Logistics warehouse system and method

Abstract

To provide a physical distribution warehouse system and its method capable of reducing the consumption of resources (including time, spatial movement, and energy) related to movement during shipping in a multi-tier automatic warehouse.SOLUTION: A physical distribution warehouse system comprises: a physical distribution co-occurrence degree regulation part D1 which regulates a physical distribution co-occurrence degree which is a value relating to a physical distribution co-occurrence property between articles; a storage place definition part which defines first places close to each other in a multi-tier automatic warehouse as storage places about a first plurality of articles in which the physical distribution co-occurrence degree regulated by the physical distribution co-occurrence degree definition part is a threshold value or higher, and defines a second place determined by a free location method as a storage place about a second plurality of articles in which the physical distribution co-occurrence degree regulated by the physical distribution co-occurrence degree definition part is less than the threshold value; and warehousing means which stores a plurality of warehousing articles in the first or second place defined by the storage place definition part when the plurality of warehousing articles are stored.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a logistics warehousing system and method for use in, for example, various logistics centers and the like, and in particular to a logistics warehousing system (including an article picking system) and method in which designated target articles are placed into a collection box among a wide variety of articles stored on storage shelves or in an automated warehouse. [Background technology]

[0002] In various logistics centers, items are handled in various ways, such as storage and sorting. For example, items are usually arranged in an automated warehouse by a free location method that does not limit the location where the items are stored, from the viewpoint of optimizing throughput while effectively utilizing the automated warehouse functions. In the case of the free location method, it is freely decided which storage location in the automated warehouse to put which item. When retrieving items thus stored, it has been common to transport a box (e.g., a returnable box) containing the target item from the storage location each time an order is placed, and the target item stored in the box is picked from the transported and retrieved box.

[0003] For example, Patent Document 1 discloses a technical concept regarding a picking system that enables efficient picking work by switching the picking operation means depending on the situation.

[0004] The technical idea disclosed in the above document appeals to the point of applying a certain ingenuity to the picking operation in order to solve the problem of making the picking work more efficient. In the case of arrangement in the above-mentioned free location method, the multiple target items to be accumulated together as one order at the time of delivery are necessarily located in mutually free positions that are unrelated to each other (in other words, positions that are relatively unrelated to each other). Therefore, it usually occurs that the target items are ordered simultaneously as one delivery order, but are located in multiple positions that are three-dimensionally separated from each other. In this case, since the multiple items to be gathered as one order are arranged at a distance from each other, in order to deliver the multiple items to the picking station continuously in the order unit (also called "orderly delivery"), that is, in order to accumulate the multiple separated boxes for the order, each box must be moved by the distance that the boxes are separated from each other. From the standpoint of minimizing transportation time and energy loss, it is logical to place items that should be gathered together as one order in locations close to each other, but because one order is placed after each item is stored in its respective storage location, it is not known at the time of warehousing which items will be ordered together as one order. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2019-52030 A Summary of the Invention [Problem to be solved by the invention]

[0006] Because it is not possible to know in advance whether items are to be ordered together as one order, they have no choice but to be stored in a warehouse at a free location. As a result, efforts to optimize the time and energy required for transportation when shipping items out of the warehouse have had to be neglected.

[0007] The present application has been made with the intention of solving these problems in the prior art, and has as its object to provide a logistics warehousing system and method that can prevent the consumption of resources (including time, spatial movement, energy, etc.) associated with movement when goods are removed from a warehouse in an automated multi-story warehouse. [Means for solving the problem]

[0008] In order to solve the above problems, the inventor of the present invention first developed a unique concept of "flow co-occurrence" inspired by the concept of "co-occurrence." In other words, in the field of natural language processing, "co-occurrence" refers to the simultaneous appearance of two character strings in a given document or sentence.

[0009] Therefore, the inventors have defined logistics co-occurrence as follows: when items x and y are simultaneously shipped (picked) for an order, there is logistics co-occurrence between the two items x and y. More specifically, the logistics co-occurrence degree Pxy, which is a concept that quantifies the logistics co-occurrence between items x and y, is mathematically defined by the following formula 1. Pxy=log[p(x, y) / {p(x)*p(y)}] (Equation 1) Here, p(x) and p(y) are the probabilities that items x and y are picked, and p(x, y) is the probability that items x and y are picked simultaneously into the same collection bin. Probabilities such as p(x) can be calculated using probabilities over a certain period of time, for example {(number of x picked per day) / (number of all items picked per day)}, and can also be extended to sets of three or more items. It is also possible to obtain the degree of logistics co-occurrence by a method other than the above, for example, machine learning as described below, and this is also included in the technical idea of ​​the present invention.

[0010] Next, the above-mentioned problems are summarized using the concept of logistics co-occurrence. For example, in order to store items in an automated warehouse using a free location method, it has not been possible to place each item (or its box) on a storage shelf while considering items with a certain or higher logistics co-occurrence value (or a certain threshold value). Similarly, items with logistics co-occurrence have not been stored on a storage shelf, even artificially. This is because the concept of logistics co-occurrence or logistics co-occurrence itself has not been conceived. One reason for this is that items with a certain or higher logistics co-occurrence value are not limited to simple items such as pencils and erasers, but often include unexpected items such as sofas and chainsaws. In other words, the basis for a certain or higher logistics co-occurrence (i.e., the logistics co-occurrence value is a certain threshold or higher) may not only be that the items are used together, but may also include seasonal factors, psychological factors, or factors that cause logistics co-occurrence as an indirect cause that is not directly related to the item. In summary, it can be said that it is sometimes difficult to artificially predict which items have a logistics co-occurrence, and that there were no standards in logistics centers and other facilities for handling items that have a logistics co-occurrence.

[0011] The inventors have noticed that a solution to the above problem can be achieved by using the above-mentioned concept of logistics co-occurrence or logistics co-occurrence degree to adopt a warehousing method that reflects logistics co-occurrence.

[0012] Therefore, in order to solve the above-mentioned problems, a logistics warehouse system according to a first aspect of the present invention comprises a logistics co-occurrence specification unit that specifies a logistics co-occurrence degree, which is a value related to the logistics co-occurrence between items, a storage location definition unit that defines a first location that is close to each other in an automated multi-story warehouse as a storage location for a first plurality of items whose logistics co-occurrence degree specified by the logistics co-occurrence specification unit is equal to or greater than a threshold, and defines a second location determined by a free location method as a storage location for a second plurality of items whose logistics co-occurrence degree does not meet the threshold, and a warehousing means that, when a plurality of incoming items are received, stores the multiple incoming items in the first or second location defined by the storage location definition unit.

[0013] In this case, the degree of logistics co-occurrence may be obtained by calculation using Formula 1 or by machine learning. According to the above configuration, a plurality of incoming items whose degree of logistics co-occurrence is equal to or greater than a certain threshold are regarded as a plurality of items with high logistics co-occurrence and are arranged in advance in a nearby location in the automated warehouse, so that the items with high logistics co-occurrence can be efficiently guided to the picking station, thereby enabling accurate, fast and efficient picking.

[0014] As a logistics warehouse system according to a second aspect of the present invention, in the above first aspect, the logistics co-occurrence determining unit may be configured to determine the logistics co-occurrence by calculating using a first and second probability that a first item and a second item are picked in a certain period, respectively, and a third probability that the first and second items are picked simultaneously into the same collection box in the certain period. With this configuration, the logistics co-occurrence can be defined for each item, and specifically, items with high logistics co-occurrence can be efficiently guided to a picking station.

