JP2002154615A - Article shipment management system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To formularize each working hour for the optimization of a size of a batch of shipment order and to find a method for optimizing an SA method(Simulated Annealing Method) which is one of a combined optimization method automatically to improve shipment lead time. SOLUTION: The method in which factors (item, case number, etc.), of shipment working hour are analyzed to shorten the shipment lead time and work hour is formularized using multiple regression analysis based on these factors to apply the SA method(Simulated Annealing Method) so as to find a combination of shipment demand (shipment order) by which the shipment lead time becomes minimum using this as an evaluation function is provided to improve the shipment lead time by the introduction of a warehouse work management system program based on the method.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention aims to shorten the total work time (shipment lead time) from the storage position of each article item in a warehouse or inventory section of goods and parts to the shipping position in a cargo handling area. And an article shipment management system.

[0002]

2. Description of the Related Art In recent years, the needs of consumers have become more and more diversified, and in response to this, in shipping centers and sales warehouses, the type of shipment of small items of various types and small items by single item management has been increasing. Furthermore, in addition to the conventional baggage storage function in warehouses, the work time (shipment lead time) until shipment and delivery in response to a shipment request from a customer
Has been required as a very important function.

[0003] For example, when ordering goods using the Internet or mail order, the producer must individually deliver the goods to the consumers, and the delivery time is often problematic. Also, in convenience stores, etc., orders are made to the producer several times a day in small lots. In this case, the shipping order from the storage position of the product is appropriately grouped, the picking performed collectively, and the There is a shipping operation up to carrying out the secondary sorting at the shipping and delivering position (for each convenience store or each truck) at the shipping and delivering position, and these warehouse shipping operations also need to be performed in a short time. (This is excluded this time because the delivery company will load the trucks after the secondary sorting.)

Further, at a parts inventory site at a production site,
With the adoption of the so-called "Toyota Kanban system" or the like, it is necessary to dispatch goods in a short time and in a timely manner in response to a shipment request.

On the other hand, it has been adopted to reduce the work time (shipment lead time) from the product storage position to shipping / delivery by installing an automatic warehouse or the like.
Such an automatic warehouse system has a very large equipment cost and is not applicable to any other than a large-scale warehouse. Therefore, in warehouses that do not have a hardware configuration such as automated warehouses, material handling equipment such as transport trolleys and conveyors can be installed, and these equipment configurations, equipment layouts can be improved, and warehouse work management system programs can be introduced to improve shipping lead times. However, no significant improvement effect was obtained.

For example, Japanese Unexamined Patent Publication No. Hei 9-118312 discloses a method of automatically sorting, picking, packing, and shipping goods received at a distribution center accurately and promptly according to the order of a client without manpower. Attach the product number,
1 item is set on a shelf with 1 product placed in 1 tray
Each storage space is stored by computer using shuttle work etc., and it is taken out of the storage space using shuttle work etc. and packed according to the order slip (shipping form) stored in the computer. A system for automatically sending a shipment number to a lane according to a shipping area and shipping the same.

In Japanese Patent Application Laid-Open No. 2000-142926, in a shipping process in which various products having different shapes and dimensions are picked, a shipping package made of those products is created, temporarily stored, and then shipped, the shipping process is made more efficient. In order to measure this, the management processing unit that collects packing work result data and temporary storage result data, specifies the temporary storage area for temporarily storing shipments based on the collected data, and the storage space for the specified temporary storage area A temporary storage area change unit is provided for storing the shipped product in another temporary storage area when the product cannot be stored due to a shortage. Further, a progress display unit is provided for displaying the packing work result data and the temporary storage result data managed by the management processing unit in comparison with the vehicle allocation plan data.

Further, in Japanese Patent Application Laid-Open No. Hei 6-286823, in order to automatically pick up a plurality of types of articles in a short time, the central control device stores data on the articles to be handled in a storage device. Based on this, a stacking layout corresponding to a shipping request from the input device is determined, and a command is issued to each device via the communication device. And various other techniques have been proposed. However, such a conventional technique aims to increase the efficiency and time of the picking work of articles, but does not attempt to shorten the entire shipping lead time.

