Expiration Near Analysis Module for Medical Inventory Management

Abstract

A method, apparatus, and system for determining items approaching expiration are provided. Historical data of items is received. The near-expiration analysis module is trained using training data, and the trained near-expiration analysis module is generated. The trained near-expiration analysis module is configured to receive item data regarding an item and generate a near-expiration prediction for one or more items regarding the item data. The bar includes a platform at the distal end of the bar. The platform is configured to contact an item provided within the housing. The sensor is configured to generate a signal indicating the filling level of the housing based on the contact between the item provided within the housing and the platform. An operation is performed to avoid the item being unused and exceeding the target date.

Description

Technical Field

[0001] This application claims the priority of a U.S. provisional application filed on June 23, 2022, with application number 63 / 354915 and invention title "SOON-TO-EXPIRE ANALYSIS MODELS FOR MEDICAL INVENTORY MANAGEMENT", and all its contents are incorporated herein by reference.

[0002] The subject matter described in this disclosure generally relates to data processing, and in particular to soon-to-expire (STE) analysis and dynamic analysis models for medical inventory management.

Background Art

[0003] Modern medical inventory management software applications provide various solutions for monitoring the supply and use of stored pharmaceuticals. These types of software applications are integrated with usage management software applications and exist within an overall pharmacy management software suite, and are deployed in various medical environments such as pharmacies, clinical trial labs, and medical facilities to track the storage, distribution, consumption, and disposal of various pharmaceuticals, equipment, and other consumables. In many cases, medical inventory management software applications are implemented to reduce operation costs and waste. For example, many medical inventory management software applications are configured to track the shipping, delivery, storage, prescription, dispensing, administration, and disposal of individual pharmaceuticals and generate corresponding one or more electronic records (e.g., cost, lot number, expiration date, patient name).

Summary of the Invention

[0004] Systems, methods, computer program products, and apparatuses are provided for medical inventory management having a soon-to-expire (STE) analysis function. In some embodiments, an inventory controller may apply an STE analysis model to identify items that are likely to expire unused at a current stocking location. Accordingly, the inventory controller may take one or more corrective actions to avoid one or more items from expiring unused. For example, the one or more items (or a predetermined amount of the one or more items) may be repositionable to different locations where they are likely to be consumed before their expiration date. Alternatively and / or additionally, the inventory controller may prioritize dispensing of one or more items from a first location where the one or more items are likely to expire unused over dispensing of one or more items from a second location where the one or more items are less likely to expire unused.

[0005] Embodiments of the subject matter protected can include articles that are operable to cause one or more machines (such as computers, etc.) to operate in a manner consistent with the description provided in this disclosure and that include a specifically embodied machine-readable medium for implementing one or more of the described features. Similarly, the described computer system may include one or more processors and one or more memories coupled to the one or more processors. The memory, which may include a non-transitory computer-readable storage medium or a machine-readable storage medium, may embody, encode, store, etc., one or more programs for causing the one or more processors to execute one or more operations described in this disclosure. A computer-implemented method consistent with one or more embodiments of the subject matter protected may be implemented by one or more data processors present in a single computing system or multiple computing systems. Such multiple computing systems may be connected via one or more connections and can exchange data and / or commands or other instructions, etc., and the one or more connections include connections over a network (such as the Internet, a wireless wide area network, a local area network, a wide area network, a personal area network, a peer-to-peer network, a mesh network, a wired network, etc.) such as a direct connection between one or more computing systems.

[0006] Details of one or more variations of the subject matter protected described in this disclosure are set forth in the accompanying drawings and the following description. Other features and advantages of the subject matter protected described in this disclosure will become apparent from the description and drawings, as well as from the claims. The specific features of the subject matter protected disclosed herein are described for illustrative purposes in connection with a medical inventory that includes pharmaceuticals, devices, and consumables, but it will be readily understood that there is no intention to limit such features. The claims that follow this disclosure are intended to define the scope of the subject matter protected.

[0007] The accompanying drawings are incorporated herein and form a part hereof, showing specific aspects of the subject matter disclosed herein and, together with the description, serving to explain some of the principles of the disclosed embodiments.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2A

Figure 2B

Figure 3

Figure 4A

Figure 4B

Figure 5

Modes for Carrying Out the Invention

[0009] In practice, like reference numerals indicate like structures, features, and elements.

[0010] One of the primary objectives of a medical inventory management software application is to reduce operating costs and waste through the storage, distribution, consumption, and disposal of various pharmaceuticals, devices, and other consumables. For example, a medical inventory management software application deployed at a healthcare facility may perform an analysis to predict within the current inventory of medical supplies at that location, supplies that may expire, or supplies that may be moved to another location for use within a certain period. In some embodiments, once an item is moved or identified for movement, the location or item may be excluded from further near-expiry analysis. The exclusion may be time-limited (e.g., 10 days after movement are excluded) or event-limited (e.g., excluded until an inventory update event such as restocking). Medical supplies identified as near expiry (STE) are medical supplies expected to expire within a certain period. These near-expiry items are typically removed from storage before being discarded from their current location or having their inventory reduced. In the latter case, the medical supplies for which inventory has been removed may, for the removed medical supplies, be replenished in inventory at another location or returned for partial refund.

[0011] As used in this disclosure, an item may expire (or an expiration date may be set) based on the period during which the item is expected to continue to be used effectively and / or safely. It should be understood that the expiration date may vary depending on the item. Some items, such as certain pharmaceuticals, e.g., pharmaceuticals that are compounded and / or repackaged at a local site (such as a hospital, pharmacy, or other permitted facility), may have a "short-term" expiration date, i.e., the expiration date of these items is a very short period (not months or years, but hours, days, weeks, etc.), and the item is considered near expiration within the expiration date. Also, the cost of an item may be very high, in which case it may be necessary to move and store the item at the location where it is most likely to be used before the expiration date. On the other hand, in the case of low-cost items, the cost of moving the item to another location to replenish inventory may exceed the cost of the item itself. In this case, the time from when the item is considered near expiration until it is moved to a new location may be shortened and approach the expiration date of the item. In the case of items that are out of stock, the need to avoid waste due to the expiration date may take precedence over the cost of the item. Therefore, the time from when an out-of-stock item is considered near expiration until it is moved to a new location may be shorter than in the case of an item with sufficient inventory.

[0012] However, conventional medical inventory management software applications are insufficient for several reasons. For example, conventional medical inventory management software applications perform soon-to-expire (STE) analysis based on the earliest expiration date of the current inventory at a particular location. However, in addition to the fact that clinicians are not obligated to reduce inventory with the earliest expiration dates, and the verification and updating of the earliest expiration date of the current inventory are not always performed as part of the inventory replenishment workflow, the earliest expiration date is often not the correct value. Also, the results of the soon-to-expire (STE) analysis impede efforts to reduce operational costs and waste by performing after-the-fact measures such as the disposal of expired medical supplies and medical supplies with an expiration date that is too close to be replenished to other locations.

