JP2001034321A - Method and device for preparing optimum production plan - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prepare a production plan capable of averaging loads in each unit of packet in both of a production assembling process and a source process to be a bottleneck. SOLUTION: The load is averaged in each unit of packet in both of a product assembling process and a source process to be a bottleneck on the basis of order receiving information necessary for preparing a production plan 11 consisting of plural processes, basic data of a product included in the order receiving information, the basic data of a unit 10 constituting the product and source process information for preparing the unit 10.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for creating an optimum production plan in which a load is leveled in bucket units in both a product assembling process and a bottleneck source process. .

[0002]

2. Description of the Related Art In recent years, a scheduling system has attracted attention as a means for automatically creating a production plan that maintains a high operation rate of existing equipment and reduces the number of times of switching while complying with order delivery dates.

[0003] In the past, the level of research and development was limited, but in recent years the cost and speed of computers have been dramatically reduced, and software technologies that support automatic planning such as artificial intelligence and simulation methods have been developed. With its completion, many practical scheduling system packages have been proposed.

[0004] A production planning method using a conventional scheduling system will be described below. 7, 1 is an input device for inputting order information, 2 is a storage device for storing data, 3 is a central processing unit, 4 is a display device for displaying results, and 5 is a printing device for printing results. Reference numeral 6 denotes order data in which the order information input by the input device 1 is registered for a certain period of time, 7 denotes equipment surplus capacity data, 8 denotes a product master, 11 denotes a created production plan, and 22 denotes a production plan drafting means.

Next, the operation will be described. First, the person in charge registers order information received by telephone, facsimile or the like in the order data 6 using the input device 1. Note that the order information includes product data, delivery date information, quantities, and the like necessary for production planning. The registration result is accumulated for a certain period of one day to one month.
The order data 6 and the surplus capacity data 7
The production plan 11 is created by the production plan drafting unit 22 based on the product master 8 and the product master 8. In the production plan 11, display on the screen and printing on paper are performed using the display device 4 and the printing device 5.

[0006] The production plan drafting means 22 drafts a production plan to complete the production within the due date in consideration of the load obtained by multiplying the quantity included in the order data 6 and the takt time included in the product master 8.

[0007]

SUMMARY OF THE INVENTION However, the above-mentioned production plan creating apparatus aims at creating a single-step production plan for producing an order. For example, creation of an assembly process production plan for a customer order or creation of a source process production plan for an assembly process production plan.

In recent years, it has been required to reduce the total production lead time from the source process to the assembly process, and the need to create a production plan as a whole factory has been increasing.

The procedure for creating a factory-wide production plan using the above-described production plan creation apparatus is as follows. 1) Create an assembly process production plan for a customer order 2) Create an assembly process production plan as an order and create a source process production plan for it 3) If there are multiple source processes, repeat step 2) When the production capacity of the process is balanced, a factory-wide production plan can be created by the above procedure.

However, in an actual production site, the production capacity is often not balanced between the assembling process and the source process. In such a state, if a production plan is created by the above procedure, the production plan for the assembling process is created. Even if possible, the following problems occur in the production planning of the source process. 1) Delay in order delivery of assembly process due to insufficient production capacity 2) Order of assembly process is advanced in anticipation of delay of source process 3) As a result of advance, load of source process is concentrated in the first half,
Insufficient equipment capacity for the entire process can be resolved by extending the operation, adding equipment, or outsourcing production.However, load leveling within a certain period can be achieved without considering the above process production planning stage. Can not. However, the conventional production plan creation device does not consider the load situation on the source process side, so that load leveling is impossible.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned conventional problems, and provides an optimum production plan creation method and apparatus for creating a production plan in which a load is leveled in bucket units in both a product assembly process and a bottleneck source process. It is intended to provide.

[0012]

According to the present invention, there is provided an optimum production plan creating method comprising: order information necessary for creating a production plan of a plurality of processes; basic data on a product included in the order information; Based on basic data about the unit and information on the source process that produces the unit, load leveling is performed for each bucket in both the product assembly process and the bottleneck source process. It is possible to create a load-balanced production plan for both source processes. Here, the bucket is a unit of a period in which orders are allocated, for example, on a weekly basis, and according to the present invention, orders can be allocated by leveling the load for each bucket. Note that the size of the bucket differs depending on the factory to which the production plan creation method is applied.

