

[0277] TC max Maximum total cost for wind turbine manufacturers

[0278] Normalized total cost of all spare parts for wind turbine manufacturers

[0279] c P，i，k Central warehouse i spare parts k unit supplier allocation cost

[0280] c D，i，k Central warehouse spare parts k unit allocation cost

[0281] c L Regional warehouse spare parts unit transfer cost

[0282] c S，i，k The cost per order for spare parts k in central warehouse i and all its regional warehouses.

[0283] c T，i，k The unit transportation cost of spare parts k for central warehouse i and all its regional warehouses.

[0284] c H，i，k Unit holding cost within the inspection cycle of central warehouse i and its corresponding regional warehouse spare parts k

[0285] C P，i，k The cost of spare parts k being transferred from the supplier to the central warehouse i for wind farms.

[0286] C D，i，k The cost of transferring spare part k from the central warehouse i to its corresponding wind farm.

[0287] C L，i，k The cost of transferring spare parts k from all regional warehouses under the central warehouse i to their corresponding wind farms.

[0288] C S，i，k The ordering cost of spare parts k for central warehouse i and all its regional warehouses.

[0289] C T，i，k Spare parts ordering and shipping costs for central warehouse i and all its regional warehouses k

[0290] C H，i，k The total holding cost of central warehouse i and its corresponding regional warehouse spare parts k

[0291] C H，i，j，k : The holding cost of spare parts k in regional library j corresponding to central library i.

[0292] C H，C，i，k Total holding cost of spare parts k in the central warehouse i

[0293] W: Total Cost of Ownership for Wind Turbines

[0294] W min Minimum total cost for wind turbine owners

[0295] W max Maximum total cost for wind turbine owners

[0296] Normalized total cost of wind turbine owners

[0297] E i : The hourly power generation loss of the wind farm corresponding to central reservoir i

[0298] T S，i，k : Spare parts k are transferred from the supplier to the central warehouse i. Downtime of the wind farm

[0299] T C，i，k The downtime when spare part k is transferred from the central warehouse i to the corresponding wind farm.

[0300] T L Downtime when spare parts are transferred from the regional warehouse to the corresponding wind farm

[0301] W S，i，k The total downtime cost of spare part k being transferred from the supplier to the central warehouse i corresponds to the wind farm's total downtime cost.

[0302] W C，i，k The total downtime cost of spare parts k being allocated from the central warehouse i to the corresponding wind farm.

[0303] W L，i，k The total downtime cost of spare parts k being transferred from all regional warehouses under central warehouse i to its corresponding wind farm.

[0304] Downtime specified for central warehouse supply chain reliability

[0305] Downtime specified for regional warehouse supply chain reliability

[0306] R i，k : Supply chain reliability of the regional warehouse spare parts k corresponding to central warehouse i

[0307] R i，L，k Supply chain reliability of central warehouse i and its corresponding regional warehouse spare parts k

[0308] R L Supply chain reliability for all spare parts in all regional warehouses

[0309] R L，C The supply chain reliability of all spare parts corresponding to all regional and central warehouses.

[0310] The contract specifies the supply chain reliability threshold for all spare parts in the regional warehouse.

[0311] The contract specifies the supply chain reliability thresholds for all spare parts in regional and central warehouses.

[0312] Inventory plan for all spare parts.