[0013] In some embodiments, the inventory management device may perform soon-to-expire (STE) analysis by applying a soon-to-expire (STE) analysis model trained to identify one or more items that are unlikely to be used at the current location, where the items are pharmaceuticals, devices, and other consumables, etc. The inventory management device may apply various embodiments of the soon-to-expire (STE) analysis model, such as a heuristic-based model or a hybrid model that combines one or more heuristic models and a machine learning model. In some cases, the soon-to-expire (STE) analysis model may be trained based on historical data points associated with a medical facility and / or one or more similar medical facilities. The historical data points captured by the soon-to-expire (STE) analysis model may be associated with one or more of dispensing events, locations, and inventory levels for various items stored as inventory in the medical facility. Examples of such data points may include the quantity of items removed during the current period for a particular item stored in the medical facility, the inventory level of the item during the current period (e.g., measured value), the quantity of items consumed during the previous period, and the interval until the earliest expiration date associated with the item during the current period.

[0014] In some embodiments, an inventory management apparatus that applies an expiration approaching (STE) analysis model may pre-identify one or more items that are unlikely to be used at the current location. Therefore, if it is determined that an item is unlikely to be used, the inventory management apparatus may perform various improvement operations to prevent the item from becoming obsolete at the current location. For example, the inventory management apparatus may apply an expiration approaching (STE) analysis model and determine that a particular item is likely to expire unused at a first location rather than at a second location. In this case, the inventory management apparatus may relocate the item from the first location to the second location. Alternatively and / or additionally, the inventory management apparatus may prioritize dispensing of an item from a first location where the item is likely to expire unused over a second location where the item is less likely to expire unused. In some cases, when the quantity of items at the first location that are present and likely to expire unused is greater than the quantity of items at the second location that are present and likely to expire unused, the first location may be prioritized as the dispensing location. Also, in some cases, the inventory management apparatus may apply an expiration approaching (STE) analysis model, determine the quantity of items that are likely to remain unused, and adjust the quantity of items ordered for inventory replenishment according to the determination result. Further, for future inventory replenishment, the reorder point and the maximum (or minimum) quantity of items at that location may be adjusted, and furthermore, it may be proposed to completely remove the items from the storage location.

[0015] In some embodiments, the near-expiration (STE) analysis model can be trained to recognize the association between individual items and different storage locations. In particular, depending on the item and the location where the item is stored, it may exhibit a combination of characteristics that affect the likelihood that the item will expire unused at that location. Thus, the near-expiration (STE) analysis model can be trained to identify this combination of characteristics, which is for identifying items that are likely to expire unused and their quantities at a specific storage location. In this context, the term "location" is used in the same sense as the term "storage location" and may refer to a medical environment at any level of specificity. For example, the location where an item is stored can be a specific device that houses the item (e.g., a dispensing cabinet, a shelf, etc.), the building (or part of the building) where the device is located, a department, a room within the department, a unit, or the care area of the facility associated with the item, the facility itself, the geographical area where the facility is located, a network including the facility and one or more other facilities, etc.

[0016] The likelihood that an item will expire unused at its current storage location can change over time due to changes in various factors, such as the rate at which the item is used at that location, the types of items stored at that location, the cost of the item, the expiration pattern of the item, the movement pattern of the item, etc. Thus, in some embodiments, the inventory management device may subject the near-expiration (STE) analysis model to periodic updates to account for changes in the factors that affect the near-expiration (STE) analysis of the items stored in a particular medical facility. For example, in some cases, after the near-expiration (STE) analysis model is trained at a first time t0, the near-expiration (STE) analysis model can be updated at one or more consecutive time points. For example, at a second time t1, the near-expiration (STE) analysis model may be updated by training based on the data points from the first time t0 to the second time t1.

[0017] FIG. 1 is a system diagram showing an illustration of a medical inventory management system 100 based on several embodiments. Referring to FIG. 1, the medical inventory management system 100 may include an inventory management device 110, a client device 120, and one or more medical locations 130. As shown in FIG. 1, the inventory management device 110, the client device 120, and the one or more medical locations 130 may be communicatively coupled via a network 140. The client device 120 may be a specially configured processor-based device including, for example, a point of care unit (PCU), a smartphone, a tablet computer, a wearable device, a desktop computer, a laptop computer, a workstation, etc. Also, the network 140 may be a wired and / or wireless network including, for example, a Public Land Mobile Network (PLMN), a Local Area Network (LAN), a Virtual Local Area Network (VLAN), a Wide Area Network (WAN), the Internet, etc.

[0018] In some embodiments, the inventory management device 110 may apply a shelf-life expiration (STE) analysis model 115 to perform a shelf-life expiration (STE) analysis on items 135 stored at one or more locations 130. For example, the inventory management device 110 may apply the shelf-life expiration (STE) analysis model 115 to determine the likelihood that items 135 stored at one or more locations 130 are unused and past their expiration dates. Alternatively and / or additionally, the inventory management device 110 may apply the shelf-life expiration (STE) analysis model 115 to determine the quantity of items that are unused and past their expiration dates at one or more locations 130.

[0019] In some embodiments, the inventory management device 110 may apply an expiration approaching (STE) analysis model 115 to pre-judge that item 135 is likely to remain unused at its current location, and may perform various improvement operations. The improvement operations may be for preventing item 135 from becoming obsolete at its current location, or in some cases, for removing item 135 from inventory sufficiently early before its expiration date and replenishing it at another location. For example, if the inventory management device 110 determines that item 135 is likely to expire unused at the first location 130a but will be used before its expiration date at the second location 130b, item 135 may be relocated from the first location 130a to the second location 130b. In some cases, relocating item 135 from the first location 130a to the second location 130b may include moving item 135 from one decentralized dispensing location (e.g., the first dispensing table) to another dispensing location (e.g., the second dispensing table in the same facility or a different facility or a part of a facility). In other cases, relocating item 135 from the first location 130a to the second location 130b may include moving item 135 from a decentralized dispensing location (e.g., the first dispensing table) to a central dispensing location (e.g., a central pharmacy, a warehouse, etc.).

[0020] In some embodiments, the inventory management device 110 may generate one or more electronic records regarding the item to be relocated from the first location 130a to the second location 130b. For example, relocating item 135 from the first location 130a to the second location 130b may include transferring the custody obligation of item 135 from a first user associated with the first location 130a to a second user associated with the second location 130b. Thus, the inventory management device 110 may generate one or more electronic records to document a series of management obligations regarding item 135. In some cases, item 135 may be subject to specific regulatory requirements that require a specific series of management obligations. Therefore, the relocation from the first location 130a to the second location 130b may further include an intermediate location such as a pharmacy.