More specifically, a process of calculating the assembly process load and the bottleneck process load for all the order information, and a process of converting the load values of the assembly process and the bottleneck process into deviation values to match the variances. A process of calculating a load evaluation value obtained by combining the deviation values of both processes with respect to all order information using a weight coefficient W; a process of assigning an order to a bucket using the load evaluation value; The bucket allocation for obtaining the optimum load level is obtained in the process of obtaining the level and the process of returning to the process of calculating the load evaluation value by updating the weight coefficient W to improve the load level.

More specifically, it is created at the following stage.

1) An assembly process load is calculated for each order information.

Here, the assembling process load is obtained by obtaining an assembling standard man-hour from basic data on a product and multiplying it by the number of orders as shown in (Equation 1).

[0017]

(Equation 1) 2) Calculate the bottleneck process load for each order information.

Here, the load of each unit and part obtains a list and the number of units and parts produced in the bottleneck process from the basic data on the product, and then obtains the machine tact from the basic data on the unit and part. Then, as shown in (Equation 2), it is obtained by multiplying the number of orders, the number of members, and the machine tact for each of the units and parts. Next, the bottleneck process load on the order information as in (Equation 3) is obtained by summing the values in (Equation 2).

[0019]

(Equation 2)

[0020]

(Equation 3) 3) Match the load distribution state in the assembly process and the bottleneck process.

The assembly process load is generally a minute unit because of the man-hour. In addition, the source process has a tact time, and the load unit does not coincide with the number of dots in the mounting process. Also, since the load variation is different,
(Equation 1) and (Equation 3) cannot be regarded as having the same variance. Therefore, each is converted into a deviation value with respect to the original load set. For example, a conversion formula relating to the assembly process load is as shown in (Equation 4). With such a conversion, the loads of the assembly process and the bottleneck process are both 50 on average and 1 variance.
It becomes 0.

[0022]

(Equation 4) 4) A load evaluation value is calculated for each order information.

Here, the load evaluation value is given by
The deviation value of the assembly process load (Equation 1) and the deviation value of the bottleneck process load (Equation 3) are combined using a weight coefficient W. W takes a value in the range of 0 to 1. In the present invention, this W is an operation variable.

[0024]

(Equation 5) 5) Assign orders to buckets using the evaluation values in the following procedure.

5-1) Select the order with the largest load evaluation value among the unassigned orders.

5-2) Allocate the order selected in 5-1) to the bucket having the smallest load evaluation value.

5-3) Return to 5-1) until all orders are allocated to buckets. Through this process, rough leveling with the load evaluation value is performed.

6) Calculate the load level as a result of the allocation.

Here, the load level is obtained by (Equation 6). In equation (6), LA _k is the load of the assembly process bucket k, LP _k is the load of the bottleneck process bucket, and S _LA
Is the deviation value of LA _k and S _LP is the deviation value of LP _k . That is, the load level indicates how much the leveling of the load evaluation value performed in 5) is leveled with the actual load value.

[0030]

(Equation 6) 7) Update the value of W to improve the load level, and return to step 4).

Here, since the relationship between W and the load level is as shown in FIG. 4, it is difficult to determine W that minimizes the load level using a general optimization method. W is determined by the following procedure.

Preprocessing 1) W is changed at a constant width from 0 to 1, and the above procedures 4) to 6) are executed for each W.

Preprocessing 2) The value W at which the load level is minimized is set as an initial value for optimization.

(Optimization process) The value of W is updated using the simulated annealing method (SA method), and the optimum value of W is determined.

If the fixed width in the pre-processing 1 is reduced, a more appropriate optimized initial value can be obtained in the pre-processing 2, but the amount of calculation required for the pre-processing increases. FIG.
The relationship between W and the load leveling indicated by depends on the characteristics of the assembling process and the bottleneck process to be optimized, and therefore does not change significantly even if the order receiving status changes. Therefore, there is no problem even if the method of setting the fixed width is determined by an empirical rule. If the load leveling optimization in the preprocessing stage is sufficient, there is no problem even if the bucket allocation state at that time is set as the optimal solution.

Further, the SA method is an optimization method that stochastically accepts an increase in the objective function, that is, the load leveling level, when sequentially finding a neighborhood solution, and is effective in preventing convergence to a local solution. . However, in the case of multi-modality, since the optimal solution is highly likely to depend on the initial value, setting of the initial value by a rough optimization search such as preprocessing is effective.

Further, the optimum production plan creating device of the present invention
Input device for order information, product basic data, unit basic data, storage device for constantly storing these data, display device for displaying the created result, printing device for printing the created result, input device, storage And a central processing unit having a production plan creation means for executing the production plan creation method according to claim 1 or 2, and receiving order information from an input device. And a production plan in which the load is leveled in bucket units is created by the above-described production plan creation method using the product and unit basic data stored in the storage device, and the production plan can be understood by a person in charge by a display device or a printing device. Is displayed on the screen or printed on paper. Further, the data is passed as data to a scheduling system based on a conventional production planning method, and is used as original data for creating a daily sequence plan.