Next, in order to understand the shipping lead time, a shipping work flow of a distribution center or a commercial warehouse, which is a prerequisite of the present invention, will be described with reference to FIG. First, a shipping order from a customer (convenience store, consumer, head office) for a product stored in a warehouse is received online by facsimile, telephone, and terminal, and the shipping order is summarized at regular time intervals such as every evening and every morning. . Picking is performed for each consolidated shipping order, and truck dispatch is arranged. After the picking is completed, it is conveyed to a cargo handling place as needed. At the cargo handling area, customers,
Secondary sorting of luggage is performed for each truck. Then, the cargo is loaded onto the truck in order from the one where the secondary sorting is completed and the truck arrives, and the truck is shipped.

[0010] Here, as a point for solving the problem of shortening the shipment lead time, it is conceivable to optimize the wrapping of the shipment order at regular time intervals. Therefore, the shipment lead time is analyzed in order to optimize the closing of the shipment order. The shipping lead time mainly consists of the following three work times. (1) Picking work time for performing total picking by grouping a large number of shipping orders compiled by a customer into a plurality of shipping orders (bundling). (2) Pallet piled by the picking is handled at a handling area. (3) Secondary sorting time for sorting packages packed for each of a plurality of shipping orders into individual customer orders and trucks

Therefore, when the shipping lead time is composed of the total time of (picking work time + transport time + secondary sorting time), the relationship between the size of the shipment order and the above three types of work time is generally as follows. There is a proportional relationship. The larger the bundled amount of the shipping order (the larger the number of products per bundle), the shorter the picking work time and the shorter the transport time, but the longer the secondary sorting time. On the other hand, the smaller the bundled amount of the shipping order (the smaller the number of products per bundled unit), the longer the picking work time and transport time, but conversely, the shorter the secondary sorting time. That is, the squeezed amount of the shipping order, the picking work time, and the transport time are inversely proportional, and the secondary sorting time is directly proportional.

Therefore, by finding an appropriate squeezing amount such that the total shipping lead time of each of the three types of work time is minimized, the shipping lead time can be minimized. However, this deciding decision has traditionally relied on shipping workers' years of intuition and experience.

[0013]

SUMMARY OF THE INVENTION In view of the above-mentioned problems of the prior art, the present invention provides a method for optimizing the size of a shipment order by formulating each operation time and a combination optimization method. The SA method (Simulated
It is an object of the present invention to find a method of automatically optimizing by an annealing method (pseudo annealing method) and to improve a shipping lead time.

[0014]

In order to solve the above-mentioned problems, the present invention analyzes the factors (items, cases, etc.) of the shipping work time to reduce the shipping lead time, and performs multiple regression from these factors. The working time is formalized using the analysis, and this is used as an evaluation function to determine a combination of shipping requests (shipping orders) that minimizes the shipping lead time.
Method A (Simulated Annealing Me)
It is a gist of the present invention to propose a method to which the third method is applied and to improve a shipping lead time by introducing a warehouse work management system program based on the method.

That is, the present invention reduces the total work time of shipping work from the storage position of each article item in a warehouse or stocking department of goods and parts (hereinafter referred to as articles) to the secondary sorting in a shipping and handling place. In an article shipment management system, shipping orders are stored in a database at every predetermined time interval such as every morning and every evening, and shipping orders are combined for each combination of shipping orders properly selected from the storing means. It is characterized in that the operation time for each work element is obtained, and in the comparison of the total work time, a calculation means for obtaining an almost optimum combination of sizing and the shipping order is sieved based on the almost optimum combination of sizing. . As such an order storage means, a built-in or external storage means of a personal computer such as a hard disk or an MO is sufficient.
The operation of the personal computer with a ron of 500 MHz and a memory of about 128 MB and a central control device are sufficient.

