CN113674057A

CN113674057A - Batch optimization method based on hierarchical aggregation strategy

Info

Publication number: CN113674057A
Application number: CN202110952691.5A
Authority: CN
Inventors: 刘健军; 梁宝林; 彭乘风; 毛宁; 陈庆新
Original assignee: Guangdong University of Technology
Current assignee: Guangdong University of Technology
Priority date: 2021-08-19
Filing date: 2021-08-19
Publication date: 2021-11-19
Anticipated expiration: 2041-08-19
Also published as: CN113674057B

Abstract

The invention discloses a batch optimization method based on a hierarchical polymerization strategy, which constructs a batch rule according to production related data provided by enterprises, improves the utilization rate of plates, improves the production efficiency, batches orders to optimize production, and replaces a batch mode with the problems of low efficiency, long time consumption and the like caused by original manual arrangement.

Description

Batch optimization method based on hierarchical aggregation strategy

Technical Field

The invention relates to the field of order batch optimization, in particular to a batch optimization method based on a hierarchical polymerization strategy.

Background

In the furniture industry, the traditional mass and single-variety production mode is no longer suitable, and a plurality of enterprises turn to large-scale customized production. Large-scale customization has become a common mode of production in the furniture industry, particularly the panel furniture industry, which has grown enormously in recent years. Due to the customization requirements of customers, the schemes designed by designers are all universal, but the schemes are decomposed into cabinets with various functions, and the cabinets are actually assembled by different numbers of boards with different materials, colors and sizes. Thus, application conditions can be provided for large-scale customization of a large number of different orders. At present, the panel furniture industry also faces a number of challenges. On one hand, the furniture production technology content is not high, the admission threshold is low, the industry competition is intense, and the furniture produced in the market is also unsmooth. On the other hand, furniture manufacturers need to make better production plans and production modes due to the constraint of capacity.

Due to the increasingly higher personalized demands of the market, the orders of customers are complicated, and personalized production often brings two problems: namely, an increase in the non-working time of the process and waste of raw materials. How to carry out batch production on a large number of customer orders so as to improve the utilization rate and the production efficiency of plates is a problem to be solved urgently by enterprises. Under the background, the data of the invention is based on a certain furniture manufacturing plant, the limitation in the actual production and the constraint brought by the specific quantitative relation brought by materials, equipment and processes are considered, the batch optimization strategy meeting the actual needs of enterprises is explored and designed, and the production efficiency and the utilization rate of plates are improved.

Scholars at home and abroad carry out a great deal of research on the problem of order batching, however, the research on order batching in domestic research is applied to the furniture industry less, and the research on the combination of batching and stock layout problems is common in plate product workshops. The study of keen and zhanghao et al on the furniture order batch has less considering capacity constraint, focuses on plate color and plate thickness, and does not consider the area and the number of components. Most of domestic and foreign researches combine order batching with stock layout, and consider the laminated plate cutting of plates, namely, a plurality of raw materials are stacked together, and the plates are simultaneously cut at one time to obtain a plurality of same parts. However, the laminated plate cutting has high requirements on the process, the cutting accuracy needs to be ensured, the requirements on a workbench and equipment are high, the cutting can be stably carried out, all parameters are kept consistent, and the laminated plate cutting can be considered only when enough similar parts in the batch are available for enterprises. The solving process of the foreign research order batch strategy is generally complex, sometimes the optimal solution is difficult to find, and the customization is not obvious.

In order to respond to the requirements of customized and multi-variety small batches of the market, the batches of most enterprise orders are formed according to manual arrangement, and a relatively long time is consumed. Furniture orders cannot be put into production quickly and are not optimal in terms of board utilization. Aiming at the order batching requirements caused by the limitations of production fields, equipment and personnel capacity in the production field scheduling of panel furniture production enterprises, how to divide production orders into a plurality of order sets meeting the production requirements based on production scheduling plans is a difficult problem to be solved urgently by department enterprises. The plate product is a product formed by assembling a plurality of plate type accessories after plane processing by taking a plate as a main raw material. If the plate furniture factory can not realize the balance of the production plan and the operation plan, the plate furniture factory easily causes the problems of serious waste of the plate, incapability of delivering the plate according to the schedule, reduction of economic benefit and the like.

Disclosure of Invention

The invention aims to provide a batch optimization method based on a hierarchical aggregation strategy, which replaces a batch mode with the problems of low efficiency, long time consumption and the like caused by the original manual arrangement according to production related data provided by enterprises, and improves the production efficiency.