[0038]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of an optimum production plan creating method and apparatus according to the present invention will be described below with reference to FIGS. The bottleneck process will be described below by taking a mounting process as an example. In the case of mass production products, it is necessary to perform integrated production of the assembling process and the mounting process. Therefore, in principle, there is no bottleneck process in addition to the mounting process. Further, in the case of mass production products, the printed circuit boards are also integrated into one by a dedicated LSI or the like. However, in the case of high-mix low-volume production, a dedicated LSI cannot be developed in terms of cost, so that the number of printed boards per product is ten or more, and the mounting process often becomes a bottleneck process.

In FIG. 1 showing the configuration of the production plan creating apparatus of the present embodiment, 1 is an input device for inputting data, 2 is a storage device for storing data, 3 is a central processing unit, and 4 is a result. , A display device for displaying the results, a printing device for printing the results, 6 for order data, 8 for a product master storing basic data including the standard assembly man-hours for each product, and 9 for a unit serving as a bottleneck process for each product And a component / configuration master storing component configurations, 10 is a unit / component master storing basic data of units and components, 11
Is a created production plan, and 12 is a production plan creating means.

The operation of the production plan creating apparatus having the above configuration is described as follows.
This will be described with reference to FIG. 2 showing a production plan creation processing flow. First, monthly order information is acquired as order data 6 from an order management system or the like. The items included in the order data that are necessary for creating a production plan include a product number, an order quantity, and a delivery date. As for the delivery date, the production completion date is used as a condition for setting from the beginning of the month to the delivery date. From the end of production to shipment, it will be in stock at the factory.

A product master 8 in which the order data 6 and information on the product part number included in the order information are registered, a unit / part configuration master 9 in which the relation between the product part number and the unit / part in the bottleneck process is registered, and a unit / part Based on the unit / part master 10 in which the information has been registered, the load value calculation means 13 calculates a load evaluation value for each order.

The load value is calculated for each of the assembly process and the bottleneck process. The load value of the assembly process is calculated by multiplying the order quantity of the order data 6 by the standard man-hour of the product master 8. First, the load value of the mounting process is calculated based on the order data 6 and the product number based on the component configuration master 9 into the printed circuit board part numbers, and the required quantity is calculated by multiplying the ordered quantity and the number of persons for each developed order. I do. Then, a product obtained by multiplying the required amount by the mounting points of the unit / part master 10 is summed up for all the developed printed circuit boards. The total score is used as a load value in the mounting process.

Next, in order to match the load value distributions in the assembling process and the mounting process, the respective load values are set to the assembly load deviation value 15a and the mounting load deviation value 1 by the distribution matching means 14.
Convert to 5b. The above process is the pre-process for creating the production plan.

Next, a weight coefficient W for minimizing the load level is determined by the following processing. First, the initial value of W is determined by the weight coefficient initial value determining means 16. FIG. 3 is a flowchart of the weight coefficient initial value determining means 16. W value from 0 to 1
Up to a fixed width, and bucket allocation means 1 for each W
9, the orders are allocated to the buckets, and the allocation result is calculated by the load level calculating means 20 to obtain the load level. Among them, W that minimizes the load level is set as the initial value of the optimization.
If the fixed width is set to 0.1, the bucket assignment and the load leveling calculation are performed 11 times. FIG. 4 is a graph of the calculation result. In this case, W =
0.1 is adopted as an initial value.

FIG. 5 is a flowchart of the bucket allocating means.
The bucket allocating means first allocates an order whose delivery date of the order data 6 is other than the end of the month to the corresponding bucket. This is an operation in which an order with a designated delivery date is preferentially assigned to a bucket for achieving the delivery date. Here, the process of assigning to the bucket is performed using the assembly process load value converted into the deviation value and the mounting process load value converted into the deviation value. After allocating all the delivery-order-specified orders, until there are no other unallocated orders, the non-allocated orders are allocated to the buckets with the smallest total load evaluation value already assigned, from the one with the largest load evaluation value.

The load level calculating means 20 divides the value obtained by dividing the load level by the average of the variance of the total assembly man-hours per bucket in the assembling process, and the variance of the total number of mounting points per bucket in the mounting process by the average. Calculate by adding values.