[0015] As a logistics warehouse system according to a third aspect of the present invention, in the first aspect, the logistics co-occurrence definition unit may have a model generation means for generating a trained model that defines the logistics co-occurrence between a certain item and another item by machine learning using information on the types of almost all items stored in a certain time / days as input data and information on the types of multiple items that are co-occurring and released from the warehouse in a certain time / days as teacher data, an incoming item receiving means for receiving information on all incoming items that are brought in in a certain time / days, and a logistics co-occurrence output means for defining, as the logistics co-occurrence, a value output by the trained model generated by the model generation means for the incoming items received by the incoming item receiving means. According to this configuration, even for a set of items whose logistics co-occurrence is not necessarily logically clear or cannot be easily inferred, the logistics co-occurrence can be defined by reflecting the learning results by machine learning, so that it is possible to specifically and efficiently guide items with high logistics co-occurrence to a picking station.

[0016] The concept of logistics co-occurrence in each of the above aspects is not limited to fixed logistics co-occurrence of pencils and erasers, sashimi and wasabi, etc. It also includes variable logistics co-occurrence, such as high logistics co-occurrence of cakes and champagne before an event such as Christmas, high logistics co-occurrence of New Year's dishes and sake during the year-end and New Year holidays, and sudden logistics co-occurrence, such as high logistics co-occurrence of items that have become popular in the media, etc. Furthermore, for any of these logistics co-occurrences, a fluidity called regional fluctuation can be defined, and the logistics co-occurrence of each alcoholic beverage taking this fluidity into consideration is also encompassed in the technical idea of ​​the present application.

[0017] As a logistics warehousing system according to a fourth aspect of the present invention, in any one of the first to third aspects, the goods may be stored in the automated multi-story warehouse based on the logistics co-occurrence degree defined by the logistics co-occurrence degree defining unit. In this way, goods that have a high logistics co-occurrence, i.e., goods that are likely to be ordered and shipped simultaneously as one order, are stored in positions closer to each other (than in the case of the free location method), which makes it possible to reduce resources such as time / energy / effort required for movement when shipping and to increase throughput related to shipping.

[0018] In order to solve the above-mentioned problems, a control method for a logistics warehouse system according to a fifth aspect of the present invention defines a logistics co-occurrence degree, which is a value related to the logistics co-occurrence between items, defines a first location in the automated warehouse that is close to each other as a storage location for a first plurality of items whose defined logistics co-occurrence degree is equal to or greater than a threshold, defines a second location determined by a free location method as a storage location for a second plurality of items whose defined logistics co-occurrence degree is less than the threshold, and, when a plurality of incoming items are received, causes the plurality of incoming items to be stored in the first or second location defined by the storage location definition unit. According to the above configuration, a plurality of incoming items whose logistics co-occurrence degree is equal to or greater than a certain threshold are controlled to be preliminarily placed in a nearby location in the automated warehouse as a plurality of items with high logistics co-occurrence, so that items with high logistics co-occurrence can be efficiently guided to a picking station, thereby enabling accurate, rapid and efficient picking.

[0019] As a control method for a logistics warehouse system according to a sixth aspect of the present invention, in the fifth aspect, the logistics co-occurrence may be defined by calculating a first probability that a first item and a second item are picked in a certain period of time, and a third probability that the first and second items are picked simultaneously into the same collection box in the certain period of time. With this configuration, the logistics co-occurrence can be defined for each item, and specifically, items with high logistics co-occurrence can be efficiently guided to a picking station.

[0020] As a logistics warehouse system according to a seventh aspect of the present invention, in the above fifth aspect, the logistics co-occurrence may be determined by using information on the types of substantially all items received in a certain time / day as input data, information on the types of multiple items co-occurring and released in a certain time / day as training data, generating a trained model that defines the logistics co-occurrence between a certain item and another item by machine learning, receiving information on all incoming items delivered in a certain time / day, and defining a value output using the generated trained model for the received incoming items as the logistics co-occurrence. According to this configuration, even for a set of items whose logistics co-occurrence is not necessarily logically clear or cannot be easily inferred, the logistics co-occurrence can be defined by reflecting the learning results by machine learning, so that it is possible to specifically and efficiently guide items with high logistics co-occurrence to a picking station.

[0021] As a logistics warehousing system according to an eighth aspect of the present invention, in any one of the fifth to seventh aspects, the goods may be stored in the automated warehouse based on the specified degree of logistics co-occurrence. In this way, goods that have a high degree of logistics co-occurrence, i.e. goods that are likely to be ordered and shipped simultaneously as one order, are stored in positions closer to each other (than in the case of the free location method), which makes it possible to reduce resources such as time / energy / effort required for movement when shipping and to increase throughput related to shipping. Effect of the Invention

[0022] According to each aspect of the present invention, by arranging items in a multi-level automated warehouse or on storage shelves based on logistics co-occurrence, it becomes possible to perform various picking tasks accurately, quickly, and efficiently. [Brief description of the drawings]

[0023] [Figure 1] FIG. 1 is a diagram showing an overview of the configuration of deep learning of logistics co-occurrence in one embodiment of the present invention. [Diagram 2] 1 is a plan view conceptually showing a configuration of a main part of a picking device according to an embodiment of the present invention; [Diagram 3] 1 is an elevational view showing a main configuration of a storage shelf according to an embodiment of the present invention. FIG. [Figure 4] FIG. 2 is an elevational view showing the main configuration of the temporary storage shelf / guide lamp / transport conveyor according to one embodiment of the present invention. [Diagram 5] 1 is a diagram showing the configuration of a main part of a removal indicator installed on a temporary storage shelf according to an embodiment of the present invention; [Figure 6] 1 is a diagram showing the main configuration of a temporary placement display device installed on a storage shelf in one embodiment of the present invention; [Figure 7] 1 is a diagram showing an overview of a picking system configuration according to an embodiment of the present invention. [Figure 8] FIG. 1 is a diagram showing an overview of a deep learning configuration according to one embodiment of the present invention. [Figure 9] 2 is a diagram showing an outline of the configuration of an item warehousing instruction unit according to an embodiment of the present invention; FIG. [Figure 10] 11 is a diagram illustrating a file layout of item storage information for each storage shelf according to an embodiment of the present invention. FIG. [Figure 11] FIG. 13 is an explanatory diagram illustrating a file layout of logistics co-occurrence of items according to an embodiment of the present invention. [Figure 12] 11 is a flowchart for explaining the storage of items having logistics co-occurrence into a storage shelf according to an embodiment of the present invention. [Figure 13] 4 is a flowchart for explaining the operation of the picking system according to one embodiment of the present invention. [Figure 14] FIG. 1 is a perspective explanatory view for explaining an automated warehouse according to an embodiment of the present invention. [Figure 15] FIG. 1 is a perspective explanatory view for explaining an automated warehouse according to an embodiment of the present invention. [Figure 16] FIG. 1 is a plan view for explaining an automated warehouse according to an embodiment of the present invention. [Figure 17]FIG. 1 is a plan view for explaining an automated warehouse according to an embodiment of the present invention. [Figure 18] FIG. 2 is a plan view illustrating the operation of the automated warehouse according to one embodiment of the present invention. [Figure 19] FIG. 1 is a plan view illustrating picking according to an embodiment of the present invention. [Figure 20] FIG. 1 is a diagram showing an overview of a deep learning configuration according to one embodiment of the present invention. [Figure 21] FIG. 2 is an explanatory diagram for explaining placement of items in an automated warehouse according to an embodiment of the present invention. [Figure 22] 1 is a diagram illustrating a file layout of parts picked from an automated warehouse according to an embodiment of the present invention. FIG. [Diagram 23] FIG. 13 is an explanatory diagram illustrating a file layout of logistics co-occurrence of items according to an embodiment of the present invention. [Figure 24] 4 is a flowchart for explaining the operation of the automated multi-level warehouse according to one embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0024] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0025] As described above, the inventors of the present application have discovered and focused on the fact that work can be carried out efficiently if logistics co-occurrence is taken into consideration when handling items handled at a logistics center or the like. First, the logistics co-occurrence of items in this embodiment will be described here. Logistics co-occurrence includes items that have a predetermined relationship among a collection of items ordered by an orderer. There are multiple orderers, and related items are examined for each orderer, and items with a co-occurrence rate higher than a predetermined threshold are automatically determined to have logistics co-occurrence.