[0021] When item 135 is stored in multiple locations, if a particular item is likely to expire unused at location 130a rather than at location 130b, and / or if the quantity of items present at location 1 is likely to expire unused and is greater than the quantity of items present at location 2 that are likely to expire unused, item 135 may be dispensed at location 130a. For example, it is conceivable that location 130a is the first drawer of an automated dispensing unit and location 130b is the second drawer of the automated dispensing unit. The physical item 135 can be stored in both locations. When a clinician requests the dispensing of item 135, inventory management device 110 may apply the STE model (or consider the information generated by the STE model) to determine at which location to release and dispense item 135. Inventory management device 110 may permit dispensing at the location identified as having the highest likelihood of expiring closest to the current date. In some cases, inventory management device 110 applies the near-expiration (STE) analysis model 115 to determine the quantity of item 135 that is likely to remain unused and may adjust the quantity of item 135 ordered for inventory replenishment at location 130a and / or location 130b according to the determination result.

[0022] In some embodiments, when identifying items 135 that are unused and likely to expire at one or more locations 130, the inventory management device 110 may send one or more corresponding notifications 125 to the client device 120 associated with the one or more locations 130. For example, the inventory management device 110 may send a notification 125 having an instruction to review the expiration date of the items 135 stored at the first location 130a to the first location 130a, reduce the inventory of the items 135 from the first location 130a, and / or instruct to remove the items 135 from the first location 130a. In some cases, the instruction may further specify a predetermined quantity of the items 135 to be reduced and removed from the first location 130a and a predetermined quantity of the items 135 to be moved and stored at the second location 130b. In some cases, the inventory management device 110 may send the notification 125 at a specific time to ensure that the items 135 are sufficiently removed from the first location 130a well before the expiration date and the inventory of the items 135 is replenished and consumed at the second location 130b. For example, if the item 135 is a high-cost item that expires in a short period and is stored at the first location 130a having a history indicating few dispensing activities, the inventory management device 110 can send the notification 125 early and / or send the notification 125 relatively frequently.

[0023] In some embodiments, the soon - to - expire (STE) analysis model 115 can be trained to recognize the relationship between individual items, such as item 135, and different storage locations 130. In particular, depending on item 135 and one or more locations 130 where item 135 is stored, a combination of characteristics that affect the likelihood that item 135 will become unused and expire can be shown. For example, item 135 stored in the first location 130a may exhibit a different set of characteristics compared to item 135 stored in the second location 130b. Examples of these characteristics may include the rate at which item 135 is used at each location 130, the cost of item 135, the expiration - date pattern of item 135, the movement pattern of item 135, and the like. After training, the soon - to - expire (STE) analysis model 115 can identify item 135 well in advance as being likely to expire at the first location 130a, for example, identify item 135 that should be removed from inventory at the first location 130a and moved to the second location 130b where it can be consumed before the expiration date. This can achieve a significant reduction in both business costs and waste.

[0024] At least because the likelihood that item 135 will become unused and expire in the current storage location can change over time due to changes in various factors, in some embodiments, the inventory management device 110 may periodically update the soon - to - expire (STE) analysis model 115. The various factors are, for example, the rate at which an item is used at the location, the change in the rate at which an item is used at the location, the cost of the item, the expiration - date pattern of the item, the movement pattern of the item, and the like. For example, if the soon - to - expire (STE) analysis model 115 is trained at the first time t0, the inventory management device 110 may continuously update the soon - to - expire (STE) analysis model 115 at the second time t1 by training the soon - to - expire (STE) analysis model 115 based on data points between the first time t0 and the second time t1.

[0025] Changes in one or more of the foregoing factors can occur frequently because they can be caused by simple reasons such as changes in the packaging, manufacturing, storage conditions, and / or other characteristics of the item. For example, if item 135 is a composite item, a repackaged item, and / or an item that requires special storage (such as refrigeration, etc.), item 135 may be associated with a relatively short expiration date. By updating the soon-to-expire (STE) analysis model 115 through periodic retraining of the soon-to-expire (STE) analysis model 115, the soon-to-expire (STE) analysis model 115 can be made to accommodate the foregoing changes and maintain the accuracy of the soon-to-expire (STE) analysis.

[0026] In some embodiments, the soon-to-expire (STE) analysis model 115 may be trained to perform a soon-to-expire (STE) analysis for item 135 based on various data points, which may include, for example, the cost of item 135, the speed or usage rate of item 135, the quantity of item 135 removed, the current inventory level of item 135, the quantity of item 135 consumed in the previous period, the interval until the earliest expiration date associated with the current inventory of item 135, and the like. Other data points that may be incorporated into the soon-to-expire (STE) analysis of item 135 may include the packaging of item 135, the storage conditions, whether the item is a custom compound, and the like.

[0027] At least some of the aforementioned data points may be represented as categorical values. For example, the cost of item 135 may be represented as a first binary value indicating whether item 135 is a high-cost item or a low-cost item, the current inventory level of item 135 may be represented as a second binary value indicating whether item 135 is associated with a low inventory value or a high inventory value, and the usage rate of item 135 may be represented as a third binary value indicating whether item 135 has a low usage rate or a high usage rate. On the other hand, the quantity of item 135 removed due to reasons such as being obsolete or out of stock may be represented as a category selected from a plurality of categories regarding the removal percentage (for example, the ratio that the first quantity, which is the quantity of item 135 removed, occupies in the second quantity, which is the quantity of item 135 in the inventory at the start of that period).

[0028] Thresholds for categorical values representing the aforementioned data points, for example, whether it is a low or high unit cost, whether it is a low or high inventory dollar value, and whether it is a low or high usage rate, as well as percentiles of different categories regarding the removal rate, can be determined based on the corresponding error (such as mean absolute error (MAE) or different error metrics). For example, in some cases, the inventory management device 110 may determine a threshold for separating the high-value category and the low-value category based on the corresponding historical data from each of one or more locations 130 as part of the training of the soon-to-expire (STE) analysis model 115. Thus, for example, the high or low of the threshold can be specified as a percentile (for example, the 75th, 80th, or 85th) of an existing dataset associated with the minimum error (such as mean absolute error (MAE) or different error metrics).

[0029] In some embodiments, when the system identifies item 135 for relocation or inventory reduction, or when identifying location 130a, 130b as containing item 135 that is near expiration for relocation or inventory reduction, the system may secure the location until the relocation or inventory reduction occurs. For example, if a first location (e.g., a drawer, pocket, or bin) of an automated dispensing cabinet contains items that are considered near expiration (e.g., items within a predetermined number of days from a date predicted to be near expiration, items within a predetermined expiration confidence range, etc.), the automated dispensing cabinet may be configured to prevent further dispensing from the first location unless a clinician has made an access request for inventory reduction or relocation. The access request may be identifiable based on qualification information or other user identification information provided by the clinician accessing the automated dispensing cabinet. The access request may also be based on an action selected at the automated dispensing cabinet. For example, a clinician may activate a control element on a user interface and activate an inventory reduction mode or a relocation mode. In this mode, a request to unlock or lock a location is distinguishable from a dispensing request for a particular patient.

