Replenishment planning device and replenishment planning method

Abstract

To generate a replenishment plan that reduces both a decrease in service level of storage equipment and the number of times of replenishment.SOLUTION: A replenishment planning device generates a directed graph including vertices each corresponding to a combination of a replenishment timing in a target period and a replenishment pattern of a replenishment target object and an arc from a vertex whose corresponding replenishment timing precedes to a subsequent vertex, calculates and adds a weight indicating the length of a period in which a service level of storage equipment decreases between replenishment timings corresponding to an output source and an output destination of each arc and the number of times of replenishment at a replenishment timing corresponding to the output destination of the arc, based on a demand result of the replenishment target object and a replenishment pattern corresponding to the output source of the arc, and generates a replenishment plan indicating a replenishment timing and a replenishment pattern corresponding to each vertex included in the shortest route from a starting point to an end point in the directed graph in which the weight is added to each arc.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a replenishment planning device and a replenishment planning method. [Background technology]

[0002] As a background technique in this technical field, there is JP 2004-206413 A (Patent Document 1). This publication states that "the system is configured to include a demand forecasting unit which forecasts the demand per unit time for cash and other items as replenishment items in an automated teller machine, an input unit which inputs the number of replenishment operations or specific replenishment operation days for the automated teller machine within a specified period, and a plan creating unit which determines an operation plan including replenishment operation days including specific replenishment operation days and the replenishment item quantities for the replenishment operation days, based on the demand forecast result by the demand forecasting unit and the number of replenishment operations or specific replenishment operation days input from the input unit, so that replenishment operations are performed for the number of replenishment operations within a specified period" (see abstract). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2004-206413 A Summary of the Invention [Problem to be solved by the invention]

[0004] When the storage device that stores the replenishment objects runs out of replenishment objects or when the content amount of the replenishment objects in the storage device reaches the upper limit, the storage device may experience downtime, etc., which may result in a decline in the service level. If the number of times the storage device is replenished with the replenishment objects increases in order to reduce the frequency of such declines in the service level of the storage device, the operating costs associated with the replenishment work will increase. In other words, there is a trade-off between suppressing the decline in the service level and suppressing the number of times the replenishment objects are replenished.

[0005] However, Patent Document 1 does not disclose a method for creating a plan for simultaneously suppressing a decrease in the service level in such an automated teller machine and suppressing the number of times that refills of refill items are performed.

[0006] Therefore, one aspect of the present invention generates a replenishment plan that simultaneously suppresses a decrease in the service level of a storage device and suppresses the number of times that the storage device is replenished with replenishment target items. [Means for solving the problem]

[0007] In order to solve the above problem, one aspect of the present invention employs the following configuration: A replenishment planning device that generates a replenishment plan for replenishing replenishment objects in a storage device includes a processor and a memory, the memory holds demand information indicating actual demand results for each type of the replenishment objects and replenishment patterns indicating combinations of quantities of each type of the replenishment objects, the processor generates a directed graph including a start point, an end point, and vertices corresponding to each combination of the replenishment timing within a target period of the replenishment plan and the replenishment pattern, as well as an arc from the start point to the vertex, an arc from the vertex to the end point, and an arc from a vertex whose corresponding replenishment timing is earlier to a vertex whose corresponding replenishment timing is later, and for each arc included in the directed graph, generates a replenishment timing corresponding to an output source and an output destination of the arc. weights indicating the length of a period during which the service level for the replenishment objects in the storage equipment will decrease between timings and the number of replenishments at the replenishment timing corresponding to the output destination of the arc are calculated and assigned based on the actual demand and the replenishment pattern corresponding to the output source of the arc, the shortest path from the start point to the end point in the directed graph in which each of the arcs has been assigned a weight is searched for based on a predetermined algorithm, and a replenishment plan is generated as a plan for changing the quantity of each type of replenishment object stored in the storage equipment to the quantity indicated by the replenishment pattern corresponding to the vertex at the replenishment timing corresponding to each vertex included in the searched shortest path. Effect of the Invention

[0008] According to one aspect of the present invention, it is possible to generate a replenishment plan that achieves both prevention of a decrease in the service level of storage equipment and prevention of the number of times that the storage equipment is replenished with replenishment target items.

[0009] Problems, configurations and effects other than those described above will become apparent from the following description of the embodiments. [Brief description of the drawings]

[0010] [Figure 1] 1 is a block diagram showing an example of the configuration of a replenishment planning device according to a first embodiment. [Diagram 2] FIG. 4 is a diagram illustrating an example of a data configuration of demand information in the first embodiment. [Diagram 3] FIG. 4 is a diagram illustrating an example of a data configuration of demand forecast information in the first embodiment. [Figure 4] 6 is a diagram illustrating an example of a data configuration of replenishment pattern information according to the first embodiment. FIG. [Diagram 5] FIG. 4 is a diagram illustrating an example of a data configuration of optimal replenishment plan information in the first embodiment. [Figure 6] 1 is a flowchart illustrating an example of a demand forecasting process according to the first embodiment. [Figure 7] 1 is a flowchart showing an example of an optimal replenishment plan calculation process in the first embodiment. [Figure 8] FIG. 2 is an explanatory diagram showing an example of a directed graph created in the optimal replenishment plan calculation process in the first embodiment. [Figure 9] FIG. 11 is a block diagram showing an example of the configuration of a replenishment planning device according to a second embodiment. [Figure 10] 13 is a flowchart illustrating an example of a demand forecasting process according to the second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings. In this embodiment, the same components are generally designated by the same reference numerals, and repeated explanations are omitted. Note that this embodiment is merely an example for realizing the present invention, and does not limit the technical scope of the present invention.

[0012] In this embodiment, a replenishment planning device that generates a replenishment plan for replenishing cash to an ATM (Automatic Teller Machine), which is an example of a replenishment plan, will be described. The ATM has one or more cassettes, and any one cassette stores one denomination of cash. In other words, any one cassette does not store multiple denominations of cash. Also, cash of the same denomination may be stored in multiple cassettes. For example, the ATM has a cassette that stores only 1,000 yen notes, a cassette that stores only 5,000 yen notes, and a cassette that stores only 10,000 yen notes.

[0013] Each denomination of cash is an example of a replenishment object. Additionally, each cassette that stores each denomination of cash is an example of a storage unit that stores the replenishment object. Additionally, an ATM having a cassette is an example of a storage device having a storage unit.

[0014] If an ATM receives a lot of cash, the cassette will fill up and no more cash can be deposited (at least from a cassette with a full cash capacity), and if an ATM receives a lot of cash, the cassette will run out of cash and no more cash can be withdrawn (at least from an empty cassette).

[0015] Hereinafter, a state in which the amount of cash contained in a cassette of each denomination in an ATM reaches its upper limit, or a cassette of that denomination is empty, is also referred to as ATM downtime for that denomination. Note that a state in which the amount of cash contained in a cassette of each denomination in an ATM becomes equal to or greater than a first predetermined value (predetermined upper limit), or equal to or less than a second predetermined value (predetermined lower limit) (where 0≦second predetermined value<first predetermined value≦upper limit of amount contained in cassette) may also be defined as ATM downtime for that denomination. When downtime occurs, the service level for that denomination decreases.