[0026] In this embodiment, the specific value of the threshold value for judging the co-occurrence of logistics is not limited, but each logistics center, etc., experimentally determines and operates a criterion for judging the co-occurrence of logistics (that is, no matter how the threshold value is determined, it is included in the technical idea of ​​the present invention). That is, for example, it is possible to adopt or control a method of arranging items close to each other in order of the degree of co-occurrence of logistics. By doing so, items with low co-occurrence of logistics will end up being arranged at relatively distant positions from each other. There are various algorithms for realizing such an arrangement, and regardless of the algorithm, arranging items based on the co-occurrence of logistics is included in the technical idea of ​​the present invention. Furthermore, items with co-occurrence of logistics are not limited to two items that correspond one-to-one as described above, but may be, for example, a combination of three or more items. In this way, the technical idea of ​​the present invention also includes an aspect of defining the co-occurrence of logistics between any multiple items and arranging each item in a warehousing / storage position based on the judgment of the co-occurrence of logistics.

[0027] It is also possible to automatically obtain such logistics co-occurrence by machine learning. FIG. 1 is a diagram showing an overview of the configuration of deep learning of logistics co-occurrence according to an embodiment of the present invention. As shown in FIG. 1, the logistics co-occurrence determination system for goods that determines the logistics co-occurrence of goods handled in a logistics center includes an goods storage unit A1 that associates and stores goods with collection boxes (boxes in which ordered goods are collected), a data input unit B1 that accepts goods stored in the goods storage unit A1, and a data output unit G1 that predicts and outputs the logistics co-occurrence of goods estimated from goods input to the data input unit B1 using a determination model F1 generated by a model generation unit E1 that connects goods having logistics co-occurrence from the association between a plurality of goods and generates a determination model F1. The output information is stored in a goods logistics co-occurrence information storage unit D1. Here, the determination model F1 may be generated without teacher data, for example, or may be generated with teacher data.

[0028] In the data input section B1, situation information (including at least one of information related to situations such as season, regional characteristics, events, advertisements, Christmas, New Year's holidays, Obon, and consecutive holidays) may be input together with the collection box in association with the collection box. This makes it possible to define items having logistics co-occurrence by reflecting the situation information.

[0029] 2 is a plan view conceptually showing the main configuration of a picking device constituting a picking system according to an embodiment of the present invention. In the following explanation, a temporary shelf is added to the picking system, but the items put into the temporary shelf are not related to logistics co-occurrence. Note that in this embodiment, the picking system is explained including the temporary shelf, but it is not limited to a picking system including the temporary shelf, and it goes without saying that a picking system not including the temporary shelf is also included. Any type of picking system is encompassed in the technical concept of the present invention.

[0030] The facility is equipped with at least storage shelves 1a to 1f (six storage shelves are shown, but the number is of course not limited to six) for storing various items and supplying them to picking positions, a conveyor 3 for transporting collection boxes 4 (in FIG. 4, there are actually many collection boxes, but for simplicity's sake, 14 collection boxes 401 to 414 are shown), and a temporary storage shelf 5 for temporarily storing items taken out from storage shelves 1a to 1f. The space between storage shelves 1a to 1i and the conveyor 3 becomes the work space P where worker 2 performs picking. According to the conveying direction C of the collection boxes, for example, the picking areas of each worker 2k, 2l, 2m, 2n, etc. are defined as Pk, Pl, Pm, Pn, etc., and the picking zones are defined as Zk, Zl, Zm, Zn, etc., and storage shelves 1a to 1i and temporary storage shelves 5 corresponding to each are arranged.

[0031] FIG. 3 is an elevational view showing the main configuration of a storage shelf according to an embodiment of the present invention. When the storage shelves 1a to 1i are viewed from the front in FIG. 2 (the side of the picker) (i.e., from the side indicated by 5 in FIG. 2 to the side indicated by 1), for example, as shown in FIG. 3, the items are stored in four rows and three tiers of shelves (e.g., 101 to 112) for each picking zone, and a take-out indicator 6 (e.g., 601 to 612) is provided for each of the shelves, and the shelf where the items to be taken out are located is indicated (i.e., the take-out indicator 6 relating to the shelf where the items to be taken out at that time are located is highlighted, lit, or blinks, etc.), and depending on the situation, the indicator control unit 6220 (described later) indicates the direct input and the quantity to be directly input, whether to temporarily store the items, and (if temporarily stored) the quantity to be temporarily stored, and the completion of the take-out is displayed. In this case, as described above, the items may be directly input into the outgoing box without being temporarily stored.

[0032] Fig. 4 is an elevational view showing the main configuration of the temporary shelf / guide lamp / transport conveyor according to one embodiment of the present invention. When the temporary shelf 5 is viewed from the front in Fig. 2 (i.e., from side 1 to side 5 in Fig. 2), as shown in Fig. 4, for example, guide lamps 7 are disposed on the conveyor 3 that transports collection boxes 4 (collection boxes 401 to 414) to which articles are to be inserted, and the temporary shelf 5 is disposed above them. The guide lamps are disposed above the conveyor 3, and indicate the position of the collection boxes 401 to 414 into which articles are to be inserted. ByA number of such boxes are arranged in fixed positions, and the guide lamps are switched on and off in accordance with the movement of the target collection box on the conveyor (i.e., so that the timing of the movement and the timing of the light movement are approximately synchronized), and are controlled to indicate the movement position of the collection box. A temporary placement display 8 is provided under each box for each temporary storage shelf 5 (for example, one temporary storage shelf has one vertical row and four horizontal rows of boxes arranged). When a temporary placement instruction is issued on storage shelves 1a to 1i, the location of the temporary placement is indicated by the display control unit 6220 (described later), and the location where the items to be put from the temporary storage shelves 5 (for example, 501 to 504) are temporarily placed in collection boxes 401 to 414, as well as the quantity to be put in, are indicated by the display control unit 6220 (described later), and when the putting in is completed, the completion of the putting in is displayed. Also, the guide lamp 7 indicates the destination (target) of the insertion, for example, the collection boxes 401 to 414, under the control of the display control unit 6220 (described later).

[0033] FIG. 5 is a block diagram of an embodiment of the present invention. storage FIG. 6 is a diagram showing the main configuration of a removal indicator installed on a shelf according to one embodiment of the present invention. Provisional 6 is a diagram showing the main configuration of a temporary placement display installed on a shelf. The above-mentioned removal display 6 and temporary placement display 8 (e.g., 801 to 804) operation instructions are given by displays such as those shown in Fig. 5 and Fig. 6, respectively. The display operations of these displays are controlled by a display control unit 6220 (described later).

[0034] First, the storage shelves are divided into four rows and three levels for each picking zone, and the items are stored on the storage shelves. The take-out indicators 6 installed on each shelf include a direct take-out indicator 63 that indicates by lighting up that the items on the shelf where the take-out indicator 6 is located should be put directly into the collection box 4 (i.e., directly into the collection box 4 without going through the temporary storage shelf), a direct take-out quantity indicator 64 that indicates by lighting up the quantity of items on the shelf that are to be put directly into, and a target collection box for putting items on the shelf within the area covered by the worker 2. The system is equipped with a temporary placement and removal indicator 61 that lights up to indicate that the items on that shelf should be placed on the temporary shelf 5 (temporary placement) when there is no box, etc., a temporary placement and removal quantity indicator 62 that lights up to indicate the quantity of items on that shelf to be placed on the temporary placement table, a removal location indicator (bay lamp) 65 that lights up to show the location on the shelf where the items to be removed are located, whether for temporary placement or direct insertion, and a storage shelf removal completion reset button 66 that lights up when pressed by the operator when removal of the items is completed, whether for temporary placement or direct insertion.