[0030] In some embodiments, the near - expiration (STE) analysis model 115 can be implemented as a heuristic model or as a hybrid model that combines one or more heuristic models and a machine - learning model. FIG. 2A is a schematic diagram showing an example of the near - expiration (STE) analysis model 115 implemented as a hybrid model 200, and FIG. 2B shows another example of the near - expiration (STE) analysis model 115 implemented as a heuristic model 250. In some cases, the inventory management device 110 may select to implement the near - expiration (STE) analysis model 115 that is relatively suitable for one or more locations 130. For example, the inventory management device 110 may select to implement the near - expiration (STE) analysis model 115 associated with a relatively low prediction error (such as mean absolute error (MAE) or different error metrics). For example, when the near - expiration (STE) analysis model 115 is implemented as a heuristic model, the inventory management device 110 may identify the one with the lowest prediction error from the options of heuristic models incorporating different combinations of data points and category values. The following Table 1 shows examples of heuristic models and corresponding data points.

[0031]

Table 1

[0032] The inventory management device 110 may evaluate the models as follows: providing data regarding a controlled group of items to each model, comparing the model's prediction for each item with the actual output or desired output, and based on the comparison results, selecting the model that has the highest success rate in predicting the actual output or desired output, thereby evaluating the models.

[0033] FIG. 4A is a schematic diagram showing an exemplary logic flow 400 of an expiration approaching (STE) analysis model 115 implemented as a heuristic model, which determines the percentage of the current inventory of items 135 at one or more locations 130 that are unused and likely to expire. In the example shown in FIG. 4A, the logic flow 400 may include a combination of factors such as current inventory level, unit price, inventory value, and earliest expiration date, and the combination of factors is used to determine the percentage of the current inventory of items 135 that are unused and likely to expire (i.e., the percentage that the items 135 occupy in the current inventory). As described above, the combination of factors can be selected as part of the training of the expiration approaching (STE) analysis model 115 based on being associated with the least error (e.g., error measured by mean absolute error (MAE) or different error metrics). Further, as shown in FIG. 4A, the expiration approaching (STE) analysis model 115 may apply one or more of the aforementioned category values associated with the items 135. It should be understood that the inventory management device 110 may perform one or more improvement actions based on the percentage of items 135 that are likely to remain unused, and the percentage may include, for example, the percentage related to the quantity corresponding to the inventory reduction and / or relocation of the items 135.

[0034] In the example of the logic flow 400 shown in FIG. 4A, at block 402, the expiration approaching (STE) analysis model 115 may first determine whether any quantity of items 135 remains at one or more locations 130. If some quantity of items 135 remains at one or more locations 130, then at block 404, the expiration approaching (STE) analysis model 115 may determine whether the items 135 are high-cost items or low-cost items. As described above, the threshold used to classify the items 135 as high-unit-price items or low-unit-price items can be determined as part of the training of the expiration approaching (STE) analysis model 115.

[0035] Also, in the example shown in FIG. 4A, when it is determined that item 135 is a high - unit - price item, the near - expiration (STE) analysis model 115 may determine that 100% of the current inventory of item 135 is likely to be unused and past its expiration date. Alternatively, when it is determined that item 135 is a low - unit - price item, the near - expiration (STE) analysis may, at block 406, determine whether item 135 is associated with a high inventory value (e.g., the unit price of the item multiplied by the quantity of the item in the inventory at that location, the holding cost (e.g., for dangerous pharmaceuticals, high - risk pharmaceuticals, or divertible pharmaceuticals, which is the holding cost multiplied by the quantity of the item in the inventory at that location), the cost required for replenishment or replacement of the item (e.g., the time required for reconstitution or ordering, which is multiplied by the quantity of the item in the inventory at that location)). If item 135 is associated with a high inventory value, the near - expiration (STE) analysis model 115 may determine that 45% of the current inventory of item 135 is likely to be unused and past its expiration date. However, even when item 135 is not associated with a high inventory value, the near - expiration (STE) analysis model 115 may, at block 408, determine that 45% of the current inventory of item 135 is likely to be unused and past its expiration date when the current inventory of item 135 is associated with the earliest expiration date. On the other hand, when item 135 is not associated with a high inventory value and does not have the earliest expiration date, the near - expiration (STE) analysis model 115 may determine that 0% of the current inventory of item 135 is likely to be unused and past its expiration date, which indicates that no inventory adjustment is necessary.

[0036] FIG. 4B is a schematic diagram showing another exemplary logic flow 450 of the near-expiration (STE) analysis model 115 based on some embodiments. The exemplary near-expiration (STE) analysis model 115 shown in FIG. 4B is configured to determine the withdrawal rate of items 135 stored at one or more locations 130 based on a combination of factors including the unit price of item 135 and the velocity of item 135. In some cases, the inventory management device 110 may perform one or more improvement actions based on this withdrawal rate, and the improvement actions may include, for example, reducing and / or reallocating the corresponding quantity of item 135 in inventory. Again, it should be understood that this combination of factors can be selected based on the corresponding error (e.g., mean absolute error (MAE) or different error metrics) in the training of the near-expiration (STE) analysis model 115. Also, each factor can be evaluated as a categorical value, and the threshold values for the categorical values (e.g., the threshold for separating high unit price and low unit price, the threshold for separating fast movement and slow movement, etc.) are selected as part of the training of the near-expiration (STE) analysis model 115 based on the corresponding error (e.g., mean absolute error (MAE) or different error metrics).

[0037] Referring again to FIG. 4B, the near-expiration (STE) analysis model 115 that implements the logic flow 450 may first determine whether item 135 is a high-value item or a low-value item before determining whether item 135 is a fast-moving item (i.e., an item that moves quickly) or a slow-moving item (i.e., an item that moves slowly). As shown in FIG. 4B, if it is determined that item 135 is a high-value and fast-moving item, the near-expiration (STE) analysis model 115 may specify that 0% of the current inventory of item 135 should be withdrawn. On the other hand, if it is determined that item 135 is a slow-moving and high-value item, the near-expiration (STE) analysis model 115 may specify that 60% of the current inventory of item 135 will be withdrawn. The near-expiration (STE) analysis model 115 may specify that 25% of the current inventory of item 135 will be withdrawn if item 135 is a fast-moving and low-value item, and may also specify that 56% of the current inventory of item 135 will be withdrawn if it is determined that item 135 is a slow-moving and low-value item.