[0016] When an ATM is to be replenished with cash, a cash replenishment staff member goes to the ATM's location with a cassette for each denomination that contains cash of that denomination (or is empty), replaces the cassette for that denomination in the ATM with the cassette of the corresponding denomination that he or she brought with him, and takes away the cassette that was in the ATM.

[0017] The cassettes brought by the cash replenisher contain a number of cash sheets selected from a number of different numbers of cash sheets for each denomination that are determined in advance. Specifically, for example, the cash replenisher brings, for each denomination, a cassette containing 0 cash sheets of that denomination (an empty cassette for that denomination), a cassette containing 2,000 cash sheets of that denomination, and a cassette containing 4,000 cash sheets of that denomination.

[0018] For example, if an ATM's cassette for 1,000 yen notes has 500 remaining notes, the cassette for 5,000 yen notes has 1,000 remaining notes, and the cassette for 10,000 yen notes has 0 remaining notes, and a cash replenishment staff member brings a cassette containing 2,000 1,000 yen notes, a cassette containing 2,000 5,000 yen notes, and a cassette containing 4,000 10,000 yen notes to replenish cash (i.e., replace cassettes of the same denomination), the ATM will have been replenished with 1,500 1,000 yen notes, 1,000 5,000 yen notes, and 4,000 10,000 yen notes.

[0019] Also, for example, if the number of remaining 1,000 yen notes in an ATM's cassette is 4,000, and a cash replenisher brings a cassette containing 2,000 1,000 yen notes, then 2,000 1,000 yen notes will have been removed. Hereinafter, the cassette brought by the cash replenisher will be referred to as the replenishment cassette.

[0020] In addition, in the present embodiment, when the ATM is replenished with cash, all cassettes in the ATM are replenished (i.e., all cassettes in the ATM are replaced with replenishment cassettes of the same denomination). In addition, the timing of cash replenishment in the ATM in this embodiment is, for example, the end of the target day for cash replenishment, but may be other timing (for example, the start of the target day for cash replenishment, etc.).

[0021] The replenishment planning device in this embodiment is not limited to the above-mentioned ATM cash replenishment plan, but can also be applied to generating replenishment plans for any replenishment object and any storage equipment having a storage unit, such as a product replenishment plan for a vending machine (in this case, the replenishment object is the product, the storage unit is each section of a rack that stores the products installed in the vending machine, and the storage equipment is the vending machine) or a product replenishment plan for a store (in this case, the replenishment object is the product, the storage unit is each section of a shelf that stores the products installed in the store, and the storage equipment is the shelf). EXAMPLES

[0022] 1 is a block diagram showing an example of the configuration of a replenishment planning device 100. The replenishment planning device 100 is configured by a computer having a CPU (Central Processing Unit) 101, a memory 102, an auxiliary storage device 103, an input device 104, an output device 105, and a communication device 106, which are connected to each other by an internal communication line 107 such as a bus.

[0023] The CPU 101 includes a processor and executes programs stored in the memory 102. The memory 102 includes a ROM (Read Only Memory), which is a non-volatile storage element, and a RAM (Random Access Memory), which is a volatile storage element. The ROM stores immutable programs (e.g., a Basic Input / Output System (BIOS)). The RAM is a high-speed, volatile storage element such as a DRAM (Dynamic Random Access Memory), and temporarily stores programs executed by the CPU 101 and data used when the programs are executed.

[0024] The auxiliary storage device 103 is a large-capacity non-volatile storage device such as a magnetic storage device (HDD (Hard Disk Drive)) or a flash memory (SSD (Solid State Drive)), and stores programs executed by the CPU 101 and data used when the programs are executed. That is, the programs are read from the auxiliary storage device 103, loaded into the memory 102, and executed by the CPU 101.

[0025] The input device 104 is a device such as a keyboard or a mouse that receives input from an operator. The output device 105 is a device such as a display device or a printer that outputs the results of program execution in a format that can be viewed by the operator.

[0026] The communication device 106 is a network interface device that controls communication with other devices according to a predetermined protocol, and may include a serial interface such as a Universal Serial Bus (USB).

[0027] A part or all of the programs executed by the CPU 101 may be provided to the replenishment planning device 100 from a removable medium (CD-ROM, flash memory, etc.) which is a non-transitory storage medium, or from an external computer equipped with a non-transitory storage device via a network, and stored in a non-volatile auxiliary storage device 103 which is a non-transitory storage medium. For this reason, the replenishment planning device 100 may have an interface for reading data from the removable medium.

[0028] The replenishment planning device 100 is a computer system configured on one physical computer, or on multiple logically or physically configured computers, and may operate in separate threads on the same computer, or may operate on a virtual computer constructed on multiple physical computer resources.

[0029] The CPU 101 includes, for example, a demand forecasting unit 111 and an optimal replenishment plan calculation unit 112. The demand forecasting unit 111 predicts future demand for cash of each denomination at an ATM. The optimal replenishment plan calculation unit 112 calculates a plan for replenishing cash to an ATM. In particular, the optimal replenishment plan calculation unit 112 calculates a replenishment plan that achieves both a reduction in ATM downtime (i.e., suppression of a decrease in the service level of an ATM) and a reduction in the workload of replenishing cash to an ATM.

[0030] For example, CPU 101 functions as a demand forecasting unit 111 by operating according to a demand forecasting program loaded into memory 102, and functions as an optimal replenishment plan calculation unit 112 by operating according to an optimal replenishment plan calculation program loaded into memory 102. The relationships between programs and functional units are similar for other functional units included in CPU 101 in other embodiments described later.

[0031] Some or all of the functions of the functional units included in the CPU 101 may be realized by hardware such as an Application Specific Integrated Circuit (ASIC) or a Field Programmable Gate Array (FPGA).

[0032] The auxiliary storage device 103 holds, for example, demand information 131, demand forecast information 132, replenishment pattern information 133, and optimal replenishment plan information 134. The demand information 131 indicates the demanded and remaining amounts of cash of each denomination at the ATM. The demand information 131 is generated, for example, based on the past usage record of the ATM. The demand forecast information 132 indicates, for example, the predicted future demand amount of cash of each denomination at the ATM. The demand forecast information 132 is generated, for example, based on the demand information 131.

[0033] In this embodiment, an example will be described in which the demand forecasting unit 111 generates the demand forecasting information 132, but the demand forecasting information 132 may be provided in advance. In this case, the replenishment planning device 100 does not need to include the demand forecasting unit 111 and the demand information 131, and the demand forecasting process described later may also be omitted.

[0034] The replenishment pattern information 133 indicates, for example, the amount of cash contained in a replenishment cassette for each denomination. In this way, the replenishment pattern information 133 can indicate a replenishment pattern. The replenishment pattern indicates a combination of the amounts (number of pieces) of cash of each denomination, and specifically indicates, for example, a combination of the amounts of cash contained in a replenishment cassette for each denomination.

[0035] For example, when denominations 1 to n exist, each replenishment pattern is represented by (the number of cash of denomination 1 stored in the replenishment cassette of denomination 1, the number of cash of denomination 2 stored in the replenishment cassette of denomination 2, ..., the number of cash of denomination n stored in the replenishment cassette of denomination n). Note that there may be multiple types of content quantities in replenishment cassettes of one denomination. Also, it is assumed that replenishment pattern information 133 is determined in advance.