[0035] Next, the temporary placement indicators 8 installed on each shelf of the temporary storage shelf as shown in Figure 6 are configured to include a temporary placement location indicator (bay lamp) 81 which, when a temporary placement instruction is given by the removal indicator 6, lights up to indicate which shelf of the temporary storage shelf 5 the item on the shelf where the removal indicator 6 is located will be temporarily placed on, a temporary placement input indicator 82 which lights up to indicate that items are to be inserted into the collection box from the shelf where the temporary storage shelf 5 is located, a temporary placement input quantity indicator 83 which lights up to indicate the number of items to be inserted into the collection box from the shelf where the temporary storage shelf 5 is located, and a temporary placement input completion reset button 84 which lights up when pressed by the operator when the insertion of the temporarily placed items into the collection box is completed.

[0036] The configuration of a picking system that supports the above-mentioned picking work from a system perspective will be described with reference to FIG. 7. FIG. 7 is a diagram showing an outline of the picking system configuration according to an embodiment of the present invention. As shown as an example in the figure, the upper information system 6000 is configured to include, for example, a data management computer 6100 and a data control computer 6200. The data management computer 6100 includes an inventory management section 6110, a warehousing management section 6120, and a shipping management section 6130, and the data control computer 6200 includes a conveyor control section 6210, a display control section 6220, an item replenishment control section 6230, and a collection control section 6240. The conveyor control section 6210 controls the input line, and the display control section 6220 controls the take-out display 6, the temporary placement display 8, and the guide lamp 7. When a centralized bay lamp, a picking notice display, etc. are provided, these are also controlled by the display control section 6220. The goods replenishment control unit 6230 controls the replenishment line and the storage, and the collection control unit 6240 controls the collection line and the storage. In this way, based on the vast amount of goods information stored in the upper information system, a program using a known algorithm that functionally links the conveyor 3, the above-mentioned displays 6 and 8, and the guide lamp 7 is stored in a memory unit (not shown), and the program is read from the memory unit by a CPU (Central Processing Unit) (not shown) to execute and control it.

[0037] The operations of each machine controlled by the above-mentioned higher-level information system 6000, data management computer 6100, data control computer 6200, etc. are extracted as log data by a log data extraction unit 51 and stored in a log data storage unit 53. Here, the log data includes the operation history of each machine controlled by management data, control data, etc. In this embodiment, the log data includes at least the name information of each stored item, the identification information of the item, the information of the source of storage, the information of the manufacturer, the information of the date and time of production, the cumulative number of items stored in the warehouse, the cumulative number of items removed from the warehouse, the number of items currently stored in the warehouse, the current storage location information of the target product, the date and time of storage, the date and time of picking, the amount (number) of the same item shipped at the same time, the storage shelf number information, the collection box number information, the identification information and amount (number) of other items shipped to the same shipping destination on the same day, shipping destination information, etc., and includes information that can specify from which storage shelf 1a to storage shelf 1i the target item was picked. Furthermore, the status information corresponding to the year, month, and date of the log data is acquired by the status information acquisition unit 105 and stored in the status information storage unit 107. As a result, each log data and each status information correspond one-to-one. The situation information includes at least information related to items such as internal factors such as date and locality, and external factors such as temperature, humidity, day of the week, events, advertisements, Christmas, New Year's holidays, Obon, and consecutive holidays.

[0038] The above-mentioned picking system is one example of a guide lamp type digital picking system that realizes a control system that can convert between a first picking process in which items are directly placed in a collection box from a storage shelf without going through a temporary storage shelf, and a second picking process in which items are placed in a collection box from a storage shelf via a temporary storage shelf, depending on the situation, and is not limited to this.

[0039] The above-mentioned control system further has a function of predicting logistics co-occurrence and storing the items in an even manner on storage shelves 1a to 1i. Below, machine learning for predicting a plurality of items for which logistics co-occurrence should be recognized from the items collected in collection boxes 401 to 404 will be described. FIG. 8 is a diagram showing an overview of the configuration of deep learning according to one embodiment of the present invention, and FIG. 9 is a diagram showing an overview of the configuration of an item warehousing instruction unit according to one embodiment of the present invention.

[0040] In the picking system shown in FIG. 4, which includes collection boxes 401 to 404 into which instructed items are put, a conveyor 3 for transporting the collection boxes 401 to 404, and storage shelves 1a to 1f (FIG. 2) for storing various items to be picked, the picking system includes an item storage unit 55 for storing items in association with the collection boxes 401 to 404, a data input unit 57 for receiving items stored in the item storage unit 55, and a data output unit 71 for predicting and outputting the logistics co-occurrence of items estimated from items input to the data input unit 57 using a judgment model 59 generated by a model generation unit 73 that generates a judgment model 59 by connecting items having logistics co-occurrence from the association between a plurality of items. The output information is stored in a logistics co-occurrence information storage unit 75 of items. Here, the judgment model 59 may be generated without teacher data or may be generated with teacher data.

[0041] The data input unit may input situation information (including at least one of information related to situations such as season, regional characteristics, events, advertisements, Christmas, New Year's holidays, Obon, and consecutive holidays) acquired by situation information acquisition unit 105 and stored in situation information storage unit 107 in association with collection boxes 401 to 404. This makes it possible to define items having logistics co-occurrence by reflecting the situation information.

[0042] 8 and 9, an item is stored (stored) in a storage shelf based on the logistics co-occurrence information predicted by the judgment model 59 generated by the model generation unit 73. That is, an item storage instruction unit 79 reads the logistics co-occurrence information of the items from the logistics co-occurrence information storage unit 75, and also reads the item storage information for each storage shelf from the item storage information storage unit 77. Then, an instruction is output to a storage instruction information storage unit 80 to store the logistics co-occurring items in a relatively vacant storage shelf after leveling them out.

[0043] If items with logistics co-occurrence are concentrated on one worker, and one worker puts multiple items into one collection box, while another worker puts one item into one collection box, the overall efficiency will decrease. For this reason, in the case of two types of items with logistics co-occurrence, it is preferable to have one type of item stored in one worker's storage shelf and the other type in the storage shelf of another worker. However, in the case of small items such as a yellow highlighter pen and a pink highlighter pen, which can be picked up separately with both hands at the same time, they may be stored in the storage shelf of one worker. In addition, since the number of items of each type stored in the storage shelf is known when judging logistics co-occurrence, items with a large number of items may be stored near the center of the storage shelf, for example, where the worker can easily handle them, and items with a small number of items may be stored in the corners of the storage shelf. In this way, by judging logistics co-occurrence, it is possible to efficiently handle various items and perform retrieval work. The determination model 59 generated by the model generation unit 73 is generated, for example, by clustering without using teacher data.

[0044] FIG. 10 is a diagram for explaining a file layout of item storage information for each storage shelf according to an embodiment of the present invention. As shown in FIG. 10, Table 73 is an example of a file layout of the item storage unit 55. For example, the names of items that are put together into one collection box for the same shipping destination are stored. Specifically, the names of items that are associated with and picked from the collection boxes 401 to 404 are stored. That is, the collection box 401 stores that one Christmas tree will be picked as the item name on December 24th. In addition, it stores that one towel will be picked, one pair of pajamas will be picked, two dozen soaps will be picked, 12 pairs of slippers will be picked, and one wristwatch will be picked.

[0045] In the collection box 402, the following items are stored as the names of the items to be picked on December 24th: 2 Christmas trees, 2 pieces of clothing, 5 towels, 2 pillows, 5 rolls of toilet paper, and 1 watch.

[0046] The collection box 403 stores, as items to be picked on December 24th, the following: one piece of clothing will be picked, one dozen soaps will be picked, five pairs of slippers will be picked, ten towels will be picked, and one wristwatch will be picked.

[0047] In the collection box 404, the following items are stored as the names of the items to be picked on December 24th: one wristwatch, two towels, two pairs of pajamas, and twelve rolls of toilet paper.