It is preferable that the calculating means for obtaining the almost optimum combination of groups use not a strict optimum solution but a calculating means to which a combination optimizing method for obtaining a practically optimum solution (quasi-optimal solution) is applied. Well, for example, the factor analysis of the shipping work element, quantification analysis of each work time and formulate,
Using this, arithmetic means applying a pseudo-annealing method, which is one of the combinational optimization techniques, more specifically, the arithmetic means generates an initial solution for the closing of the shipping order, and then, based on the initial solution, The total work time obtained by calculating the work time of the shipping work element is set as a pseudo-optimal time, and then a new near solution for the closing of the shipping order is newly generated at random, and the shipping work element of the shipping work element is generated based on the near solution. The total work time calculated by calculating the work time is compared with the pseudo-optimal time, and a practically optimum solution (sub-optimal solution) is obtained while sequentially updating the pseudo-optimal solution based on a predetermined adoption probability. It is preferably an arithmetic means.

[0017] Further, the shipping work element includes a picking work element for carrying out packing in accordance with the consolidation for each shipping order, and a transporting work for carrying a group of articles packed on a pallet or the like to a cargo handling place. A work element and a secondary sorting work element for secondary sorting the packed articles for each transport medium such as a truck or for each customer, so that the total work time of each of the three types of work elements is substantially minimized. In addition, it is preferable that the calculating means is an arithmetic means to which an optimization technique is applied.

Also, the article shipment management system is not an automatic warehouse in which an automatic loading / unloading system, an automatic transport trolley, an automatic sorting machine, etc. are incorporated, but a manual or forklift.
It is preferable that the system is an article shipment management system intended for a warehouse where picking, transporting, and secondary sorting are performed by a cart.

[0019]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below in detail with reference to an embodiment shown in the drawings. However, unless otherwise specified, dimensions, shapes, relative arrangements, and the like of the components described in this embodiment are not intended to limit the scope of the invention, but are merely illustrative examples. Embodiments of the present invention will be described step by step. In an automatic warehouse in which material handling equipment (automatic loading / unloading system, automatic transport trolley, automatic sorting machine, etc.) is incorporated, the working time can be estimated in some detail from the equipment performance. Is intended for warehouses in which material handling equipment is not installed and picking, transporting, and secondary sorting are performed manually or using a forklift or cart.

Procedure 1: Quantitative analysis of each working time of picking, transporting and secondary sorting Picking from the storage position, transporting to the secondary sorting site, 2
First, a method of quantification analysis and formulation based on actually measured values in order to obtain an estimated value of each operation time of the next sorting operation will be described.

First, factors affecting the working time are extracted.
Factors ・ Average number of workers (W) ・ Number of transport vehicles (T)
-Number of pallets and plastic containers (Pl)-Item storage location (L)-Number of customers included in the group (S)-Number of items included in the group (I)-Number of cases included in the group (C) (case: Cardboard boxes, wooden boxes,
The number of pieces (P) (pieces: the minimum unit of a certain item) included in the grouping, and the like. However, symbols in parentheses after each factor represent variables in a working time estimation formula described later. The item generally indicates a product type such as “dry beer” and “orange juice”.

From the factors affecting these work times, the calculation of the estimation of each element work time for each grouping of the shipping order is performed by a multiple regression model. For example (1) is an example of picking time _{t 1.} t ₁ = (a · S + b · f (L, IC + c · C + d · P + e) / W (1) a, b, c, d, e: parameters to be determined by the estimation formula f: inter-item movement time function

[0023] An example of a correlation diagram between the actual measurement value and the estimated value at the picking time t ₁ in FIG. The horizontal axis is the measured value,
An estimate vertical axis based on the working time estimation equation of t _1, which plots the data for each bundling thereto. In this example, the estimation accuracy of the working time estimation formula is high, with the correlation coefficient (R) being R = 0.9445, which indicates that the estimation accuracy is high.