[0036] The optimal replenishment plan information 134 indicates a replenishment plan that the optimal replenishment plan calculation unit 112 has generated by referring to the demand information 131 , the demand forecast information 132 , and the replenishment pattern information 133 .

[0037] In addition, some or all of the information stored in the auxiliary storage device 103 may be stored in the memory 102, or in an external database connected to the replenishment planning device 100.

[0038] In this embodiment, the information used by the replenishment planning device 100 does not depend on the data structure and may be expressed in any data structure. In this embodiment, the information is expressed in a table format, but the information can be stored in a data structure appropriately selected from, for example, a list, a database, or a queue.

[0039] In the following examples of Figures 2 to 4, for simplicity of explanation, it is assumed that all ATMs have one cassette for cash of denomination 1 and one cassette for cash of denomination 2 (there is a one-to-one correspondence between denominations and cassettes for all ATMs).

[0040] 2 is a diagram showing an example of the data configuration of the demand information 131. The demand information 131 includes, for example, an ID column 1311, a date column 1312, a demanded quantity column 1313, and a remaining quantity column 1314. The ID column 1311 holds an ID for identifying an ATM. The date column 1312 holds a date (a past date when the demanded quantity and the remaining quantity were observed). Note that a date (1 day) is an example of a unit period when the demanded quantity and the remaining quantity were observed.

[0041] The demand quantity column 1313 holds information indicating the demand quantity of cash of each denomination at the ATM with the ID indicated in the ID column 1311 on the date indicated in the date column 1312. The remaining quantity column 1314 holds information indicating the remaining quantity of cash of each denomination at the ATM with the ID indicated in the ID column 1311 on the date indicated in the date column 1312 (at the end of the date).

[0042] The demand information 131 is calculated, for example, from a log showing the usage record of the ATM. For example, for each denomination on each date at each ATM, the value obtained by subtracting the total number of deposited coins from the total number of withdrawn coins shown in the log is stored in the demand quantity column 1313 for that denomination on that date at that ATM. If the total number of deposited coins is greater than the total number of withdrawn coins, the demand quantity will be a negative value. Therefore, the demand quantity can be a positive value, 0, or a negative value.

[0043] Also, for example, the remaining amount of cash of each denomination at each ATM at the end of each date, as indicated by a log showing the ATM usage record, is stored in the remaining amount column 1314. The remaining amount can take a value of 0 or more (and less than the maximum storage amount of the cassette that stores the cash of that denomination).

[0044] 3 is a diagram showing an example of the data configuration of the demand forecast information 132. The demand forecast information 132 includes, for example, an ID column 1321, a date column 1322, and a predicted demand quantity column 1323. The ID column 1321 holds an ID for identifying an ATM. The date column 1322 holds a (future) date for predicting the demand quantity. Note that a date (1 day) is an example of a unit period for predicting the demand quantity. The predicted demand quantity column 1323 holds information indicating the predicted demand quantity of cash of each denomination at an ATM having an ID indicated in the ID column 1321 on the date indicated in the date column 1322.

[0045] 4 is a diagram showing an example of the data configuration of the replenishment pattern information 133. The replenishment pattern information 133 includes, for example, a replenishment cassette cash quantity field 1331. The replenishment cassette cash quantity field 1331 holds, for example, information indicating the number of cash pieces stored in a replenishment cassette (a cassette brought by a person in charge of cash replenishment) of each denomination.

[0046] In the example of Figure 4, when replenishment of cash, the person in charge of replenishment of cash can bring a replenishment cassette containing 0 coins of denomination 1 cash (empty replenishment cassette for denomination 1), a replenishment cassette containing 2,000 coins of denomination 1 cash, a replenishment cassette containing 4,000 coins of denomination 1 cash, a replenishment cassette containing 0 coins of denomination 2 cash (empty replenishment cassette for denomination 2), a replenishment cassette containing 2,000 coins of denomination 2 cash, and a replenishment cassette containing 4,000 coins of denomination 2 cash.

[0047] In the example of Figure 4, the person in charge of cash replenishment can bring three types of replenishment cassettes with different numbers of cash stored for each denomination 1 and denomination 2 when replenishment of cash. Therefore, when replacing each of the ATM's cassettes for denomination 1 and denomination 2 with replenishment cassettes of the same denomination, there are 3 x 3 = 9 replenishment (replacement) patterns.

[0048] For example, replenishment with a replenishment pattern of (2,000 cash coins of denomination 1, 4,000 cash coins of denomination 2) is equivalent to changing the amount of cash coins of denomination 1 in the ATM to 2,000 coins, and changing the amount of cash coins of denomination 2 in the ATM to 4,000 coins.

[0049] In addition, the number of cash pieces stored in the replenishment cassette or the number of patterns of the number of cash pieces stored in the replenishment cassette may differ for each denomination, or may be the same for some or all of the denominations.

[0050] 5 is a diagram showing an example of the data configuration of the optimal replenishment plan information 134. The optimal replenishment plan information 134 holds, for example, an ID column 1341, a date column 1342, and a replenishment cassette cash quantity column 1343. The ID column 1341 holds an ID for identifying an ATM. The date column 1342 holds a date (of a future date targeted by the replenishment plan), i.e., a planned replenishment date. Note that a date (1 day) is an example of a unit period in the replenishment plan.

[0051] The replenishment cassette cash quantity column 1343 indicates the number of cash bills stored in a replenishment cassette to be replaced with a cassette of each denomination in the ATM having the ID indicated in the ID column 1311 on the date indicated in the date column 1312 .

[0052] For example, the record in the first row of the optimal replenishment plan information 134 in Figure 5 indicates that, in the replenishment plan indicated by the optimal replenishment plan information 134, on January 13, 2022, for an ATM with an ID of "00001", the cassette of denomination 1 loaded in the ATM will be replaced with a replenishment cassette containing 2,000 pieces of denomination 1 cash, and the cassette of denomination 2 loaded in the ATM will be replaced with a replenishment cassette containing 0 pieces of denomination 2 cash.

[0053] 6 is a flowchart showing an example of a demand forecasting process. In the demand forecasting process, the demand forecasting unit 111 generates demand forecasting information 132. It is assumed that the demand information 131 has been obtained before the demand forecasting process starts.

[0054] The demand forecasting unit 111 refers to the demand information 131 and acquires the demand quantity and the remaining quantity for each ATM and denomination on each date in a past period (S601). Note that the past period may be, for example, predetermined (for example, the most recent 30 days, etc.) or may be specified by the user.

[0055] The demand forecasting unit 111 uses a predetermined machine learning algorithm such as LightGBM (Light Gradient Boosting Machine) or ARIMA (AutoRegressive Integrated Moving Average) to construct a demand forecasting model that forecasts the demand for each denomination at each ATM on a future date (S602). Explanatory variables in the demand forecasting model include, for example, the ID of the ATM, the date, the month, and / or the day of the week, and the objective variable includes the demand (number) of each denomination.

[0056] In addition, the demand forecasting unit 111 uses the information acquired in step S601 as learning data in the prediction model construction process in step S602. In addition, the demand forecasting unit 111 may construct a demand prediction model without using the remaining number as learning data (using only the demand number as learning data).