[0048] FIG. 11 is an explanatory diagram for explaining a file layout of the logistics co-occurrence of items according to an embodiment of the present invention. As shown in FIG. 11, Table 75 is a file layout of the logistics co-occurrence information storage unit 75 for items in which the judgment model 59 finds out the names of items with logistics co-occurrence, and the item storage instruction unit 79 stores instruction information for storing the items in a storage shelf. That is, the judgment model 59 specifies the items "towel" and "wristwatch" that have logistics co-occurrence. Also, it specifies the items "soap" and "slippers" that have logistics co-occurrence. Meanwhile, the item storage instruction unit 79 instructs to store the items with logistics co-occurrence in a storage shelf. Specifically, it instructs to store "towel" and "wristwatch" in positions close to each other in the storage shelf on December 24th. Also, it instructs to store "soap" and "slippers" in positions close to each other in the storage shelf on December 24th. In addition, by constructing a judgment model through machine learning, it becomes possible to learn through machine learning that, for example, a consumer who purchases a "watch" also simultaneously orders a "towel" (i.e., the logistics co-occurrence is above a certain threshold), and for two items whose logistics co-occurrence is defined as above a certain threshold, the system can be controlled to encourage the action of recommending the purchase of a "towel" to a consumer who has only purchased a "watch". Furthermore, by associating the logistics co-occurrence of items with dates, it is possible to predict what kind of co-occurrence will appear on special days such as Obon, Christmas, or the New Year holidays, and it is possible to make predictions about how much of an item should be purchased. In addition, by carrying out the above-mentioned processing through the above-mentioned thinking as an expanded (logistics) co-occurrence that takes into account the logistics co-occurrence and various attributes in a synergistic manner, various things can be predicted.

[0049] Fig. 12 is a flowchart explaining the operation of storing (replenishment of) items with logistics co-occurrence into a storage shelf, and the operation of updating a judgment model for multiple items whose logistics co-occurrence is equal to or greater than a certain threshold. "Picking" in the following refers to the retrieval operation when a worker moves an item from a storage shelf to a collection box. Note that since this is a simulation, no actual operation is performed.

[0050] First, the judgment model 59 reads information on each item scheduled to be picked up in the collection box 401 from the purchase order information (step S501). That is, it reads that one Christmas tree will be picked as the item name on December 24th. Similarly, it reads that one towel will be picked, that one pair of pajamas will be picked, that two dozen soaps will be picked, that 12 pairs of slippers will be picked, and that one wristwatch will be picked.

[0051] Information on the items to be picked up in the collection box 402 for the second time is read. It is read that two Christmas trees will be picked as the item names on December 24th. Similarly, it is read that two pieces of clothing will be picked, five towels will be picked, two pillows will be picked, five toilet paper rolls will be picked, and one watch will be picked.

[0052] The third time, the collection box 403 reads that one piece of clothing will be picked as the name of the item on December 24. Similarly, it reads that one dozen soaps will be picked, that five pairs of slippers will be picked, that ten towels will be picked, and that one watch will be picked.

[0053] For the fourth time, the collection box 404 reads that one wristwatch will be picked as the item name on December 24. Similarly, it reads that two towels will be picked, two pairs of pajamas will be picked, and 12 rolls of toilet paper will be picked.

[0054] Next, whether or not the items have logistics co-occurrence is determined by the determination model 59. If it is determined that the items have logistics co-occurrence (i.e., items whose logistics co-occurrence degree is equal to or higher than a certain level; the same applies below), the process proceeds to step S505. If it is determined that the items do not have logistics co-occurrence, the process proceeds to step S507 (step S503).

[0055] The judgment model 59 groups logistics co-occurrence items (i.e. items with a logistics co-occurrence degree equal to or higher; the same applies below) (step S505). Grouping refers to storing items with logistics co-occurrence together in the logistics co-occurrence information storage unit 75. For example, if there are two items of a given type, the two items are stored in one row. If there are three items of a given type, the three items are stored in one row. Items that are not grouped are determined to have no logistics co-occurrence.

[0056] It is determined whether or not items in all collection boxes have been read. That is, first, information on the item to be stored in collection box 401 is read, and second, information on the item in collection box 402 is read. Third, information on the item in collection box 403 is read, and fourth, information on the item in collection box 404 is read. If it is determined that information on the items in all collection boxes 401 to 404 has been read, the process proceeds to step S509. If it is determined that information on the items in all collection boxes 401 to 404 has not been read, the process returns to step S501 (step S507).

[0057] The goods warehousing instruction unit 79 determines a storage shelf for warehousing (step S509). The storage shelf is determined by identifying a storage shelf with vacant space, and generating information for warehousing goods with logistics co-occurrence in separate storage shelves.

[0058] Instruction information is generated to store the grouped items stored in a specified warehouse or the like in a storage shelf (step S511).

[0059] It is determined whether or not storage instruction information has been generated for items in all groups. If it is determined that storage instruction information has been generated for all groups, the process proceeds to step S515. If it is determined that storage instruction information has not been generated for items in all groups, the process returns to step S511 (step S513).

[0060] The model generation unit 73 updates the judgment model 59 (step S515). In this way, by updating the judgment model, the data for which storage has been completed also becomes accumulated data, and the judgment model can be updated to a more accurate one. Note that, although it is preferable to accumulate data and update it all at once, a method of updating it each time may also be used.

[0061] An example of the instructions on the displays 6 and 8 and the actions of the worker 2 when picking is performed using the above embodiment will be described as a series of picking work actions centered around the worker 2k in the work space Pk of the picking zone Zk in Fig. 2. Fig. 13 is a flowchart for explaining the operation of a picking system according to an embodiment of the present invention. In the following explanation, a temporary shelf is added to the picking system. In this embodiment, the picking system is explained including a temporary shelf, but it is not limited to a picking system including a temporary shelf, and may be a picking system that does not include a temporary shelf.

[0062] First, prior to picking, the worker 2k looks in the direction of the storage shelf 1 in the work space Pk, and starts by confirming that the removal indicator 6 is not operating (step S700).

[0063] When the collection boxes 402-406 transported to the picking zone Zk of the worker 2k's assigned range are target collection boxes into which articles are to be continuously put, the information on the information label indicating whether the mode is direct put-in or not is read by a sensor or the like (not shown) (step S701). When the direct put-in mode is recognized (Yes in step S701), the display control unit 6220 controls the take-out location indicator 65 and the storage shelf take-out completion reset button 66 on the take-out indicators 602-606 of the storage shelves 102-106 to light up as a picking order, and also sequentially lights up the direct put-in take-out indicator 63 indicating direct put-in and the direct put-in take-out quantity indicator 64 indicating the quantity to be directly put in (step S720). At the same time, the guide lamps 7 are sequentially lit by the control of the display control unit 6220 (step S721), so that the worker 2k can directly place the articles on the storage shelves 102-106 sequentially into the collection boxes 402-406 in accordance with these instructions.

[0064] In this way, when the direct input of the items 102-106 without passing through the temporary storage shelf 5 is completed, the storage shelf take-out completion reset button 66 is pressed in sequence by the worker 2k, and the system recognizes this in the same manner as above (step S722). The system 6000 recognizes this and, under the control of the display control unit 6220, turns off the direct input take-out indicator 63, the direct input take-out quantity indicator 64, the take-out location indicator 65, and the storage shelf take-out completion reset button 66 in sequence, and the completion of the input work of the items 102-106 is recognized (step S723). If the picking order is not completed (No in step S725), the process proceeds to the next picking work.

[0065] Next, the collection boxes 407-409 transported to the picking zone Zk, which is the area of ​​responsibility of the worker 2k, are not the target collection boxes into which items are to be placed (No in step S701), but if there are target collection boxes 410-412 into which the worker 2k is to place items in picking zones Zm and Zn, the computer 6200 lights up the pick-up location indicator 65 and the storage shelf pick-up completion reset button 66 as a picking order for the pick-up indicators 607-609 of the storage shelves 107-109 where the items to be placed are stored, and the display control unit 6220 simultaneously lights up the temporary placement pick-up indicator 61, which indicates temporary placement, and the temporary placement pick-up quantity indicator 62, which indicates the quantity to be temporarily placed, in sequence (step S710).