[0024] Further, the conveying operation time t ₂ affects the moving distance (D) working time or the like between the average velocity (v) items and secondary sorting field truck and transport vehicle number (T) together with the variables above based on factors, determined by the further secondary sorting based working with variables mentioned above similarly applies to the time t ₃ to factors that influence the working time of such ships number of tracks (Tr) regression model. For example, transport work time t ₂ = Σ (D1 + D2... + Dn) × v / T (2) Secondary sorting work time t ₃ = (a ′ · (S−1) × (b′P + c ′ · C + d ′ · I + e ′) ) / W (3) a ′, b ′, c ′, d ′, e ′: parameters to be determined by the estimation formula

Procedure 2: An optimization algorithm for finding a bundle of the shipping order that minimizes the shipping lead time. According to the procedure 1, the picking work time t ₁ , the transport work time t ₂ , and the transport work time t ₂ are determined by quantification analysis based on the actually measured values. since was determined each work time estimation equation of the following sorting operation time t _3, shipment lead time will be described optimization algorithm for obtaining the bundling shipment order which minimizes This is applied.
In order to find the optimal solution from a large number of shipping orders to the warehouse, it takes a huge amount of calculation time to find the combination of all shipping order bundles, calculate all the shipping lead times, and find the optimal solution. Is not practical. For this reason, S is one of the combinational optimization techniques for finding a practically optimal solution (suboptimal solution) instead of finding an exact optimal solution.
Method A was applied. The merits of this combination optimization method include the following two. A suboptimal solution can be found in a relatively short time from a huge number of solutions. Since it is possible to escape from the local optimal solution, which is common in combinational optimization problems, it is possible to search for a global optimal solution.

Next, FIG. 4 shows a flow chart of an optimization algorithm when this SA method is applied to a warehouse shipping operation. Hereinafter, the outline procedure of this processing will be described with reference to FIG. (S1) Generating an initial solution of the bundle of the load order: Generate an initial solution of the bundle of all the shipping orders, which is summarized at a time interval every morning and every evening. As a method of generating an initial solution, for example, in order to reduce the calculation time, it is considered advantageous to "start the calculation from a position close to the optimal solution" as the initial solution. Also, when arranging shipping orders, it is considered reasonable to find out in advance items having common items as much as possible. (S2) Creation of virtual list: A picking list and a secondary sorting list are created based on the initial solution of the shipping order created in (S1). (S3) Work time calculation: The work time is calculated by applying each work time estimation formula from the picking list and the secondary sorting list created in (S2). (S4) Working time: The working time obtained in (S3) is set as a pseudo-optimal solution. (S5) Combination calculation: randomly generate a new order closure (neighboring solution). (S6) Virtual list: A picking list and a secondary sorting list are created based on the shipping order created in (S5). (S7) Work time calculation: The work time is calculated by applying each work time estimation formula from the picking list and the secondary sorting list created in (S6). (S8) Comparison: The work time obtained this time is compared with the pseudo optimal solution up to the previous time. At this time, the current value is updated to the pseudo optimal solution with the following adoption probability (p). At this time, in the adoption probability (p), if the current value is larger than the previous pseudo-optimal solution, the current value is unconditionally adopted if Δ <0, but even if Δ ≧ 0, e− ^{Δ / T} Use the current value with probability. e− ^Δ / T: Δ ≧ 0, where Δ = (current working time in the neighborhood solution) − (working time in the pseudo-optimal solution) T = variable (T = e− ^{β · N} ) (β: constant, (N: number of calculations) (S9) Checking the number of calculations: Checking whether the number of calculations has reached a predetermined number. (S10) Variable calculation: The number of calculations is increased by one, and the variable T at this time is calculated.

Here, when the SA method is applied to the combinatorial optimization problem, the initial values of variables (in the case of the SA (pseudo annealing) method, the temperature is the object), the number of repetition calculations, the setting of end conditions, the initial solution Consideration is needed in the method of generating neighborhood solutions. In this study, trial calculation was performed to determine the initial value, number of calculations, and termination conditions. As for the method of generating the initial solution and the neighborhood solution, the respective solutions are generated in the following procedure.