[0038] When the near-expiration (STE) analysis model 115 is implemented as a hybrid model that combines one or more heuristic models and a machine learning model, the inventory management device 110 may train and deploy the one or more heuristic models and the machine learning model based on the quantity of data available for training the near-expiration (STE) analysis model 115. In some cases, the training data used to train the near-expiration (STE) analysis model 115 may be location-specific data. For example, the near-expiration (STE) analysis model 115 may be trained based on training data that includes historical data from the first location 130a. By doing so, the near-expiration (STE) analysis model 115 can be trained to recognize combinations of characteristics that affect the likelihood that item 135 will expire unused at the first location 130a. Also, once trained, the near-expiration (STE) analysis model 115 is applicable to current data from the first location 130a to determine whether the items 135 currently stored at the first location 130a are likely to expire unused. If the second location 130b is sufficiently similar to the first location 130a, the near-expiration (STE) analysis model 115 trained based on data from the first location 130a is also applicable to determine whether the items 135 currently stored at the second location 130b are likely to expire unused.

[0039] In some embodiments, when determining the state regarding an item approaching its expiration date based on the type of the item or clinical dispensing factors at other locations, the system can more efficiently or accurately determine the evaluation of the high or low moving speed. In such cases, the system may not only evaluate different models, but also evaluate various modelling pipelines to identify the optimal and accurate configuration for the site. Regarding optimization, not only accuracy but also the time and other resources required to generate the state regarding an item approaching its expiration date can be considered. For example, a pipeline that considers the speed of the item (high-speed movement and low-speed movement) and the unit cost may generate the state regarding an item approaching its expiration date of the test item using a speed of n (e.g., per second) or x resources (network communication, processing cycles, memory, etc.). If another configuration consumes less time or fewer resources, the system may select the alternative configuration.

[0040] For further illustration, the exemplary hybrid model 200 shown in FIG. 2A includes a machine learning model 210, a first heuristic model 220, and a second heuristic model 230, and each model has different requirements regarding the amount of training data. Therefore, when the amount of available training data meets the first threshold (e.g., when historical data of six months or more is available), the hybrid model 200 can be trained by training the machine learning model 210 based on the training data. When the amount of training data does not meet the first threshold and meets the second threshold (e.g., when data of less than six months and two months or more is available), the hybrid model 200 can be trained by training the first heuristic model 220 based on the training data. When the amount of training data does not meet the second threshold (e.g., when data of less than two months is available), the hybrid model 200 can be trained by training the second heuristic model 230 based on the training data.

[0041] FIG. 3 is a flowchart illustrating an example of a process for medical inventory management using near-expiration (STE) analysis, based on several embodiments. Referring to FIGS. 1-3, process 300 may be performed by an inventory management device 110 that applies a near-expiration (STE) analysis model 115 to determine whether an item 135 stored at one or more locations 130 is likely to be unused and expired at its current location.

[0042] In block 302, the inventory management device 110 may train a near-expiration (STE) analysis model to perform a near-expiration (STE) analysis for one or more locations. In some embodiments, the inventory management device 110 may train the near-expiration (STE) analysis model 115 to perform a near-expiration (STE) analysis for one or more locations 130. The near-expiration (STE) analysis model 115 may be implemented as a heuristic model (e.g., heuristic model 250), or a hybrid model (e.g., hybrid model 200) that combines one or more machine learning models and a heuristic model. Also, the near-expiration (STE) analysis model 115 may be trained based on historical data points associated with one or more locations 130. Examples of such data points may include, but are not limited to, for each item stored at one or more locations 130, such as item 135, the quantity of items removed during the current period, the inventory level of the item during the current period (e.g., a measured value), the quantity of items consumed in the previous period, the interval until the earliest expiration date associated with the item during the current period, etc. Through training, the near-expiration (STE) analysis model 115 can recognize combinations of characteristics that affect the likelihood that an item 135 is unused and expired at each of one or more locations. Examples of such characteristics may include, but are not limited to, the rate at which item 135 is used at each location 130, the cost of item 135, the expiration date pattern of item 135, the movement pattern of item 135, etc.

[0043] As described above, the near - expiration (STE) analysis model 115 can be implemented as one or more heuristic models and / or machine - learning models. Thus, it should be understood that all or part of one or more aspects of the described artificial intelligence can be implemented by a model that includes a machine - learning model. The training that the model undergoes may be supervised training, unsupervised training, reinforcement training, or a hybrid approach that employs multiple learning techniques to generate the model. The training of the model may include obtaining a series of training data and adjusting the characteristics of the model to obtain a desired model output. For example, three characteristics may be associated with one desired device state. In such a case, the training may include receiving the three characteristics as inputs to the model and, for each set of the three characteristics, adjusting the characteristics of the model such that the output device state matches the desired device state associated with the training data. In some cases, the training may be dynamic training, that is, the system may update the model using a series of events that have detectable properties and are used for adjusting the model.

[0044] In some cases, the near - expiration (STE) analysis model 115 may be an equation, an artificial neural network, a recurrent neural network, a convolutional neural network, a decision tree, and / or other machine - readable artificial - intelligence structures. The characteristics of the structure that can be used for adjustment during training may vary depending on the selected model. For example, if a neural network is the selected model, the characteristics may include input elements, network layers, node density, node activation thresholds, weights between nodes, weights of input values or output values, etc. When the model is implemented as an equation (such as a regression equation), the characteristics may include weights of input parameters, thresholds or limit values for evaluating output values, or criteria for selecting from a series of equations.

[0045] After the expiration approaching (STE) analysis model 115 is trained, retraining may be performed to improve or update the model to reflect additional data or specific operating conditions. For example, as described above, the inventory management device 110 may periodically update the expiration approaching (STE) analysis model to accommodate changes in factors that affect the expiration approaching (STE) analysis of items stored in a particular medical facility. The retraining may be based on one or more signals detected by the devices described in this disclosure and may be part of the methods described in this disclosure. When a specified signal is detected, the system may activate the training process to adjust the expiration approaching (STE) analysis model 115 as described above.

[0046] Further examples of machine learning and modeling functions that may be included in the above embodiments are described in "A survey of machine learning for big data processing" by Qiu et al. in "EURASIP Journal on Advances in Signal Processing (2016)", the content of which is incorporated herein by reference.

[0047] In block 304, inventory management device 110 may apply a trained soon - to - expire (STE) analysis model to identify items that are stored at one or more locations, unused, and likely to expire. The trained model may check the unit price, inventory quantity, number of days until the earliest expiration date, and usage rate of the item, and provide a soon - to - expire quantity for the item. For example, inventory management device 110 may apply a trained soon - to - expire (STE) analysis model 115 to identify items 135 that are unused and likely to expire at one or more locations 130. Inventory management device 110 may apply the soon - to - expire (STE) analysis model 115 periodically or when detecting a change in the inventory of one or more items such as item 135. Also, inventory management device 110 may apply the soon - to - expire (STE) analysis model 115 to identify items that are unused and may expire at each location or location group, and the location group may be locations within a facility, within the same geographical area, or within the same network. To apply the soon - to - expire (STE) analysis model 115, inventory management device 110 may provide information regarding at least a portion of the items currently stored at one or more locations 130, such as the current inventory level, unit price, inventory value, and earliest expiration date, to be incorporated into the soon - to - expire (STE) analysis model 115. The logic of the soon - to - expire (STE) analysis model 115 is applicable to process the current inventory information and generate an output that identifies items that are unused and likely to expire at the current location. In some cases, the output of the soon - to - expire (STE) analysis model 115 may include a list of items ranked by the likelihood that each item will expire at the current location, and as a result, one or more corrective actions to prevent the expiration of these items may be executed based on the list.