[0057] The demand forecasting unit 111 uses the demand forecasting model constructed in step S602 to forecast the demand quantity of each denomination at each ATM on each date in a future period (S604). Note that the future period may be, for example, determined in advance (e.g., 30 days from the next day, etc.) or may be specified by the user.

[0058] The demand forecasting unit 111 stores the demand numbers of each denomination at each ATM on each date in the future period predicted in step S604 in the demand forecast information 132 (S605), and ends the demand forecasting process.

[0059] 7 is a flowchart showing an example of an optimal replenishment plan calculation process. In the optimal replenishment plan calculation process, the optimal replenishment plan calculation unit 112 generates optimal replenishment plan information 134. It is assumed that the demand information 131, the demand forecast information 132, and the replenishment pattern information 133 are obtained before the optimal replenishment plan calculation process starts.

[0060] Further, the optimal replenishment plan calculation process in Fig. 7 is executed for each of the ATMs for which the optimal replenishment plan is calculated. Therefore, for example, the ATMs for which the optimal replenishment plan is calculated (for example, it may be all ATMs or it may be an ATM specified by a user) are selected in sequence, and the optimal replenishment plan calculation process is executed in sequence for the selected ATMs. Therefore, an ATM in the following description of the optimal replenishment plan calculation process refers to one selected ATM.

[0061] The optimal replenishment plan calculation unit 112 obtains the remaining amount of cash of each denomination on the day before the start date of the ATM's planned target period (hereinafter simply referred to as the target period) from the demand information 131, and obtains the predicted demand amount of cash of each denomination for the ATM's target period from the demand forecast information 132 (S701).

[0062] The target period for an ATM may be determined in advance (for example, 30 days from the next day) or may be specified by the user (however, it is desirable that the target period be set within a range in which the remaining amount of cash of each denomination at the ATM on the day before the start date of the target period is stored in demand information 131 and the forecast demand amount of cash of each denomination at the ATM on the end date of the target period is stored in demand forecast information 132).

[0063] The optimal replenishment plan calculation unit 112 generates a weighted directed graph in steps S702 to S706. The optimal replenishment plan calculation unit 112 generates all vertices of the directed graph (S702).

[0064] Specifically, for example, the optimal replenishment plan calculation unit 112 creates a vertex s (in-degree 0) that is the start point of a path indicating the replenishment plan, and a vertex t (out-degree 0) that is the input destination of the path indicating the replenishment plan. Furthermore, the optimal replenishment plan calculation unit 112 creates vertices v corresponding to each combination of an arbitrary date d (replenishment timing) during the target period and an arbitrary replenishment pattern p. d,p Note that vertex s corresponds to the day before the start date of the target period, and vertex t corresponds to the day after the end date of the target period.

[0065] For example, assume that the length of the target period is 30 days, and the replenishment pattern information 133 in FIG. 4 is defined (i.e., there are 3×3=9 replenishment patterns). In this case, the optimal replenishment plan calculation unit 112 calculates the replenishment pattern information 133 by dividing the replenishment pattern information 133 into two vertices, vertex t and vertex s, and the 30×9 vertices v d,p That is, the number of vertices created by the optimal replenishment plan calculation unit 112 is 2 + length of the target period × number of replenishment patterns for each denomination.

[0066] The optimal replenishment plan calculation unit 112 calculates all v d,p For arc (s,v d,p For any vertices v and w, the arc from vertex v to vertex w is represented as (v, w).

[0067] For any combination of v satisfying d1 < d2 and v d_1,p_1 and v d_2,p_2 , an arc (v d_1,p_1 , v d_2,p_2 ) is created (S704). Note that "_" indicates a subscript. For example, "d1" and "d_1" have the same meaning. The optimal replenishment plan calculation unit 112 creates an arc (v d,p , t) for all v d,p (S705).

[0068] Each path from vertex s to vertex t in the directed graph represents one replenishment plan during the target period. Specifically, the replenishment plan indicated by the path from vertex s to vertex t is to perform replenishment of the replenishment pattern p on the date d corresponding to each v d,p included in the path.

[0069] FIG. 8 is an explanatory diagram showing an example of a directed graph created in steps S703 to S705 of the optimal replenishment plan calculation process. For ease of viewing the drawing, in the example of FIG. 8, some vertices and some branches of the directed graph 800 are omitted and described.

[0070] The directed graph 800 has the vertex s, the vertex t, and the vertices v of the length of the target period × the number of patterns d,p as described above. In the example of FIG. 8, in the directed graph 800, nodes indicating the same replenishment date and different replenishment patterns are arranged horizontally, and nodes indicating the same replenishment pattern and different replenishment dates are arranged vertically (the lower the position, the earlier the date).

[0071] For example, vertex 801 is a vertex indicating that on the first day of the target period, the cassette of currency type 1 loaded in the ATM is replaced with a replenishment cassette containing 0 sheets of currency of currency type 1, and the cassette of currency type 2 loaded in the ATM is replaced with a replenishment cassette containing 2000 sheets of currency of currency type 2.

[0072] Also, for example, vertex 802 indicates that on the third day of the target period, the cassette of currency type 1 loaded in the ATM is replaced with a replenishment cassette containing 2,000 pieces of currency type 1, and the cassette of currency type 2 loaded in the ATM is replaced with a replenishment cassette containing 2,000 pieces of currency type 2.

[0073] Therefore, for example, the path <vertex s, vertex 801, vertex 802, vertex t> in the directed graph 800 indicates a replenishment plan in which the above-described replenishment operation on the above-described date indicated by vertex 801 and the above-described replenishment operation on the above-described date indicated by vertex 802 are performed.

[0074] Returning to the description of FIG. 7. Note that the optimal replenishment plan calculation unit 112 may generate, for example, a replenishment plan in which replenishment is always performed within n days from the last replenishment date. Specifically, for example, in steps S703 to S705, the optimal replenishment plan calculation unit 112 satisfies d1 < d2 ≦ d1 + n (where n is a natural number of 2 or more and shorter than the length of the target period) for any combination of v d_1,p_1 and v d_2,p_2 (provided that vertex s is regarded as v with d = 0 d,p and vertex t is regarded as v with d = (the day after the end date of the target period) d,p ), an arc to arc (v d_1,p_1 , v d_2,p_2 ) may be created (that is, it is only necessary that the length of the period between d1 and d2 is always below a predetermined value).

[0075] Also, the optimal replenishment plan calculation unit 112 may generate, for example, a replenishment plan in which replenishment is not always performed on a specific day. Specifically, for example, when replenishment is not always performed on the m-th day of the target period, the optimal replenishment plan calculation unit 112 satisfies d1 < d2 and d1 ≠ m and d2 ≠ m (where m is a natural number indicating a day within the target period) for any combination of v d_1,p_1 and v d_2,p_2 (provided that vertex s is regarded as v with d = 0 d,p and vertex t is regarded as v with d = (the day after the end date of the target period) d,p ), arc (vd_1,p_1 , v d_2,p_2 It suffices to create (v

[0076] Also, the optimal replenishment plan calculation unit 112 may generate, for example, a replenishment plan in which replenishment is always performed on a specific day. Specifically, for example, when replenishment is always performed on the m-th day (specific day) of the target period, the optimal replenishment plan calculation unit 112, in steps S703 to S705, sets v such that d1 < d2 ≤ m or m ≤ d1 < d2 (where m is a natural number indicating a day within the target period). d_1,p_1 and v d_2,p_2 for any combination with v d,p (assuming the vertex s as v with d = 0 d,p and the vertex t as v with d = (the day after the end day of the target period)), d_1,p_1 , v d_2,p_2 ) It suffices to create (that is, make sure that d1 and d2 do not sandwich m).