[0066] The worker 2k moves the articles from the storage shelves 107-109 to the temporary storage shelves 501-503 in order according to the removal indicators 607-609. When this is confirmed on the system side by, for example, a sensor not shown (step S711), the temporary storage location indicator 81, which indicates on which shelf the temporary storage indicators 801-803 will be temporarily stored, and the temporary storage input completion reset button 84 are turned on under the control of the display control unit 6220. Therefore, the worker 2k can temporarily store the articles according to the instructions.

[0067] When the temporary placement is completed, the worker 2k presses the storage shelf removal completion reset button 66 of the removal indicators 607 to 609 in the order of completion. When these are confirmed on the system side by, for example, a sensor not shown (step S711), the temporary placement removal indicator 61, the temporary placement removal quantity indicator 62, the removal location indicator 65, and the storage shelf removal completion reset button 66 are turned off in sequence under the control of the display control unit 6220, and the completion of the temporary placement work is recognized (step S712).

[0068] While this temporary placement work is being performed, the collection boxes 410-412 into which the articles from the temporary placement shelves 501-503 are to be placed arrive in sequence at the picking zone Zk, which is the area covered by the worker 2k, under the control of the conveyor control unit 6210. When this is recognized by the system side by a sensor or the like (not shown) (step S713), the guide lamps 7 are sequentially lit under the control of the display control unit 6220 (step S714). At the same time, the temporary placement placement indicator 82, which instructs the placement of articles from the temporary placement shelves 501-503 to the collection boxes 410-412, and the temporary placement placement quantity indicator 83, which instructs the quantity to be placed from the temporary placement shelves 501-503 to the collection boxes 410-412, are sequentially lit under the control of the display control unit 6220. The worker 2k can complete the placement of the articles into the collection boxes 410-412 based on this instruction / display.

[0069] The items 107-109 are put into the temporary storage shelves 501-503, and the temporary placement completion reset button 84 of the temporary placement indicators 801-803 is pressed by the worker 2k in that order. The system recognizes this in the same manner as above (step S715), and the temporary placement location indicator 81, the temporary placement input indicator 82, the temporary placement input quantity indicator 83, and the temporary placement input completion reset button 84 are of When the lights are turned off in sequence under the control of the display control unit 6220 (step S716), the system recognizes the completion of the insertion of the items 107 to 109 by a sensor (not shown) and follows the instructions of the next picking order (steps S717, S718).

[0070] The following describes how, when picking specified items 500 from the automated warehouse 100 and storing them in a collection box, the items 500 can be transported more quickly by storing items 500 that have a logistics co-occurrence in storage locations close to each other in the automated warehouse 100. In other words, the movement of the conveyors and lifts is reduced, and the time required to store and retrieve the items 500 is shortened.

[0071] FIG. 14 is a perspective explanatory view for explaining a three-dimensional automatic warehouse according to an embodiment of the present invention. Shown in the figure is a conceptual diagram of a storage rack, specifically, a storage rack 211 having storage locations 22 arranged in three tiers and five rows as a specific example.

[0072] FIG. 15 is a perspective explanatory view for explaining a three-dimensional automatic warehouse according to an embodiment of the present invention. Shown in the figure is a conceptual diagram of the three-dimensional automatic warehouse 100. Specifically, the storage racks 211 are adjacent to each other at the back of the storage surface to form a set of storage racks, and a plurality of such sets are arranged. On both sides of the entire warehouse, there are outer storage racks 211 with the storage surface on one side. The three-dimensional automatic warehouse 100 having a storage warehouse section 200 and a loading / unloading section 400 composed of a conveyor and a lift is shown. The storage warehouse section 200 includes a passage 23 for a transfer mechanism between the storage racks. In this specification, the moving directions of the article 500 etc. are described using the "horizontal direction", "vertical direction", and "depth direction" / "transverse direction" (the "depth direction" and the "transverse direction" are used with the same meaning), which respectively indicate the x direction, y direction, and z direction in FIG. 15.

[0073] FIG. 16 is a plan view for explaining a three-dimensional automatic warehouse according to an embodiment of the present invention, and is a more detailed configuration diagram of the storage warehouse section 200 of the three-dimensional automatic warehouse 100. As shown in the figure, the storage racks 211 are adjacent to each other at the back of the storage surface to form a set of storage racks, and a plurality of such sets are arranged. On both sides of the entire warehouse, there are outer storage racks 211 with the storage surface on one side. And in this case, the relationship between the length (L') of the storage rack 211 in the depth direction (z direction in FIG. 15) and the maximum dimension (D) of the article 500 etc. in the depth direction is generally in the relationship of D < L' < 2D.

[0074] FIG. 17 is a plan view for explaining a three-dimensional automated warehouse according to an embodiment of the present invention, and is a configuration example of a storage warehouse unit 2 of a three-dimensional automated warehouse 100 different from the configuration shown in FIG. 16. A plurality of independent storage racks 212 having a length (L) equal to or greater than the length (L') of the storage rack 211 in FIG. 16 and having no obstacles to the movement of articles 500 such as columns and partitions in the storage rack are arranged, and outer storage racks 211 having a storage surface on one side are arranged on both sides outside the entire warehouse. Here, the length (L) of the independent storage rack 212 is shown as being approximately twice the length of the storage rack 211, but it is not limited to this length, and it has a length capable of storing a plurality (n) of articles 500. Therefore, in the independent storage rack 212 of the present invention, the length (L) of the independent storage rack 212 and the maximum dimension (D) in the depth direction of the articles 5 are in a relationship of D < L < n × D.

[0075] FIG. 18 is a plan view for explaining the operation of the automated warehouse according to one embodiment of the present invention, and is a conceptual diagram showing an example of the flow of goods, etc. 500 in the automated warehouse 100 shown in FIG. 15. As shown in the figure, the warehouse is configured with a storage warehouse section 200 including an aisle 23 between storage racks 211 and an entry / exit section 400 including a conveyor and a lift, and on the same floor, a conveying mechanism 300 runs along the aisle 23 to convey goods, etc. 500 from a predetermined position to a predetermined position. For example, arrow A shows an example of a flow from warehousing to storage, and when the storage rack 211 to be stored is on the same floor as the third conveyor 43, the goods, etc. 500 sent from the third conveyor 43 pass through, for example, the second conveyor 42 that has moved to be located on the same floor as the storage rack 211 to be stored, and then the first conveyor 41 on the same floor as the storage rack 211 to be stored, and are loaded onto the conveying mechanism 300 and stored at a predetermined position. Arrow B shows an example of the flow from storage to delivery, and in the opposite direction to arrow A, the goods 500 stored in a predetermined position on the storage rack 211 to be stored are loaded onto the transport mechanism 300 and transported to the first conveyor 41, and then moved to the subsequent process via the second conveyor 42 that has been moved to be located on the same floor as the storage rack 211 to be stored, and the third conveyor 43 on the same floor as the storage rack 211 to be stored. Here, the subsequent process includes a process of picking out the necessary goods from the goods, etc., and then returning them to the warehouse, and a process of carrying out the empty collection boxes after picking by the conveyor. Here, the loading / unloading section 400 is equipped with multiple automatic shipping devices GTP. The automatic shipping devices GTP will be explained in detail later.

[0076] Such control is performed by a higher-level information system 6000 shown in Fig. 7. Also, a log data extracting unit 51 is configured to be able to extract log data.

[0077] In this way, in the automated multi-story warehouse 100, the conveying mechanism 300 can freely convey items, etc. 500 in the horizontal direction (x direction) of the storage rack, the lift 44 can freely convey items, etc. 500 in the vertical direction (y direction) of the storage rack, and the items, etc. 5 can be freely conveyed between the storage rack facing the aisle 23 and the conveying mechanism 300.