[0048] In block 306, inventory management device 110 may perform one or more improvement operations so as to prevent items from becoming unused and expiring. After training, the near-expiration (STE) analysis model 115 is applied to identify, for example, items 135 that are likely to expire at the first location 130a well in advance for improvement operations, and ensure that items 135 are consumed before their expiration date. The improvement operation may, for example, involve reducing inventory from the first location 130a and moving it to a second location 130b where it can be consumed before the expiration date. For example, when it is identified that items 135 are likely to become unused and expire at the first location 130a rather than at the second location 130b, inventory management device 110 may relocate the items 135 from the first location 130a to the second location 130b. Alternatively and / or additionally, when items 135 are likely to become unused and expire at the first location 130a rather than at the second location 130b, inventory management device 110 may prioritize dispensing of items 135 from the first location 130a rather than from the second location 130b. In some cases, inventory management device 110 may apply the trained near-expiration (STE) analysis model to determine the quantity of items 135 that are likely to remain unused at each of the first location 130a and the second location 130b. When the quantity of items 135 present at the first location 130a that are likely to become unused and expire is greater than the quantity of items 135 present at the second location 130b that are likely to become unused and expire, inventory management device 110 may relocate and / or prioritize dispensing a predetermined quantity of items 135 from the first location 130a. In some cases, inventory management device 110 may adjust the quantity of items 135 to order for inventory replenishment to the first location 130a and / or the second location 130b based on the quantity of items 135 that are likely to remain unused at each of the first location 130a and the second location 130b.Alternatively and / or additionally, the inventory management device 110 may adjust the reorder schedule of item 135 and / or the maximum (or minimum) quantity of item 135 stored at the first location 130a and / or the second location 130b, based on the quantity of item 135 that is likely to remain unused at each of the first location 130a and / or the second location 130b. In some cases, the inventory management device 110 may propose to completely remove item 150 stored at the first location 130a and / or the second location 130b from its storage location.

[0049] FIG. 5 is a block diagram showing an illustration of a computing system 500 that conforms to an example of a currently protected subject. Referring to FIGS. 1 and 5, the computing system 500 can be used to implement the inventory management device 110 and / or any component within the inventory management device 110.

[0050] As shown in FIG. 5, the computing system 500 may include a processor 510, a memory 520, a storage device 530, and an input / output device 540. The processor 510, the memory 520, the storage device 530, and the input / output device 540 can be connected to each other via a system bus 550. The processor 510 can process instructions to execute commands in the computing system 500. The instructions thus executed can, for example, implement one or more components of the inventory management device 110. In some embodiments, the processor 510 may be a single-threaded processor. Alternatively, the processor 510 may be a multi-threaded processor. The processor 510 can process instructions stored in the memory 520 and / or the storage device 530 to display graphical information on a user interface provided via the input / output device 540.

[0051] Memory 520 is a computer-readable medium, such as a volatile or non-volatile medium, and stores information within computing system 500. Memory 520 can store, for example, a data structure representing a configuration object database. Storage device 530 can provide persistent storage for computing system 500. Storage device 530 may be a floppy disk device, a hard disk device, an optical disk device, a tape device, a solid state device, and / or any other suitable persistent storage means. Input / output device 540 provides input / output operations to computing system 500. In some embodiments, input / output device 540 includes a keyboard and / or a pointing device. In various embodiments, input / output device 540 includes a display unit for displaying a graphical user interface.

[0052] According to some exemplary embodiments, input / output device 540 can provide input / output operations to a network device. For example, input / output device 540 may include an Ethernet port or other network port to communicate with one or more wired and / or wireless networks (e.g., local area network (LAN), wide area network (WAN), Internet).

[0053] In some embodiments, computing system 500 can be used to execute various interactive computer software applications that can be used to organize, analyze, and / or store data in various formats. Alternatively, computing system 500 can be used to execute any type of software application. These applications can be used to perform various functions, such as scheduling functions (e.g., generating, managing, editing spreadsheet documents, word processing documents, and / or other objects), computing functions, communication functions, etc. The application may include various add-in functions or may include independent computing products or functions. The functions can be used to generate a user interface provided via input / output device 540 when active within the application. The user interface can be generated by computing system 500 (e.g., on a computer screen monitor, etc.) and presented to the user.

[0054] One or more aspects or features of the subject matter described in this disclosure can be implemented by digital electronic circuitry, integrated circuitry, application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), computer hardware, firmware, software, and / or combinations thereof. These various aspects or features may include implementation by one or more computer programs executable and / or interpretable on a programmable system, the programmable system including at least one programmable processor for use in a special or general purpose, the at least one programmable processor being coupled to receive data and instructions from, and to transmit data and instructions to, a memory system, at least one input device, and at least one output device. The programmable system or computing system may include a client device and a server. The client device and the server are generally remote from each other and typically interact through a communication network. The relationship between the client device and the server results from computer programs running on respective computers and having a client-server relationship to each other.

[0055] These computer programs, also referred to as programs, software, software applications, applications, components, or code, include machine instructions for a programmable processor and are executable in high-level procedural and / or object-oriented programming languages and / or in assembly / machine language. The term "machine-readable medium" as used in this disclosure refers to any computer program product, apparatus, and / or device, such as, for example, magnetic disks, optical disks, memory, and programmable logic devices (PLDs), that can be used to provide machine instructions and / or data to a programmable processor, and includes a machine-readable medium that receives the device instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal used to provide machine instructions and / or data to a programmable processor. A machine-readable medium can store such machine instructions non-transitorily, for example, a non-transitory solid-state memory, a magnetic hard drive, or any equivalent storage medium. A machine-readable medium can alternatively or additionally store such machine instructions temporarily, for example, a processor cache or other random access memory associated with one or more physical processor cores.

[0056] To provide interaction with a user, one or more aspects or features of the subject matter described in this disclosure can be implemented on a computer having a display device, a keyboard, and a pointing device, where the display device can be, for example, a cathode ray tube (CRT), a liquid crystal display (LCD), or a light emitting diode (LED) monitor for displaying information to the user, and the pointing device can be, for example, a mouse or a trackball, by which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user. For example, the feedback provided to the user can be any form of perceptual feedback, such as, for example, visual feedback, auditory feedback, or tactile feedback, and the input by the user can be received in any form, such as acoustic input, voice input, tactile input, etc. Other possible input devices include a touch screen, or other touch-sensitive devices such as a single-point resistive, multi-point resistive, or capacitive trackpad, or voice recognition hardware and software, an optical scanner, an optical pointer, a digital image capture device, and associated interpretation software, etc.