[0077] Note that the optimal replenishment plan calculation unit 112 may generate a replenishment plan in which replenishment is always performed within n days from the last replenishment day and replenishment is not always performed on a specific day, or may generate a replenishment plan in which replenishment is always performed within n days from the last replenishment day and replenishment is always performed on a specific day.

[0078] In the generation of the above-described replenishment plan in which replenishment is always performed within n days from the last replenishment day and / or replenishment is not always performed on a specific day or replenishment is always performed on a specific day, the optimal replenishment plan calculation unit 112 does not need to generate arcs related to some combinations of vertices, and thus does not need to calculate the weights of the arcs, so the amount of calculation can be reduced.

[0079] Next, the optimal replenishment plan calculation unit 112 assigns a weight to each of the arcs generated in steps S703 to S705 (S706). A specific example of a method for assigning weights to arcs in step S706 will be described later. The optimal replenishment plan calculation unit 112 calculates the shortest path from vertex s to vertex t by solving the shortest path problem from vertex s to vertex t (searching for the shortest path) using, for example, the Dijkstra method (S707).

[0080] The optimal replenishment plan calculation unit 112 stores the replenishment plan generated by associating the date and replenishment pattern indicated by each vertex through which the shortest route calculated in step S707 passes with the ID of the ATM in the optimal replenishment plan information 134 (S708), and ends the optimal replenishment plan calculation process. Note that the optimal replenishment plan calculation unit 112 may output (display) the replenishment plan stored in the optimal replenishment plan information 134 to the output device 105, or may output it to another device connected to the replenishment planning device 100.

[0081] A specific example of the method of weighting the arcs in step S706 will be described below. The set of all denominations is denoted by J, and each denomination is represented by i∈J. The set of all dates included in the target period is denoted by D, and each date is represented by d∈D. The day on which the cash of denomination i in an ATM runs out or becomes full (reaching the upper limit of the content quantity of the cassette) is defined as the day on which downtime of denomination i in the ATM occurs, that is, the day on which the service level drops. Information indicating the upper limit of the content quantity of the cassette of each denomination in each ATM is stored in advance in memory 102 or auxiliary storage device 103.

[0082] Also, a variable c indicates whether the service level of denomination i will decrease on date d. i,d (dt) is defined as follows (Equation 1).

[0083]

number

[0084] For example, if the service level of denomination i is not decreased at the start of date d but is decreased at the end of date d, or if the service level of denomination i is not decreased at the end of date d but is decreased at the start of date d, then i,d (dt) =k (where k is a real number satisfying 0≦k≦1, for example).

[0085] Also, a variable c indicates whether to replenish the ATM with cash of denomination i on date d (whether to replace the cash with a replenishment cassette). i,d (rep) is defined as follows (Equation 2).

[0086]

number

[0087] In this case, the variable c i,d (dt) and variable c i,d (rep) A replenishment plan that minimizes the objective function indicating the sum of the weighted sum of x, y, and y for all dates in the target period and all denominations (i.e., a replenishment plan that satisfies the following (Equation 3)) is a replenishment plan that suppresses the deterioration of service level and reduces the cost of replenishment operations.

[0088]

number

[0089] Note that α and β are the variables c i,d (dt) and variable c i,d (rep) , and is specified by, for example, a user of the replenishment planning system 100.

[0090] In order to calculate a replenishment plan that satisfies the above (Equation 3), the optimal replenishment plan calculation unit 112 first calculates all the arcs (s, v d,p ) (where v d,pSpecifically, for example, the optimal replenishment plan calculation unit 112 assigns weights to the number of days dt i Calculate.

[0091] The optimal replenishment plan calculation unit 112 calculates an optimal replenishment plan by multiplying the sum of the number of days for which the service level will decrease by the start time of the dth day by the weighting factor, αΣ i∈J (dt i ) into the arc (s,v d,p ) as the weight.

[0092] In addition, the optimal replenishment plan calculation unit 112 sets the remaining number of each denomination on the day before the start date of the target period obtained in step S701 as the predicted remaining number at the start of the 1st day of the target period, and calculates the predicted remaining number of each denomination at the start of each of the 2nd to dth days by sequentially adding the predicted demand number of each denomination from the 1st to the d-1st days of the target period obtained in step S701 to the predicted remaining number of each denomination at the start of the 1st day (however, the predicted remaining number of each denomination on each day will not fall below 0 and will not exceed the upper limit of the content quantity of the cassette of that denomination; in other words, if the addition results in a result below 0, the result is replaced with 0, and if the addition results in a result exceeding the upper limit, the result is replaced with the upper limit).

[0093] Therefore, the optimal replenishment plan calculation unit 112 can determine that the days on which the remaining number of each denomination is 0 or the upper limit at the start of each of the 1st to dth days are days on which the service level of the denomination will decrease, and can determine dt i can be calculated.

[0094] The optimal replenishment plan calculation unit 112 calculates all the arcs (v d_1,p_1 ,vd_2,p_2 ) (However, d1 <d2かつv d_1,p_1 ≠s and v d_2,p_2 (may be t), and assign weights as follows. Specifically, for example, the optimal replenishment plan calculation unit 112 calculates the number of days dt i Calculate.

[0095] The optimal replenishment plan calculation unit 112 calculates a weighted sum of the sum of the number of days for which the service level decreases from the start of day d1+1 to the start of day d2 by denomination and the sum of the number of replenishments performed on day d1 (at the end of day d1) by denomination (one replenishment is performed for each denomination on day d1, i.e., the same number of replenishments are performed on day d1 as the total number of denominations). i∈J (dt i )+β|J| to arc (v d_1,p_1 ,v d_2,p_2 ) as the weight.

[0096] In addition, the optimal replenishment plan calculation unit 112 sets the number of each denomination indicated by replenishment pattern p1 as the predicted remaining number of each denomination at the start of d1+1, and calculates the predicted remaining number of each denomination at the start of each of d1+2 days to d2 days by sequentially adding the predicted demand number of each denomination from day d1+1 to day d2-1 of the target period obtained in step S701 to the predicted remaining number of each denomination at the start of d1+1 (however, the predicted remaining number of each denomination on each day will not fall below 0 and will not exceed the upper limit of the content quantity of the cassette of that denomination; in other words, if the addition results in a result below 0, the result is replaced with 0, and if the addition results in a result exceeding the upper limit, the result is replaced with the upper limit).

[0097] Therefore, the optimal replenishment plan calculation unit 112 can determine, for each day from the (d1 + 1)-th day to the d2-th day at the start time, the day when the remaining number of each currency type is 0 or the upper limit value as the day when the service level of the currency type decreases. As a result, dt can be calculated as the total number of days when the service level of the currency type decreases. i can be calculated.

[0098] The optimal replenishment plan calculation unit 112 can calculate a path that minimizes the objective function in the above (Equation 3), that is, a replenishment plan, by solving the shortest path problem of the directed graph with weights assigned to the arcs by the above method.