[0078] FIG. 19 is a planar explanatory diagram for explaining picking according to one embodiment of the present invention, and is an explanatory diagram for explaining the automatic shipping device GTP. In the automatic shipping device GTP, a worker H picks necessary items from the storage box 500 in which the items stored in the automatic warehouse 100 are stored, and collects the items in a collection box 501. That is, the worker H picks necessary items from the storage box 500 that has been removed from the automatic warehouse 100. After that, items remaining in the storage box as a result of picking are again stored in the automatic warehouse 100. At this time, if the items have logistics co-occurrence, since they are arranged in close proximity to each other as described above, it is possible to put the necessary items in the collection box in one picking instead of two pickings.

[0079] FIG. 20 is a diagram showing an overview of the configuration of deep learning according to an embodiment of the present invention. For example, a picking system including a collection box into which goods, etc. are put from a specified storage box 500, a conveyor for transporting the collection box, and a three-dimensional automated warehouse 100 for storing various goods to be picked is assumed. As shown in FIG. 20, the picking system includes a picking information storage unit 155 for storing goods, etc. in association with the collection box, a data input unit 157 for receiving goods stored in the picking information storage unit 155, and a data output unit 161 for predicting and outputting the logistics co-occurrence of goods, etc. estimated from goods, etc. input to the data input unit 157 using a judgment model 159 generated by a model generation unit 163 that generates a judgment model 159 by connecting goods having logistics co-occurrence from the association between a plurality of goods, etc. The output information about goods, etc. is stored in a logistics co-occurrence information storage unit 165. Here, an example of the judgment model 159 generated without teacher data is shown, but it may be generated with teacher data. The data input unit 157 may input situation information (including at least one of information related to situations such as season, regional characteristics, events, advertisements, Christmas, New Year's holidays, Obon, and consecutive holidays) acquired by the situation information acquisition unit 105 and stored in the situation information storage unit 107 in association with the collection boxes a to d. This makes it possible to define items having logistics co-occurrence by reflecting the situation information.

[0080] FIG. 21 is an explanatory diagram for explaining the arrangement of goods in the automatic warehouse according to one embodiment of the present invention. As shown in FIG. 21, goods, etc. are stored in the automatic warehouse 100 based on the logistics co-occurrence information predicted by the judgment model 159 generated by the model generation unit 163. That is, the goods storage instruction unit 179 reads the logistics co-occurrence information related to the goods, etc. from the logistics co-occurrence information storage unit 165, and also reads the goods storage information of the automatic warehouse 100 from the goods storage information storage unit 177. Then, an instruction to store the goods, etc. 500 having logistics co-occurrence in the same storage location 22 of the automatic warehouse 100 is output to the storage instruction information storage unit 180. The judgment model 159 generated by the model generation unit 163 is generated by clustering without teacher data. Note that a method with teacher data may be used.

[0081] FIG. 22 is a diagram for explaining a file layout of parts picked from an automated warehouse according to an embodiment of the present invention. As shown in FIG. 22, table 139 is an example of a file layout of the item storage information storage unit 177. For example, information related to the name of an item to be picked in association with a collection box is stored. That is, in the collection box a, it is stored that "one ballpoint pen" will be picked as the item name on October 24th. Similarly, it is stored that "two dozen" "pencils" will be picked, that "one eraser" will be picked, that "50 clear files" will be picked, that "one set" of "ink cartridges" will be picked, and that "one USB memory" will be picked.

[0082] In collection box b, it is stored that "2 boxes" of "binders" will be picked as the item name on October 24. Similarly, it is stored that "1 dozen" of "pencils" will be picked, "2 erasers" will be picked, "2 sets" of "ink cartridges" will be picked, "1 stapler" will be picked, and "3 USB memory sticks" will be picked.

[0083] In collection box c, it is stored that "3 sets" of "ink cartridges" will be picked on October 24. Similarly, it is stored that "2" "USB memory" will be picked, "1" "fountain pen" will be picked, "100 sheets" of "paper" will be picked, and "12 dozen" of "envelopes" will be picked.

[0084] In collection box d, it is stored that "3 ballpoint pens" will be picked as the item name on October 24. Similarly, it is stored that "50 clear files" will be picked, "2 calculators" will be picked, and "1 box" of "binders" will be picked.

[0085] FIG. 23 is an explanatory diagram for explaining a file layout of the logistics co-occurrence of items according to an embodiment of the present invention. As shown in FIG. 23, table 140 is a file layout of the logistics co-occurrence information storage unit 165 for items in which the judgment model 159 finds out the names of items with logistics co-occurrence, and the item storage instruction unit 179 stores instruction information for storing the items in the same storage location 22 of the automated multi-story warehouse. That is, the judgment model 159 specifies the items "ink cartridge" and "USB memory" that have logistics co-occurrence. Also, it specifies the items "pencil" and "eraser" that have logistics co-occurrence. Meanwhile, the item storage instruction unit 179 instructs the items with logistics co-occurrence to be stored in the same or adjacent storage locations 22 of the same multi-story warehouse. Specifically, on October 24, it instructs the multi-story warehouse to store the ink cartridge and the USB memory in the same or adjacent storage locations 22. Also, it instructs the automated multi-story warehouse to store the pencil and the eraser in the same or adjacent storage locations 22 of the same multi-story warehouse on October 24. Furthermore, by constructing a machine learning judgment model, it becomes possible to, for example, encourage a purchaser who has only purchased a wristwatch to also purchase towels. Furthermore, by relating the logistics co-occurrence of items to dates, it becomes possible to predict what kind of co-occurrence will appear on special days such as Obon, Christmas, or the New Year holidays, and it becomes possible to make predictions about how many items should be purchased. In addition, by carrying out the above-mentioned processing after considering the above-mentioned thinking as an expanded (logistics) co-occurrence that takes into account the synergistic relationship between logistics co-occurrence and various attributes, various predictions become possible.

[0086] 24 is a flowchart for explaining the operation of the automated warehouse according to one embodiment of the present invention, more specifically, the flowchart for explaining the generation of information on items with logistics co-occurrence, the storing (replenishment of items) of items in the same or adjacent storage locations 22 of the automated warehouse, and the updating of the judgment model for items with logistics co-occurrence. Note that since this is a simulation, no actual operation is performed.

[0087] First, the judgment model 159 reads information on each item scheduled to be collected into the collection box a from the order information (step S801). Specifically, it reads that one ballpoint pen will be picked as the item name on October 24th. Similarly, it reads that two dozen pencils will be picked, one eraser will be picked, 50 clear files will be picked, one set of ink cartridges, and one USB memory will be picked.

[0088] Read information about the items to be picked up in collection box B for the second time. Specifically, read that 2 boxes of binders will be picked as the item names on October 24th. Similarly, read that 1 dozen pencils will be picked, 2 erasers will be picked, 2 sets of ink cartridges will be picked, 1 stapler will be picked, and 3 USB memory sticks will be picked.

[0089] The third time, collection box c reads that 3 sets of ink cartridges will be picked as the item names on October 24. Similarly, it reads that 2 USB memory sticks will be picked, 1 fountain pen will be picked, 100 sheets of paper will be picked, and 12 dozen envelopes will be picked.

[0090] For the fourth time, collection box d reads that 3 ballpoint pens will be picked as the item names on October 24. Similarly, it reads that 30 clear files will be picked, 1 calculator will be picked, and 1 box of binders will be picked.

[0091] Next, the judgment model 159 judges whether or not the items have logistics co-occurrence (step S803). If it is judged in step S803 that the items have logistics co-occurrence, the process proceeds to step S805. If it is judged that the items do not have logistics co-occurrence, the process proceeds to step S807.