In the end judgment 17, if the leveling at the stage of determining the initial value of W is sufficient, the allocation result based on W obtained here is adopted as the result. If further leveling is required at the initial value stage, the value of the fixed width used for the initial value search may be reduced.

If the leveling at the initial value is not sufficient, the load leveling optimizing means 18 optimizes the load leveling from the initial value of W obtained by the weighting coefficient initial value determining means 16, that is, W which minimizes the load leveling. Ask for. FIG. 6 is a flow chart of the load level optimizing means. The optimization of W is performed by a simulated annealing method (SA method). This SA method is an optimization method for avoiding convergence to a local solution by stochastically accepting an increase in the objective function in the optimization process. Further, since the optimization is performed by sequentially obtaining the neighborhood solution, a derivative is not required. When the processing is completed under the optimization end condition, the load level at the stage of the initial value and the load level at the end are compared. This is a countermeasure when the local solution converges to another local solution from the starting point and the level of the local solution is degraded from the starting point.

With the above processing, a production plan can be made in which the load is leveled in bucket units in both the assembling process and the mounting process. The production plan by the above method is a level for determining which bucket is to be produced for an order. Therefore, based on this plan, we will first clarify orders for excess capacity, eliminate excess capacity by taking measures such as adding capacity by extending operation and transferring to outside, and then use the existing scheduling system to plan equipment and daily production sequence. Will be drafted. By using the two devices in combination in this manner, it is possible to create an optimal production plan for the entire factory.

[0050]

According to the method and apparatus for creating an optimum production plan of the present invention, as described above, a production plan is created in which the load is leveled in bucket units in both the product assembling process and the bottleneck source process. Can be. In addition, because the load of the bottleneck process is leveled, the excess capacity is clarified, and by extending the operation or transferring the excess capacity to the outside, the production plan can be executed within the allocated bucket period Becomes In other words, avoiding an increase in production lead time by moving the allocation of orders that exceed the source process capacity, which is the bottleneck, to the surplus load before the initially allocated bucket, and the delivery time for the product assembly process due to excess capacity Delays can be avoided.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a production plan creation device according to an embodiment of the present invention.

FIG. 2 is a flowchart of a production plan creation process in the embodiment.

FIG. 3 is a flowchart of a weight coefficient initial value determining unit in the embodiment.

FIG. 4 is a relationship diagram between a weighting factor and a bucket load level in the embodiment.

FIG. 5 is a flowchart of a bucket allocating unit in the embodiment.

FIG. 6 is a flowchart of a load leveling optimization unit in the embodiment.

FIG. 7 is a configuration diagram of a conventional production plan creation device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Input device 2 Storage device 3 Central processing unit 4 Display device 5 Printing device 6 Order data 8 Product master 9 Part configuration master 10 Unit / part master 11 Production plan 12 Production plan creation means 13 Load evaluation value calculation means 14 Distributed matching means 15a Assembly load deviation value 15b Mounting load deviation value 16 Weight coefficient initial value determination means 18 Load levelness optimization means 19 Bucket allocation means 20 Load levelness calculation means

 ────────────────────────────────────────────────── ─── Continued on the front page (72) Inventor Koichi Narahara 1006 Kazuma Kadoma, Kadoma City, Osaka Prefecture F-term in Matsushita Electric Industrial Co., Ltd. 3C042 RJ02 RJ03 RJ20 RL01 RL12 5H269 AB17 AB27 BB08 EE11 NN17 QB15 QD06

Claims (3)

[Claims] An order information necessary for creating a production plan of a plurality of processes, basic data on a product included in the order information, basic data on a unit constituting the product, and source process information on a unit to be produced are included. An optimal production plan creation method based on equalizing the load in bucket units in both the product assembly process and the bottleneck source process. 2. A process of calculating an assembly process load and a bottleneck process load with respect to all order information, a process of converting load values of the assembly process and the bottleneck process into deviation values and matching variances, Calculating the load evaluation value by combining the deviation values of both processes with the order information using the weighting factor W, allocating the orders to the buckets using the load evaluation value, and determining the load levelness of the allocation result. 2. A bucket allocation that obtains an optimum load level in a step of obtaining and a step of returning to a step of calculating a load evaluation value by updating a weight coefficient W to improve the load level. The optimal production plan creation method described. 3. An input device for receiving order information, product basic data and unit basic data, a storage device for constantly storing these data, a display device for displaying a created result, and a printing device for printing the created result. And a central processing unit having a production plan creation means for executing the production plan creation method according to claim 1 or 2, wherein the central processing unit manages the input device, the storage device, and the printing device in an integrated manner. Planning equipment.