[0099] Note that in the calculation of dt by the optimal replenishment plan calculation unit 112 described above, for example, when the service level of currency type i does not decrease at the start time of date d but decreases at the end time of date d, or when the service level of currency type i does not decrease at the end time of date d but decreases at the start time of date d, if it is determined that the service level of currency type i has decreased for k days (k is a real number satisfying 0 < k < 1, for example) on date d, then in the above (Equation 1), c i = k can also be taken, and it is also possible to handle cases corresponding to this. i,d (dt) = k can also be taken, and it is also possible to handle cases corresponding to this.

[0100] Note that when there is only one currency type, since the sum for the currency type can be ignored, it is sufficient to calculate a replenishment plan that satisfies the following (Equation 4).

[0101]

Equation

[0102] When the optimal replenishment plan calculation unit 112 calculates a replenishment plan that satisfies the above (Equation 4), as the weight of an arbitrary arc (s, v d,p ), αdt i is calculated, and as the weight of an arbitrary arc (v d_1,p_1 , v d_2,p_2 ), αdt i+β can be calculated.

[0103] In the following, the variable c indicates the remaining amount (amount) of denomination i at the ATM on date d. i,d (cf) The less cash remaining in the ATM, the greater the free cash flow of the ATM owner. In this case, the variable c i,d (dt) and variable c i,d (rep) and variable c i,d (cf) A replenishment plan that minimizes the objective function, which indicates the sum of the weighted sum of x, y, and y for all dates in the target period and all denominations (i.e., a replenishment plan that satisfies the following equation (5)), is a replenishment plan that suppresses the decline in service level, reduces the cost of replenishment operations, and increases the free cash flow of ATM owners.

[0104]

number

[0105] where γ is the variable c i,d (cf) , and is specified by, for example, a user of the replenishment planning system 100.

[0106] When calculating a replenishment plan that satisfies the above (Equation 5), the optimal replenishment plan calculation unit 112 calculates weights for arcs as follows.

[0107] The optimal replenishment plan calculation unit 112 calculates all the arcs (s, v d,p ) for dt i The total cash balance of denomination i at the start of each of the first to d days of the target period, calculated by multiplying the predicted remaining number of each denomination by the unit price of that denomination, is calculated as cf i The optimal replenishment plan calculation unit 112 calculates the d,p ) as weights, i∈J (dt i )+γΣ i∈J(cf i ) can be calculated.

[0108] Similarly, the optimal replenishment plan calculation unit 112 calculates all the arcs (v d_1,p_1 ,v d_2,p_2 ) for dt i The total cash balance of denomination i at the start of each of the target period from d1+1 to d2 is calculated by multiplying the predicted remaining number of each denomination at the start of each of the target period from d1+1 to d2 by the unit price of that denomination. i The optimal replenishment plan calculation unit 112 calculates an arbitrary arc (v d_1,p_1 ,v d_2,p_2 ) as weights, i∈J (dt i )+β|J|+γΣ i∈J (cf i ) can be calculated.

[0109] In the above (Equation 3) and (Equation 5), the value of the objective function increases as the number of denominations whose service levels have decreased on the same day increases. In order to give the same evaluation whether the service levels of many denominations or few denominations have decreased on the same day, the Σ i∈J αc i,d (dt) Let αmax{c i,d (dt) |i∈J} - It is sufficient to minimize the replaced objective function.

[0110] Σ in the objective functions of (Eq. 3) and (Eq. 5) i∈J αc i,d (dt) is αmax{c i,d (dt) |i∈J} - When minimizing the replaced objective function, the optimal replenishment plan calculation unit 112 calculates the weight of each arc as described above, and calculates the number of days dt i Instead of , calculate the number of days dt for which the service level declines for at least one denomination, and add αΣi∈J (dt i ) can be replaced with αdt.

[0111] In addition, in the above (Equation 3) and (Equation 5), the value of the objective function increases as the number of denominations of cash replenished on the same day increases. In order to give the same evaluation whether a large number of denominations of cash are replenished on the same day or a small number of denominations of cash are replenished, the Σ i∈J βc i,d (rep) βmax{c i,d (rep) |i∈J} - It is sufficient that the replaced objective function is minimized.

[0112] Σ in the objective functions of (Eq. 3) and (Eq. 5) i∈J βc i,d (rep) βmax{c i,d (rep) |i∈J} - When minimizing the replaced objective function, the optimal replenishment plan calculation unit 112 simply replaces β|J| included in the weight of each arc with β in the calculation of the weight of each arc described above (i.e., it can be assumed that one replenishment is performed regardless of the number of denominations in the ATM).

[0113] After generating the replenishment plan by the above-mentioned method, the replenishment planning device 100 may update the generated replenishment plan, for example, at predetermined intervals (for example, every day, etc.). Specifically, for example, the replenishment planning device 100 acquires demand information 131 up to the time point when the predetermined period has elapsed, and executes a demand forecasting process based on the acquired demand information 131, thereby regenerating demand forecast information 132 for the remaining period of the target period.

[0114] Furthermore, the replenishment planning device 100 executes an optimal replenishment plan calculation process (regenerates a weighted directed graph with the period from the lapse of the predetermined period to the end of the target period as a new target period, and re-searches for the shortest path from the start point to the end point of the directed graph) based on the demand information 131, the regenerated demand forecast information 132, and the replenishment pattern information 133, thereby generating a replenishment plan for the new target period. The replenishment planning device 100 updates the replenishment plan for the remaining period of the original target period with the replenishment plan for the new target period.

[0115] As described above, the replenishment planning device 100 of this embodiment can calculate a replenishment plan that achieves both a reduction in ATM downtime (i.e., suppression of a decrease in ATM service level) and a reduction in the workload of replenishment of cash to ATMs, by solving the shortest path problem of a weighted directed graph corresponding to the optimization problems of (Equation 3) and (Equation 4). Furthermore, by performing replenishment according to the replenishment plan, ATM downtime is reduced, which can suppress a decrease in customer satisfaction of end users of ATMs and a loss of opportunities to collect fees for ATM deposits and withdrawals, and the reduction in the workload of replenishment of cash to ATMs can reduce the operating costs of the replenishment work.

[0116] Furthermore, the replenishment planning device 100 of this embodiment can calculate a replenishment plan that realizes a reduction in ATM downtime (i.e., suppression of deterioration in ATM service level), a reduction in the amount of work required to replenish cash to ATMs, and an increase in the free cash flow of ATM owners, by solving a shortest path problem in a weighted directed graph that corresponds to the optimization problem of (Equation 5).

[0117] Furthermore, past ATM usage records reflect factors such as day of the week and seasonal factors, and the replenishment planning device 100 of this embodiment predicts demand for each day based on past ATM usage records and calculates a replenishment plan based on the predicted demand, so that the replenishment plan takes into account factors such as day of the week and seasonal factors.

[0118] Furthermore, when the replenishment planning device 100 of this embodiment solves the shortest path problem using the Dijkstra algorithm, the amount of calculation required to calculate the replenishment plan is small. EXAMPLES

[0119] Below, differences from the first embodiment will be mainly described, and descriptions of points in common with the first embodiment will be omitted as appropriate.