[0092] Next, the judgment model 159 groups the items that are determined to have logistics co-occurrence (step S805). Grouping refers to storing items that have logistics co-occurrence together in the logistics co-occurrence information storage unit 165. For example, if there are two items of a given type, two items are stored in one row. If there are three items of a given type, three items are stored in one row.

[0093] Next, it is determined whether or not the items in all collection boxes have been read (step S807). That is, first, information on the item to be stored in collection box a is read, and second, information on the item in collection box b is read. Third, information on the item in collection box c is read, and fourth, information on the item in collection box d is read. If it is determined in step S807 that information on the items in all collection boxes a to d has been read, the process proceeds to step S809. If it is determined that information on the items in all collection boxes a to d has not been read, the process returns to step S801.

[0094] Next, the goods warehousing instruction unit 179 determines the storage location 22 to be stored (step S809). The storage location 22 is determined by identifying a storage location 22 with vacant space, and storing the goods having logistics co-occurrence in that storage location 22.

[0095] Next, instruction information is generated to store the automatically grouped items stored in a predetermined warehouse or the like into the storage location 22 (step S811).

[0096] Next, it is determined whether instruction information for storing the items, etc. 500 of all groups has been generated (step S813). If it is determined in step S813 that instruction information for storing the items, etc. 500 of all groups has been generated, the process proceeds to step S815. If it is determined that instruction information for storing the items, etc. 500 of all groups has not been generated, the process returns to step S811.

[0097] Next, the model generating unit 163 updates the judgment model 159 (step S815). In this way, by updating the judgment model 159, data for which warehousing has been completed also becomes accumulated data, and the judgment model 159 can be updated to a more accurate one. Note that the update is performed by accumulating data and updating it all at once. This provides an effect of providing an automated warehouse system that can quickly and appropriately pick out co-occurring items, etc. 500 from the storage location 22. Note that, although the update is performed by accumulating data and updating it all at once, it may also be performed each time. In the automated warehouse of this embodiment, for example, one aisle 23 shown in FIG. 18 and a pair of shelves facing the aisle are regarded as one storage aisle space, and one GTP is provided in each storage aisle space. In such a warehouse, by storing a set of items with a high degree of logistics co-occurrence in one storage aisle space, the operation of the conveying mechanism in one aisle can be used to consolidate the set of items with a high degree of logistics co-occurrence into one GTP, thereby providing an automated multi-story warehouse system that can quickly and appropriately pick items, etc. 500 that co-occur logistics-wise from the storage location 22.

[0098] Although one embodiment of the present invention has been described in detail above, the technical idea of ​​the present application is not limited to a specific embodiment, and various modifications, changes, expansions, substitutions, etc. are possible. In addition, it is also possible to combine all or a plurality of components according to the above-described embodiment. [Explanation of symbols]

[0099] C. Conveying direction 1 Storage shelf Storage shelves installed in picking zones 101-112 Zk 2 workers (...2k2l2n2m...) 3. Conveyor belt 4 Collection box (...401~414...) 5 Temporary storage shelves Temporary storage shelves installed in picking zones 501-504 Zk Z Picking zone assigned to worker (···ZkZlZmZn···) P workspace (...PkPlPmPn...) 6 Ejection indicator 61 Temporary removal indicator 62 Temporary storage removal quantity indicator 63 Direct loading / unloading indicator 64 Direct loading / unloading quantity indicator 65 Removal location indicator (bay lamp) 66 Storage shelf removal completion reset button 601~612 Picking indicators installed in picking zones Zk 7 Guide Lamp 8 Temporary display 51 Log data extraction unit 53 Log data storage unit 55 Article storage section 57 Data Entry Section 59 Judgment model (judgment part) 71 Data output section 73 Model Generation Unit 75 Logistics co-occurrence information storage unit 81 Temporary storage location indicator (bay lamp) 82 Temporary placement indicator 83 Temporary input quantity indicator 84 Temporary placement completion reset button 105 Status Information Acquisition Unit 107 Situation information storage unit 109 Teacher Data Input Section 111 Decision Model 113 Teacher data output unit 114 Model Generation Unit 115 Log Data Analysis Unit 116 Goods quantity storage section 117 Data Entry Section 119 Data output section 121 Picking amount prediction information storage unit 123 Processing capacity memory unit 125 Matching section 127 Work instruction information output section 129 Work instruction information storage unit 155 Storage unit for picking information from multi-level shelves 157 Data Entry Section 159 Decision Model 161 Data output section 163 Model Generation Unit 165 Co-occurrence information storage unit Temporary placement indicators installed in picking zones 801 to 804 Zk

Claims

1. A logistics co-occurrence determination unit that uses a determination model generated by connecting items that have logistics co-occurrence based on the associations between multiple items using machine learning to generate a determination model, and outputs the logistics co-occurrence of the items that is estimated from the items input to a data input unit, thereby determining a logistics co-occurrence degree related to the logistics co-occurrence between items; a storage location determination unit which determines a first location close to each other in the automated warehouse as a storage location for a first plurality of items whose logistics co-occurrence degree defined by the logistics co-occurrence degree definition unit is equal to or greater than a threshold, and which determines a second location determined by a free location method as a storage location for a second plurality of items whose logistics co-occurrence degree does not satisfy the threshold; a storage means for storing the plurality of incoming items in the first or second location determined by the storage location determination unit when storing the plurality of incoming items; Equipped with The logistics co-occurrence determining unit a model generation means for generating a trained model by machine learning that specifies a degree of logistics co-occurrence between a certain item and another item, using information on substantially all types of items received in a certain period of time / number of days as input data and information on types of multiple items that are co-occurring and despatched out in a certain period of time / number of days as training data; An incoming goods receiving means for receiving information on all incoming goods to be brought in during a certain period of time / days; a logistics co-occurrence output means for defining a value outputted using the trained model generated by the model generation means as a logistics co-occurrence for the incoming goods received by the incoming goods receiving means; having In the logistics co-occurrence determination unit, situation information including at least one of seasonality, regionality, event-relatedness, and promotional nature is input to the data input unit, and a logistics co-occurrence of the items estimated from the items input to the data input unit and the situation information is output, thereby determining a logistics co-occurrence degree between items that reflects the situation information.

2. 2. The logistics warehousing system according to claim 1, wherein articles are stored in the automated warehouse based on the logistics co-occurrence degree defined by the logistics co-occurrence degree defining unit.

3. Using a determination model generated by linking together articles having logistics co-occurrence based on the associations between a plurality of articles using machine learning, the logistics co-occurrence of the articles estimated from the articles input to a data input section is output, thereby determining a logistics co-occurrence degree, which is a value related to the logistics co-occurrence between articles; For a first plurality of items whose specified logistics co-occurrence is equal to or greater than a threshold, a first location that is close to each other in the automated warehouse is set as a storage location, and for a second plurality of items whose specified logistics co-occurrence does not satisfy the threshold, a second location that is determined by a free location method is set as a storage location; When the plurality of incoming items are received, the plurality of incoming items are stored in the determined first or second location. A control method for a logistics warehouse system, comprising: The logistics co-occurrence degree is Using information on the types of substantially all items received during a certain period of time / days as input data and information on the types of multiple items co-occurring and de-stocked during a certain period of time / days as training data, a trained model is generated by machine learning that specifies the degree of logistics co-occurrence between a certain item and another item; Accepting information on all incoming goods during a certain time / day, A value outputted using the generated trained model for the received incoming goods is defined as a logistics co-occurrence degree; A control method for a logistics warehouse system, wherein situation information including at least one of seasonality, regionality, event-relatedness, and promotional nature is input to the data input unit, and a logistics co-occurrence degree between items that reflects the situation information between items is defined by outputting the logistics co-occurrence of the items input to the data input unit and the situation information estimated from the situation information.

4. 4. The method for controlling a logistics warehouse system according to claim 3, wherein the articles are stored in an automated warehouse based on the defined degree of logistics co-occurrence.

Images