(1) Method of Generating Initial Solution In order to shorten the calculation time, rather than giving a completely random order of the shipping order as the initial solution, it is better to “start the calculation from a place close to the optimal solution” as the initial solution. Seems to be more advantageous. In the case of grouping shipping orders, it is conceivable that the picking time can be shortened by grouping items having the same common item as much as possible. Thus, the calculation time is shortened by extracting in advance the shipping orders having common items, and arranging these items as an initial solution. That is, in order to find a common order for merchandise items, a matching rate threshold (α) based on a predetermined threshold is defined, and orders having a high matching rate are grouped. The matching rate (α ') is a constant peculiar to the warehouse. After a certain optimal solution is obtained, the matching rate (α') is adjusted based on the following equation 1 so as to be close to the order closure at that time. decide.

[0029]

(Equation 1)

FIG. 5 is a flowchart for determining the initial combination of shipping orders by the matching rate method. First, every morning,
Store order data (orders) for each store compiled at time intervals such as every evening into a database. (S21) Next, one is extracted from all the store-specific order data (orders) collected for each time interval taken into the database and is set as a master. (S22) The orders are taken out one by one from the orders that are not bundled, and the matching rate is calculated based on the above formula. (S2
3)

If the calculated matching rate (α ′) is smaller than the threshold value (α), the above operation is repeated until the calculated matching rate (α ′) exceeds the threshold value (α). (S24) When the calculated matching ratio (α ′) exceeds the threshold value (α), and when the number of cases such as boxes and bags of the item is equal to or less than the maximum number of cases, this order is assigned to the same group as the master. Wrap it up. (S25) The above operation is repeated until the number of cases reaches the maximum number of cases. (S26) When the maximum number of cases has been reached, in order to create a group of another group, one of the orders not grouped is selected.
One is taken out as a master, and the above operation is repeated to create a group of another group. (S27)

Then, it is determined whether or not all the orders have been viewed (S28). Next, it is determined whether or not all the orders have been bundled, and then the generation of the initial solution of the bundle of the shipping orders is terminated.
(S29)

(2) Method of Generating Neighborhood Solution The following procedure is adopted in order to generate a bundle of shipping orders as a neighborhood of a random and pseudo-optimal solution as a neighborhood solution of the SA method (calculation of the above (S5)).

(S31): From among all shipping orders,
One shipping order is selected at random. (S32): One bundle is randomly selected from the bundles of all shipping orders. (S33): The shipping order selected in (S31) is placed in the bundle of the shipping order selected in (S22).
If the shipping order is the same, a new shipping order is generated, and the shipping order selected in (S31) is entered therein.

(Embodiment) An application example of the algorithm for optimizing the shipment lead time by the SA method will be described. First, the warehouse shipping conditions applied are shown below. -The number of shipping orders: 172-The number of shipping items: 227-The number of shipping cases: 1,964-The number of shipping pieces: 744-The number of shipping areas: 35-The initial solution is based on the case where the shipping orders are bundled in units of shipping areas (current state). Adopted as. (In this example, the initial solution was set for each shipping area for comparison with the current situation.)

FIG. 6 shows an example of the calculation result. In this figure, the horizontal axis shows the number of calculations, the vertical axis shows the picking work time,
The table shows a transport time, a secondary sorting operation time, a shipping operation time (shipment lead time) totaling these, and the number of times of picking. The calculation results are shown below.

The number of bundled shipment orders: 71 (average 2.42 order / bundled) Shipping work time: shortened by about 2 hours Calculation time: about 10 minutes Computer used: Cellron 500 MHz, 128 MB
Memory, development language: Visual C ++ 6.0

Further, the SA method according to the embodiment of the present invention, the conventional total picking for each shipping direction (initial solution in FIG. 6), and a shipping work time when order picking is performed for each shipping order as a comparative example. FIG. 7 shows a comparison diagram of (shipment lead time). As can be seen from this figure, the SA method of the present embodiment was able to find the optimal shipping order squeezing in which the working time was shorter than the conventional total picking for each area or order picking for picking for each shipping order.