[0120] When calculating weights for arcs of a directed graph, the replenishment planning device 100 of the first embodiment predicts the days on which the service level will decline and calculates the number of days on which the service level will decline. However, the replenishment planning device 100 of the present embodiment calculates an expected value for the number of days on which the service level will decline in a period corresponding to an arc of a directed graph, based on the probability that the service level will decline on each day.

[0121] 9 is a block diagram showing an example of the configuration of the replenishment planning device 100. The replenishment planning device 100 of the present embodiment differs from the replenishment planning device 100 of the first embodiment in that forecast error information 135 is further stored in the auxiliary storage device 103. The forecast error information 135 indicates the forecast error of the demand forecasting model.

[0122] 10 is a flowchart showing an example of a demand forecasting process. Following step S602, the demand forecasting unit 111 stores the forecast error (e.g., RMSE (Root Mean Squared Error) or MSE (Mean Squared Error)) of the demand forecasting model constructed in step S602 in the forecast error information 135 (S1002), and proceeds to step S603. Note that when the demand forecasting model was constructed in step S602, the forecast error of the demand forecasting model was also calculated.

[0123] The optimal replenishment plan calculation process of this embodiment will be described below. In the optimal replenishment plan calculation process of this embodiment, the objective function to be minimized is different from that of embodiment 1. Therefore, the weights given to the arcs of the weighted directed graph are also different.

[0124] The replenishment planning device 100 of the present embodiment uses Σ i∈J αc i,d (dt) Σ i∈J αP(c i,d (dt) ) to generate a replenishment plan that minimizes the objective function P(c i,d (dt) ) denotes the probability that the service level of denomination i decreases on date d.

[0125] An example of a method for calculating weights to be assigned to arcs of a weighted directed graph will be described below. In the following, the MSE is stored as a prediction error in the prediction error information 135, and the value is σ 2 Let us assume that.

[0126] The optimal replenishment plan calculation unit 112 calculates the optimal replenishment plan for an arbitrary arc (s, v d,p ), the predicted remaining number of each denomination at the start of each of the 1st to dth days is calculated in the same manner as in the first embodiment.

[0127] The optimal replenishment plan calculation unit 112 generates a probability density function for each denomination at the start of each of the first to d days, based on the calculated predicted remaining number of the denomination for that day, in which the remaining number of the denomination at the start of that day is a random variable. The calculated predicted remaining number of the denomination for that day is averaged, σ 2 A normal distribution with variance of σ and a logistic distribution with the calculated predicted remaining number of coins of the denomination on the day as the position and the scale of (σ×√3) / π are both examples of such probability density functions.

[0128] The optimal replenishment plan calculation unit 112 calculates the expected number of days for which the service level of the denomination i at the start of each of the 1st to dth days will decrease Edt iSpecifically, for the start time of each of the first to d days, the optimal replenishment plan calculation unit 112 calculates the sum of the expected value at which the value of the random variable in the probability density function corresponding to that day is 0 or less and the expected value at which the value of the random variable is equal to or greater than the upper limit of the content quantity of the cassette of the denomination i of the ATM, and calculates the sum of the calculated sums from the first to d days as an arc(s,v d,p ) to Edt i It is calculated as follows.

[0129] In addition, the optimal replenishment plan calculation unit 112 calculates an arbitrary arc (v d_1,p_1 ,v d_2,p_2 When calculating the weights for (d1+1) and (d2), the predicted remaining number of each denomination at the start of each of the d1+1 to d2 days is calculated in the same manner as in the first embodiment. For each denomination at the start of each of the d1+1 to d2 days, the optimal replenishment plan calculation unit 112 generates a probability density function (similar to the above-mentioned probability density function) based on the calculated predicted remaining number of each denomination for that day, in which the remaining number of that denomination at the start of that day is used as a random variable.

[0130] The optimal replenishment plan calculation unit 112 calculates the expected number of days that the service level of the denomination i will decrease at the start of each of the days d1+1 to d2, E i Specifically, for the start time of each of the d1+1th day to the d2th day, the optimal replenishment plan calculation unit 112 calculates the sum of the expected value at which the value of the random variable in the probability density function corresponding to the day is 0 or less and the expected value at which the value of the random variable is equal to or greater than the upper limit of the content quantity of the cassette of the denomination i of the ATM, and calculates the sum of the calculated sums from the d1+1th day to the d2th day as an arc(v d_1,p_1 ,v d_2,p_2 ) to Edt i It is calculated as follows.

[0131] The optimal replenishment plan calculation unit 112 calculates the optimal replenishment plan for an arbitrary arc (s, v d,p ) and any arc (v d_1,p_1 ,v d_2,p_2 ) in Example 1 i The above Edt i The weights are replaced by

[0132] The optimal replenishment plan calculation unit 112 may learn a probability model for directly calculating the probability that the service level will decrease on each date (downtime probability) by using a predetermined algorithm such as logistic regression. For example, explanatory variables in the probability model include the ID of the ATM, the date, month, and day of the week for which the probability of the service level decreasing is calculated, the date on which the ATM was last replenished, and the number of coins of each denomination remaining in the ATM immediately after the replenishment, and the like, and the objective variable includes the probability that the service level will decrease on each date.

[0133] Specifically, for example, the optimal replenishment plan calculation unit 112 calculates an arbitrary arc (s, v d,p ) for any denomination i and any date d k (0 <d k ≦d), if the remaining number of coins of denomination i at the start of the first day is the predicted remaining number described above (and no cash replenishment of denomination i is performed thereafter), then d k A probabilistic model showing the probability that the service level of the denomination i will decrease (downtime probability) at the start of the day is calculated using the demand number of the denomination i indicated by the demand information 131 as learning data.

[0134] Furthermore, the optimal replenishment plan calculation unit 112 calculates an arbitrary arc (v d_1,p_1 ,v d_2,p_2 ) for any denomination i and any date d j (d1 <d j ≦d2), and for any replenishment pattern p, if the remaining number of coins of denomination i at the start of day d1+1 is the number of coins of denomination i in replenishment pattern p (and no cash replenishment of denomination i is performed thereafter), then d j A probabilistic model showing the probability that the service level of the denomination i will decrease (downtime probability) at the start of the day is calculated using the demand number of the denomination i indicated by the demand information 131 as learning data.

[0135] The optimal replenishment plan calculation unit 112 calculates the sum of the probabilities indicated by the above-mentioned probability model for the denomination i by date, thereby obtaining the above-mentioned Edt i can be calculated.

[0136] In addition, when the optimal replenishment plan calculation unit 112 learns a probabilistic model for directly calculating the probability of a decrease in service level (downtime probability), the demand forecasting process does not need to be executed, i.e., the demand forecasting information 132 does not need to be generated.

[0137] As described above, the replenishment planning device 100 of this embodiment assigns weights indicating the expected value of the number of days for which the service level will decline to the arcs of the directed graph, and therefore can calculate a replenishment plan that more accurately evaluates the number of days for which the service level will decline.

[0138] The present invention is not limited to the above-mentioned embodiment, and various modifications are included. For example, the above-mentioned embodiment has been described in detail to clearly explain the present invention, and is not necessarily limited to those having all the configurations described. It is also possible to replace a part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. It is also possible to add, delete, or replace a part of the configuration of each embodiment with another configuration.