Claims

1. A method of controlling a multi-level inventory, wherein, The structure of the multi-level inventory includes a plurality of central warehouses storing spare parts allocated from a supplier and a plurality of regional warehouses storing the spare parts allocated from the plurality of central warehouses, each of the plurality of central warehouses radiates at least one of the plurality of regional warehouses, and the control method includes: checking an inventory amount of each central warehouse, an out-of-stock amount of each central warehouse, an inventory amount of each regional warehouse, and an out-of-stock amount of each regional warehouse according to a check period; determining an inventory state steady state probability based on the inventory amount of each central warehouse, the out-of-stock amount of each central warehouse, the inventory amount of each regional warehouse, and the out-of-stock amount of each regional warehouse, the inventory state steady state probability being used to represent an inventory state when the multi-level inventory reaches a steady state; and determining an inventory strategy corresponding to the plurality of central warehouses and the plurality of regional warehouses based on the inventory state steady state probability, wherein the step of determining the inventory state steady state probability includes determining an inventory state transition probability based on the inventory amount of each central warehouse, the out-of-stock amount of each central warehouse, the inventory amount of each regional warehouse, and the out-of-stock amount of each regional warehouse, the inventory state transition probability being used to represent an inventory change of the spare parts in the multi-level inventory, and determining the inventory state steady state probability based on the inventory state transition probability, wherein the check period is divided into a pre-arrival stage and a post-arrival stage based on a pre-set lead time, the out-of-stock amount of each regional warehouse includes an out-of-stock amount of each regional warehouse before arrival and an out-of-stock amount of each regional warehouse after arrival, the inventory state transition probability includes an end-of-period inventory state transition probability of each regional warehouse, an end-of-period inventory state transition probability of each central warehouse, and a pre-arrival-post-arrival inventory state transition probability of each regional warehouse, and wherein the step of determining the inventory state transition probability includes obtaining the end-of-period inventory state transition probability of each regional warehouse according to the inventory amount of each regional warehouse, the out-of-stock amount of each regional warehouse before arrival, and the out-of-stock amount of each regional warehouse after arrival, obtaining the end-of-period inventory state transition probability of each central warehouse according to the inventory amount of the central warehouse and the out-of-stock amount of the regional warehouses radiated by the central warehouse before arrival and after arrival, obtaining the pre-arrival-post-arrival inventory state transition probability of each regional warehouse according to the inventory amount of each regional warehouse, the out-of-stock amount of each regional warehouse before arrival, and the out-of-stock amount of each regional warehouse after arrival, and determining the inventory state transition probability based on the end-of-period inventory state transition probability of each regional warehouse, the end-of-period inventory state transition probability of each central warehouse, and the pre-arrival-post-arrival inventory state transition probability of each regional warehouse.

2. The control method according to claim 1, characterized by, The inventory strategy includes a check point and a maximum inventory amount of each central warehouse and each regional warehouse, wherein when the inventory amount of any central warehouse or any regional warehouse is lower than the corresponding check point, the inventory amount of the any central warehouse or the any regional warehouse is increased to the corresponding maximum inventory amount by increasing the spare parts.

3. The control method according to claim 1, characterized by, The step of determining the inventory strategy corresponding to the plurality of central warehouses and the plurality of regional warehouses based on the inventory state steady state probability includes: obtaining a supply chain reliability and a cost based on at least one of the inventory state transition probability and the inventory state steady state probability; and constructing an objective function based on the cost, with the supply chain reliability as a constraint condition, and determining an inventory strategy corresponding to the plurality of central warehouses and the plurality of regional warehouses by minimizing the objective function.

4. The control method according to claim 3, characterized by During each inspection period, the inventory amount of each regional warehouse or the inventory amount of each central warehouse includes an initial inventory amount corresponding to the start time of the inspection period, the inspection period is divided into a pre-arrival stage and a post-arrival stage based on a pre-set lead time, and the supply chain reliability is obtained by the following steps: determining the quantity satisfied within a predetermined time of each regional warehouse based on the demand amount of each regional warehouse before arrival, the demand amount of each regional warehouse after arrival, the initial inventory amount of each regional warehouse, and the maximum inventory amount of each regional warehouse; for each central warehouse, obtaining the quantity satisfied within a predetermined time of the central warehouse based on the initial inventory amount of the central warehouse, the maximum inventory amount of the central warehouse, and the quantity satisfied within a predetermined time of the regional warehouses radiated by the central warehouse; and obtaining the supply chain reliability based on the inventory state steady state probability, the quantity satisfied within a predetermined time of each regional warehouse, and the quantity satisfied within a predetermined time of each central warehouse.

5. The control method according to claim 4, characterized by The cost includes a manufacturer total cost and an owner total cost of the spare parts, wherein the manufacturer total cost includes at least one of a supplier allocation cost, an allocation cost of each central warehouse, an allocation cost of each regional warehouse, an ordering cost, a transportation cost, and a holding cost.

6. A computer readable storage medium characterized by, The computer readable storage medium stores instructions or programs which, when executed by a processor, implement the control method according to any one of claims 1 to 5.