[0057] In the foregoing description and claims, the terms "at least one" or "one or more" may be followed by elements or features listed thereafter. The term "and / or" may appear in two or more elements or features listed. Such terms are intended to refer to any individually listed element or feature, or any combination of any described element or feature with any other described element or feature, as long as it is not implicitly or explicitly inconsistent with the context in which it is used. For example, each of the terms "at least one of A and B", "one or more of A and B", and "A and / or B" is intended to mean "only A, only B, or both A and B". A similar interpretation is also intended for lists containing three or more items. For example, each of the terms "at least one of A, B, and C", "one or more of A, B, and C", and "A, B, and / or C" is intended to mean "only A, only B, only C, both A and B, both A and C, both B and C, or all of A, B, and C". The term "based on" in the foregoing description and claims is intended to mean "based at least in part on", and features or elements not described are also allowed.

[0058] As used herein, the term "user interface" (also referred to as an interactive user interface, graphical user interface, or UI) may refer to a network-based interface that includes data fields and / or other control elements for receiving input signals, providing electronic information, and / or providing information to a user in response to received input signals. The control elements may include dials, buttons, icons, selectable regions, or other perceptible presentations presented via the UI, and the UI may initiate data exchange with the device presenting the UI when interacted with (e.g., clicked, touched, selected, etc.). All or part of the UI may be implemented using technologies such as Hypertext Markup Language (HTML), FLASH®, JAVA®, .NET®, C, C++, web services, or Rich Site Summary (RSS). In some embodiments, the UI may be included in a stand-alone client device (e.g., a thick client, fat client), and the client device may be configured to communicate (e.g., send and receive data) according to one or more of the foregoing manners. The communication may occur between a medical device and a server communicating therewith.

[0059] Further non-limiting aspects or examples are described in the following numbered examples.

[0060] <Example 1>The method includes receiving training data including historical data regarding items under the control of one or more processors, training an expiration approaching analysis module using the training data, and generating a trained expiration approaching analysis module, where the trained expiration approaching analysis module is configured to receive item data regarding an item and generate an expiration approaching prediction for one or more items regarding the item data, receiving inventory information for the item, processing at least a portion of the inventory information using the trained expiration approaching analysis module to determine items stored at one or more locations that are unused and likely to exceed a target date, and providing an output to an inventory management device to perform an operation to avoid the item being unused and exceeding the target date.

[0061] <Example 2>In the method according to Example 1, the operation to avoid the item being unused and exceeding the target date includes the inventory management device moving at least some of the items from one or more locations to a usage location.

[0062] <Example 3>In the method according to Example 1 or 2, determining items stored at one or more locations that are unused and likely to exceed a target date includes comparing a category value with a threshold value.

[0063] <Example 4>In the method according to any one of Examples 1 to 3, the category value includes at least one of the unit cost, usage rate, and movement rate of the item.

[0064] <Example 5>In the method according to any one of Examples 1 to 4, the historical data includes inventory fluctuations at multiple locations.

[0065] <Example 6>In the method according to any one of Examples 1 to 5, the expiration approaching analysis module includes a machine learning model and one or more heuristic modules.

[0066] <Example 7>In the method according to any one of Examples 1 to 6, training the approaching expiration analysis module includes determining whether the historical data meets a first threshold, and performing training using a machine learning model in response to determining that the historical data meets the first threshold.

[0067] <Example 8>In the method according to any one of Examples 1 to 7, training the approaching expiration analysis module includes determining whether the historical data meets a first threshold, determining whether the historical data meets a second threshold in response to determining that the historical data does not meet the first threshold, and performing training using one or more heuristic modules in response to determining that the historical data meets the second threshold.

[0068] <Example 9>In the method according to any one of Examples 1 to 8, the one or more heuristic modules include a model that defines the relationship between the unit price of an item and the item usage rate, the relationship between the item value and the item usage rate, or the relationship between the unit price of an item and the interval to the earliest expiration date.

[0069] <Example 10>In the method according to any one of Examples 1 to 9, the training includes any one of supervised training, unsupervised training, reinforcement training, or dynamic training.

[0070] <Example 11> In the method according to any one of Examples 1 to 10, training the expiration approaching analysis module includes obtaining training item data, generating a first expiration approaching analysis module including a first processing pipeline, in the first processing pipeline, a machine learning model receives the training item data, and a heuristics module receives at least a part of the first output from the machine learning model as a first input, generating a second expiration approaching analysis module including a second processing pipeline, in the second processing pipeline, the heuristics module receives the training item data, and the machine learning model receives at least a part of the second output from the heuristics module as a second input, measuring resource usage for processing at least a part of the training item data using the first expiration approaching analysis module and the second expiration approaching analysis module, and selecting any one of the first expiration approaching analysis module and the second expiration approaching analysis module as the expiration approaching analysis module based on the resource usage.

[0071] <Example 12> In the method according to any one of Examples 1 to 11, the target date is at least one of the expiration date for the item, a predetermined period from the current date, or the planned inventory update date.

[0072] <Example 13> In the method according to any one of Examples 1 to 12, the first physical item can be obtained at a first location managed by an inventory management device, the second physical item of the item can be obtained at a second location managed by the inventory management device, and the operation to avoid the first physical item of the item being unused and exceeding the target date includes setting the inventory management device to ship the item from the first location when a dispensing request for the item is received.

[0073] <Example 14>The system includes at least one data processor and at least one memory that stores instructions. When the instructions are executed by the at least one data processor, the system receives training data including historical data regarding items, trains an expiration approaching analysis module using the training data, applies the expiration approaching analysis module to determine items stored at one or more locations that are unused and likely to pass the target date, and provides an output to perform operations to avoid the items being unused and exceeding the target date, and the memory.

[0074] <Example 15>In the system according to Example 14, the operation to avoid the items being unused and exceeding the target date includes moving at least some of the items from one or more locations to a place of use by an inventory management device, and the historical data includes inventory fluctuations at multiple locations.

[0075] <Example 16>In the system according to Example 14 or 15, determining items stored at one or more locations that are unused and likely to pass the target date includes comparing a category value with a threshold value, and the category value includes at least one of the unit cost, usage rate, and movement speed of the item.

[0076] <Example 17>In the system according to any one of Examples 14 to 16, the expiration approaching analysis module includes a machine learning model and one or more heuristic modules, and the one or more heuristic modules include models that define the relationship between the unit price of an item and the usage rate of the item, the relationship between the value of the item and the usage rate of the item, or the relationship between the unit price of the item and the interval to the earliest expiration date.