[0139] In addition, the above-mentioned configurations, functions, processing units, processing means, etc. may be realized in part or in whole by hardware, for example, by designing them as integrated circuits. In addition, the above-mentioned configurations, functions, etc. may be realized in software by a processor interpreting and executing a program that realizes each function. Information such as the program, table, file, etc. that realizes each function can be stored in a memory, a recording device such as a hard disk or SSD (Solid State Drive), or a recording medium such as an IC card, SD card, or DVD.

[0140] In addition, the control lines and information lines shown are those that are considered necessary for the explanation, and not all control lines and information lines in the product are necessarily shown. In reality, it can be considered that almost all components are connected to each other. [Explanation of symbols]

[0141] 100 replenishment planning device, 101 CPU, 102 memory, 103 auxiliary storage device, 104 input device, 105 output device, 106 communication device, 111 demand forecasting section, 112 optimal replenishment plan calculation section, 131 demand information, 132 demand forecast information, 133 replenishment pattern information, 134 optimal replenishment plan information, 135 forecast error information

Claims

1. A replenishment planning device that generates a replenishment plan for replenishing replenishment objects in a storage device, A processor and a memory, The memory includes: Demand information indicating actual demand for each type of the replenishment object; A replenishment pattern indicating a combination of the quantities of each type of replenishment object is stored; The processor, generating a directed graph including a start point, an end point, and vertices corresponding to each combination of the replenishment timing and the replenishment pattern within the target period of the replenishment plan, as well as an arc from the start point to the vertex, an arc from the vertex to the end point, and an arc from a vertex with an earlier replenishment timing to a vertex with a later replenishment timing; For each arc included in the directed graph, a weight indicating the length of a period during which a service level for the replenishment object in the storage device decreases between replenishment timings corresponding to the output source and output destination of the arc and the number of replenishments at the replenishment timings corresponding to the output destination of the arc are calculated and assigned based on the demand record and a replenishment pattern corresponding to the output source of the arc, searching for a shortest path from the start point to the end point in a directed graph in which each of the arcs is weighted based on a predetermined algorithm; a replenishment planning device that generates, as the replenishment plan, a plan for changing the quantity of each type of replenishment object stored in the storage device to the quantity indicated by the replenishment pattern corresponding to each vertex at the replenishment timing corresponding to the vertex included in the searched shortest route.

2. 2. The replenishment planning system of claim 1, the storage device has storage sections corresponding to the respective types of the replenishment objects, Each of the storage sections stores a corresponding type of replenishment object, The processor generates, as the replenishment plan, a plan to replace storage units corresponding to each type of replenishment object held by the storage equipment with storage units storing the same type of replenishment object in the quantity indicated by the replenishment pattern corresponding to the vertex at the replenishment timing corresponding to each vertex included in the searched shortest route.

3. 2. The replenishment planning system of claim 1, The replenishment planning device, when the processor determines that the quantity of each type of replenishment object stored in the storage device reaches a predetermined upper limit value or a predetermined lower limit value, determines that the service level for the replenishment object of that type in the storage device will decrease.

4. 2. The replenishment planning system of claim 1, A replenishment planning device, wherein the length of the period during which the service level will decrease, indicated by the weight assigned to each arc included in the directed graph, indicates the length of the period during which the service level for at least one type of replenishment object in the storage equipment will decrease between the replenishment timings corresponding to the output source and output destination of the arc.

5. 2. The replenishment planning system of claim 1, A replenishment planning device, in which the number of replenishments indicated by the weights assigned to each arc included in the directed graph is one, regardless of the number of denominations.

6. 2. The replenishment planning system of claim 1, The processor predicts, for each arc included in the directed graph, a remaining amount of the replenishment object stored in the storage device between replenishment timings corresponding to an output source and an output destination of the arc, based on the demand record and a replenishment pattern corresponding to the output source of the arc; The replenishment planning device further comprises a weight for each of the arcs that is indicative of the remaining quantity calculated for that arc.

7. 2. The replenishment planning system of claim 1, The processor, At a predetermined timing during the target period, the demand record up to the predetermined timing is acquired; regenerating the directed graph in which the weights are assigned to each arc, based on the acquired demand record and the replenishment pattern, for a new target period from the predetermined timing to the end of the target period; re-searching the shortest path in the regenerated directed graph based on the predetermined algorithm; generating a replenishment plan for the new target period based on the re-searched shortest path; A replenishment planning device updates the replenishment plan for the original target period with the replenishment plan for the new target period.

8. 2. The replenishment planning system of claim 1, For each arc included in the directed graph, the interval between the replenishment timings indicated by the output source and output destination of the arc is equal to or less than a predetermined value.

9. 2. The replenishment planning system of claim 1, A replenishment planning device, wherein for each arc included in the directed graph, none of the replenishment timings indicated by the output source and output destination of the arc coincide with a predetermined timing.

10. 2. The replenishment planning system of claim 1, A replenishment planning device, wherein for each arc included in the directed graph, no predetermined timing is included between the replenishment timings indicated by the output source and output destination of the arc.

11. 2. The replenishment planning system of claim 1, the memory holds forecast demand information that is generated based on the demand information and indicates a demand forecast for each type of the replenishment object at each replenishment timing; The processor calculates a weight corresponding to each of the arcs based on the actual demand, the demand forecast, and a replenishment pattern corresponding to an output source of the arc.

12. 12. The replenishment planning system of claim 11, The memory holds a prediction error of a model generated by learning the demand information, the model predicting the demand at each replenishment timing for each type of the replenishment object, The processor, For each of said arcs, generating a probability density function based on the demand forecast between replenishment timings corresponding to the output source and output destination of the arc and the forecast error; calculating an expected value of the length of a period during which a service level for the replenishment object in the storage device decreases between replenishment timings corresponding to the output source and output destination of the arc, based on the generated probability density function; The replenishment planning device calculates a weight corresponding to the arc based on the calculated expected value.

13. A replenishment planning method using a replenishment planning device that generates a replenishment plan for replenishing replenishment objects in a storage device, comprising: The replenishment planning device includes a processor and a memory, The memory includes: Demand information indicating actual demand for each type of the replenishment object; A replenishment pattern indicating a combination of the quantities of each type of replenishment object is stored; The replenishment planning method includes: The processor generates a directed graph including a start point, an end point, and a vertex corresponding to each combination of the replenishment timing and the replenishment pattern within a target period of the replenishment plan, as well as an arc from the start point to the vertex, an arc from the vertex to the end point, and an arc from a vertex whose corresponding replenishment timing is earlier to a vertex whose corresponding replenishment timing is later; The processor calculates and assigns weights to each arc included in the directed graph, the weights indicating the length of a period during which a service level for the replenishment object in the storage device decreases between replenishment timings corresponding to the output source and output destination of the arc, and the number of replenishments at the replenishment timings corresponding to the output destination of the arc, based on the demand record and a replenishment pattern corresponding to the output source of the arc; The processor searches for a shortest path from the start point to the end point in a directed graph in which weights are assigned to each of the arcs based on a predetermined algorithm; The replenishment planning method includes generating, as the replenishment plan, a plan for changing the quantity of each type of replenishment object stored in the storage device to the quantity indicated by the replenishment pattern corresponding to each vertex included in the searched shortest route at the replenishment timing corresponding to that vertex.

Images