7. A control device for a multi-level inventory, wherein, The multi-level inventory structure includes a plurality of central warehouses storing spare parts allocated from a supplier and a plurality of regional warehouses storing the spare parts allocated from the plurality of central warehouses, each of the plurality of central warehouses radiates at least one of the plurality of regional warehouses, and the control device includes: a parameter acquisition unit configured to inspect an inventory amount of each central warehouse, a shortage amount of each central warehouse, an inventory amount of each regional warehouse, and a shortage amount of each regional warehouse according to an inspection period; a probability determination unit configured to determine an inventory state steady state probability based on the inventory amount of each central warehouse, the shortage amount of each central warehouse, the inventory amount of each regional warehouse, and the shortage amount of each regional warehouse, the inventory state steady state probability being used to represent an inventory state when the multi-level inventory reaches a steady state; and a strategy determination unit configured to determine an inventory strategy corresponding to the plurality of central warehouses and the plurality of regional warehouses based on the inventory state steady state probability, wherein the probability determination unit is configured to determine an inventory state transition probability based on the inventory amount of each central warehouse, the shortage amount of each central warehouse, the inventory amount of each regional warehouse, and the shortage amount of each regional warehouse, the inventory state transition probability being used to represent an inventory change of the spare parts in the multi-level inventory, and determine the inventory state steady state probability based on the inventory state transition probability, The inventory state transition probability includes an end-of-period inventory state transition probability of each regional warehouse, an end-of-period inventory state transition probability of each central warehouse, and a pre-delivery-to-post-delivery inventory state transition probability of each regional warehouse, and The probability determination unit is further configured to: obtain the end-of-period inventory state transition probability of each regional warehouse according to the inventory quantity of each regional warehouse, the pre-delivery shortage quantity of each regional warehouse, and the post-delivery shortage quantity of each regional warehouse; for each central warehouse, obtain the end-of-period inventory state transition probability of the central warehouse according to the inventory quantity of the central warehouse and the pre-delivery shortage quantity and the post-delivery shortage quantity of the regional warehouses radiated by the central warehouse; obtain the pre-delivery-to-post-delivery inventory state transition probability of each regional warehouse according to the inventory quantity of each regional warehouse, the pre-delivery shortage quantity of each regional warehouse, and the post-delivery shortage quantity of each regional warehouse; and determine the inventory state transition probability based on the end-of-period inventory state transition probability of each regional warehouse, the end-of-period inventory state transition probability of each central warehouse, and the pre-delivery-to-post-delivery inventory state transition probability of each regional warehouse.

8. The control device of claim 7, wherein The inventory strategy includes a check point and a maximum inventory quantity of each central warehouse and each regional warehouse, wherein when the inventory quantity of any central warehouse or any regional warehouse is lower than the corresponding check point, the inventory quantity of the any central warehouse or the any regional warehouse is increased to the corresponding maximum inventory quantity by increasing the spare parts.

9. The control device of claim 7, wherein The strategy determination unit is configured to: obtain a supply chain reliability and a cost based on at least one of the inventory state transition probability and the inventory state steady-state probability; and construct a target function based on the cost, determine an inventory strategy corresponding to the plurality of central warehouses and the plurality of regional warehouses by minimizing the target function with the supply chain reliability as a constraint condition.

10. The control device of claim 9, wherein During each inspection period, the inventory quantity of each regional warehouse or the inventory quantity of each central warehouse includes an initial inventory quantity corresponding to the start time of the inspection period, the inspection period is divided into a pre-delivery stage and a post-delivery stage based on a pre-set lead time, and the strategy determination unit is further configured to: determine a quantity satisfied within a predetermined time of each regional warehouse based on a pre-delivery demand quantity of each regional warehouse, a post-delivery demand quantity of each regional warehouse, an initial inventory quantity of each regional warehouse, and a maximum inventory quantity of each regional warehouse; for each central warehouse, obtain a quantity satisfied within a predetermined time of the central warehouse based on an initial inventory quantity of the central warehouse, a maximum inventory quantity of the central warehouse, and the quantity satisfied within a predetermined time of the regional warehouses radiated by the central warehouse; and obtain the supply chain reliability based on the inventory state steady-state probability, the quantity satisfied within a predetermined time of each regional warehouse, and the quantity satisfied within a predetermined time of each central warehouse. ​ 11. The control device of claim 10, wherein The cost includes a manufacturer total cost and an owner total cost of the spare part, wherein the manufacturer total cost includes at least one of a supplier allocation cost, an allocation cost of each central warehouse, an allocation cost of each regional warehouse, an ordering cost, a transportation cost, and a holding cost. The cost includes a manufacturer total cost and an owner total cost of the spare part, wherein the manufacturer total cost includes at least one of a supplier allocation cost, an allocation cost of each central warehouse, an allocation cost of each regional warehouse, an ordering cost, a transportation cost, and a holding cost. The cost includes a manufacturer total cost and an owner total cost of the spare part

Images