[0077] <Example 18> In the system according to any one of Examples 14 to 17, training the expiration approaching analysis module includes determining whether the historical data satisfies a first threshold, and performing training using a machine learning model in response to determining that the historical data satisfies the first threshold, or determining whether the historical data satisfies a second threshold in response to determining that the historical data does not satisfy the first threshold, and performing training using one or more heuristic modules in response to determining that the historical data satisfies the second threshold.

[0078] <Example 19> In the system according to any one of Examples 14 to 18, training includes any one of supervised training, unsupervised training, reinforcement training, or dynamic training.

[0079] <Example 20> A non-transitory computer-readable medium is a non-transitory computer-readable medium that stores instructions, and when the instructions are executed by at least one data processor, receives training data including historical data regarding items, trains an expiration approaching analysis module using the training data, applies the expiration approaching analysis module to determine items stored in one or more locations that are unused and likely to exceed the target date, and performs an operation that provides an output to avoid the items being unused and exceeding the target date.

[0080] The subject matter described in this disclosure can be implemented in a system, apparatus, method, and / or article, depending on the desired configuration. The examples described in the foregoing description do not represent all examples that are consistent with the subject matter described in this disclosure. Rather, these examples are merely some examples that are consistent with aspects regarding the described subject matter. Although some variations have been described in detail above, other variations and additional examples are possible. In particular, additional features and / or variations can be provided in addition to what is described in this disclosure. For example, the foregoing examples can be directed to combinations and sub-combinations with the features described above, and / or combinations and sub-combinations with some additional features. Also, the logic flows shown in the accompanying figures and / or described in this disclosure do not necessarily require the particular order, or sequential order, shown to achieve desirable results. Other examples can be included within the appended claims.

Claims

1. Receiving training data including historical data regarding items under the control of one or more processors; Training a near - expiration analysis module using the training data to generate a trained near - expiration analysis module, the trained near - expiration analysis module being configured to receive item data regarding the item and generate a near - expiration prediction for one or more items regarding the item data; Receiving inventory information for the item; Processing at least a portion of the inventory information using the trained near - expiration analysis module to determine items stored at one or more locations that are unused and likely to exceed a target date; Providing an output to an inventory management device to perform an operation to avoid the item being unused and exceeding the target date; Including A method.

2. The operation to avoid the item being unused and exceeding the target date is The inventory management device moving at least some of the items from the one or more locations to a place of use Including The method according to claim 1.

3. Determining items stored at the one or more locations that are unused and likely to exceed the target date includes Comparing a category value with a threshold Including The method according to claim 1.

4. The category value includes at least one of the unit cost, usage rate, and movement rate of the item The method according to claim 3.

5. The historical data includes inventory fluctuations at multiple locations The method according to claim 1.

6. The near - expiration analysis module includes a machine learning model and one or more heuristic modules The method according to claim 1.

7. Training the near - expiration analysis module includes Determining whether the historical data meets a first threshold; Performing training using the machine learning model in response to determining that the historical data meets the first threshold; Including The method according to claim 6.

8. Training the near - expiration analysis module includes Determining whether the historical data meets the first threshold; ​ In response to determining that the history data does not satisfy the first threshold, determining whether the history data satisfies a second threshold; In response to determining that the history data satisfies the second threshold, performing training using the one or more heuristic modules; including The method according to claim 7.

9. The one or more heuristic modules include a model that defines a relationship between an item unit price and an item usage rate, a relationship between an item value and the item usage rate, or a relationship between an item unit price and an interval to the earliest expiration date. The method according to claim 6.

10. The training includes any one of supervised training, unsupervised training, reinforcement training, or dynamic training. The method according to claim 1.

11. Training the near-expiration analysis module obtaining training item data; generating a first near-expiration analysis module including a first processing pipeline, in which the machine learning model receives the training item data in the first processing pipeline, and the heuristic module receives at least a part of a first output from the machine learning model as a first input; generating a second near-expiration analysis module including a second processing pipeline, in which the heuristic module receives the training item data in the second processing pipeline, and the machine learning model receives at least a part of a second output from the heuristic module as a second input; measuring resource usage for processing at least a part of the training item data using the first near-expiration analysis module and the second near-expiration analysis module; selecting either one of the first near-expiration analysis module and the second near-expiration analysis module as the near-expiration analysis module based on the resource usage; including The method according to claim 6.

12. The target date is at least one of an expiration date for the item, a predetermined period from the current date, or a planned inventory update date. The method according to claim 1.

13. The first physical object of the item is obtainable at a first location managed by the inventory management device. The second physical item can be obtained at a second location managed by the inventory management device, such operation to avoid the first physical item of the item being unused and exceeding the target date, setting the inventory management device to ship the item from the first location when receiving a dispensing request for the item including The method according to claim 1.

14. At least one data processor, At least one memory storing instructions, when the instructions are executed by the at least one data processor, receiving training data including historical data regarding an item, training an expiration approaching analysis module using the training data, applying the expiration approaching analysis module to determine items stored at one or more locations that are unused and likely to exceed the target date, providing an output to perform an operation to avoid the item being unused and exceeding the target date, A memory that performs an operation including including System.

15. Such operation to avoid the item being unused and exceeding the target date, causing the inventory management device to move at least some of the items from the one or more locations to the place of use including The historical data includes inventory fluctuations at multiple locations. The system according to claim 14.

16. Determining items stored at the one or more locations that are unused and likely to exceed the target date, including comparing a category value with a threshold value including The category value includes at least one of the unit cost, usage rate, and movement speed of the item. The system according to claim 14.

17. The expiration approaching analysis module includes a machine learning model and one or more heuristic modules, The one or more heuristic modules include models that define the relationship between the unit price of an item and the usage rate of the item, the relationship between the value of the item and the usage rate of the item, or the relationship between the unit price of the item and the interval until the earliest expiration date. The system according to claim 14.

18. Training the expiration approaching analysis module, determining whether the historical data satisfies a first threshold value, Performing training using the machine learning model in response to determining that the history data satisfies the first threshold, or including, In response to determining that the history data does not satisfy the first threshold, determining whether the history data satisfies a second threshold, Performing training using the one or more heuristic modules in response to determining that the history data satisfies the second threshold, including, The system according to claim 14.

19. The training includes any one of supervised training, unsupervised training, reinforcement training, or dynamic training. The system according to claim 14.

20. A non-transitory computer-readable medium storing instructions, when the instructions are executed by at least one data processor, Receiving training data including historical data regarding an item, Training an expiration approaching analysis module using the training data, Applying the expiration approaching analysis module to determine items stored at one or more locations that are unused and likely to have passed the target date, Providing an output and performing an operation to avoid the item being unused and exceeding the target date, including performing an operation, Non-transitory computer-readable medium.

Images