[0039]

As described above, according to the present invention, for the purpose of shortening the shipping lead time of the warehouse, the factor analysis of the shipping work and the quantification analysis of each working time are formalized, and the combination is used to optimize the combination. As a result of applying the SA method, which is one of the generalization methods, and conducting a case study, it is possible to automatically generate a shipping order wrapper that can reduce the work time compared to conventional total picking and order picking in a short time. did it. The present invention can also be applied to various types of warehouse work including material handling automation equipment.

[Brief description of the drawings]

FIG. 1 is a shipping work flow diagram of a distribution center or a commercial warehouse, which is a premise of the present invention.

FIG. 2 shows a relationship between each work time of a picking work time t ₁ , a transfer work time t ₂ , and a secondary sorting work time t ₃ and the number of rounds of a shipping order; FIG. 4 is a graph of an optimization algorithm for calculating.

FIG. 3 shows a picking time t which is one of shipping operation elements.
₁ shows an example of a correlation diagram between an actually measured value and an estimated value in FIG.

FIG. 4 is a flow chart of an optimization algorithm for applying the SA method to a warehouse shipping operation to determine an optimal combination of shipping orders.

FIG. 5 shows a flowchart of determining an initial combination of shipping orders by a matching rate method.

FIG. 6 shows an example of a specific calculation result in an application example of an algorithm for optimizing a shipment lead time by the SA method.

FIG. 7 shows the SA method according to the embodiment of the present invention, the conventional total picking for each shipping area (initial solution in FIG. 6), and a shipping operation time (shipping) when order picking is performed for each shipping order as a comparative example. Lead time)

Claims (6)

[Claims] 1. Article shipping for shortening the total work time of shipping work from a storage position of each article item in a warehouse or stocking department of goods and parts (hereinafter referred to as articles) to a secondary sorting at a shipping handling place. In the management system, at every predetermined time interval such as every morning, every evening, etc., an order storage means for storing the shipping order in the database during the predetermined time interval; An operation means for obtaining a work time and, in comparison of the total work time, an arithmetic means for obtaining an almost optimum combination of sizings; and an article shipment management system for performing sizing of a shipping order based on the almost optimum combination of sizings. . 2. The arithmetic means for obtaining the almost optimal combination of groupings is not a strict optimal solution but an arithmetic means to which a combination optimization technique for obtaining a practically optimum solution (suboptimal solution) is applied. The article shipping management system according to claim 1, wherein: 3. A shipping work element, wherein a picking work element for carrying out packing in accordance with a wrapping for each shipping order, and a transport for carrying a group of articles packed on a pallet or the like to a cargo handling place. A work element and a secondary sorting work element for secondary sorting the packed articles for each transport medium such as a truck or for each customer, so that the total work time of each of the three types of work elements is substantially minimized. And an arithmetic unit to which an optimization method is applied.
Article management system according to the above. 4. An arithmetic unit to which the optimization method is applied,
The arithmetic means for performing a factor analysis of the shipping work element and a quantification analysis of each work time, formalizing the formula, and applying a pseudo-annealing method, which is one of combination optimization techniques, using the formula. An article shipping management system according to claim 1. 5. A total operation time obtained by calculating an operation time of the shipping operation element based on the initial solution after generating an initial solution of a shipping order by applying the optimization method. Is the pseudo-optimal time, and then a new near solution for the order of the shipping order is newly generated at random, and the total working time and the pseudo-optimal time obtained by calculating the working time of the shipping work element based on the near solution are calculated. 2. An article shipping management system according to claim 1, wherein the calculating means calculates a practically optimum solution (sub-optimal solution) while sequentially updating the pseudo-optimal solution based on a predetermined adoption probability. system. 6. The article shipment management system is not an automatic warehouse in which an automatic entry / exit system, an automatic transport trolley, an automatic sorting machine, etc. are incorporated, but a manual or forklift.
2. The article shipment management system according to claim 1, wherein the article shipment management system is a goods shipment management system intended for a warehouse where picking, transporting, and secondary sorting are performed by a cart.