In order to realize the task, the invention adopts the following technical scheme:

a batch optimization method based on a hierarchical polymerization strategy comprises the following steps:

step 1, inputting an order information table and a component information table; inputting the quantity of components in the batch, the quantity of orders in the batch and the upper and lower limit ranges of the batch number;

step 2, acquiring the input part quantity and order quantity, and calculating the maximum part quantity standard in a batch and the maximum order quantity standard in the batch;

step 3, carrying out data preprocessing on the input order information table and the input component information table, and constructing an order-component information array, an order-component information decision array, a material-order information decision array and an order selection decision matrix;

step 4, judging whether the current batch has the batched orders/components, if not, selecting the first order according to the order selection decision matrix; otherwise, turning to step 5;

step 5, selecting materials of the distributed orders in the current batch, finding order numbers related to the materials from the non-batch orders according to the order-component information decision array, calculating the order purity, and constructing an order-purity matrix; completing the selection of the order according to the order-purity matrix;

step 6, deleting the batched orders from the order-component information array, and correspondingly finishing updating the order selection decision matrix, the order-component information decision array and the material-order information decision array;

step 7, counting the number of orders and the number of components in the current batch, judging whether the number of the components exceeds the standard of the maximum number of the components in the batch and the standard of the maximum number of the orders in the batch, and if the number of the orders exceeds the standard of the maximum number of the orders in the batch, taking another batch;

step 8, counting the number of completed orders, judging, and completing batching if the number of completed orders is equal to the total number of orders; otherwise, returning to the step 4.

Further, the calculating a maximum number of parts in a batch standard and a maximum number of orders in a batch standard includes:

counting the total number of parts in the current production order from the order information table:

the actual maximum batch number is min (input end maximum batch number, total number of parts in current production order/minimum number of parts in input end batch);

the actual minimum batch number is max (input end minimum batch number, total number of parts in current production order/maximum number of parts in input end batch);

the maximum part quantity standard in the batch is the total quantity of parts in the current production order/the actual minimum batch quantity;

the minimum part quantity standard in the batch is the total quantity of parts in the current production order/the actual maximum batch quantity;

the actual maximum batch order number is the total order number/the actual minimum batch number;

the actual minimum batch order number is the total order number/the actual maximum batch number;

the maximum order quantity standard in batch is min (maximum order quantity in input end batch, total quantity of parts in current production order/minimum part quantity in input end batch);

the standard of the maximum number of the parts in the batch and the standard of the maximum number of the orders in the batch are used as the basis of starting another batch.

Further, the order-component information matrix construction method comprises the following steps:

obtaining order data from the order information table, searching and counting each order information, and obtaining the component information belonging to the current order: material information, height information, width information, area information, part type number and part combination number, and stored in the form of cells.

Further, the construction method of the order-component information decision array comprises the following steps:

acquiring order numbers and indexing: order quantity, order placing material information, order placing component size information and component quantity; and then, searching a component set placed in the order by using the material type according to the material information placed in the order, counting the quantity of components using the current material type and the total area of the components, and generating an order-component decision matrix, wherein the matrix comprises the following information: order number, order quantity, material number, quantity of parts using the same material for placing the order, area corresponding to the quantity of the parts, variety quantity of the materials, and total quantity of the order parts.

Further, the method for constructing the material-order information decision matrix comprises the following steps:

acquiring a component detail table, extracting an order number from the component detail table according to the material number, and performing de-duplication processing on the order number to acquire the number of different order numbers;

and constructing a matrix, wherein information in the matrix comprises a material number, a current material order placing number set, the number of components placed according to the corresponding order number, the number of components using the material under the order number, the number of orders not counting repeated orders, the number of orders using the current material as a unique material, the total area of the components using the material under the order, and the total area of the components using the current material as the unique material.

Further, in the material-order information decision matrix, a material number is used as a row, and an order number set, a component number, an order number without counting repeated orders, an order number, a component total area and a component use total area are used as columns;

and matching according to the order number and the material number in the matrix creating process, determining the number of parts using materials of the current material-order, the total number of parts corresponding to the order, the total area of the parts using the current material and the total area of the parts placed in the order, judging whether different orders are the only using orders of the current material in the statistical process, entering the order number of the only material, and otherwise, not recording the order in the number of the only material orders.

Further, the order selection decision matrix construction method comprises the following steps:

retrieving from a material information matrix through a material number, using an order of the material, and placing the number of components using the material in the order; the material is used and is the only material order;

calculating the material order layer purity: order quantity of orders that have used the material and are the only material order/order quantity of orders that have used the material;

calculating the part layer purity of the material: retrieving from the material-order information decision matrix a total number of parts related to placing the material order, a total number of parts related to placing the material order using the material; the purity of the component layer is the total quantity of the components using the material related to the material order/the total quantity of the components related to the material order;

calculating the area layer purity of the material: retrieving from the material-order information decision matrix a total area of components involved in placing a material order, the total area of components involved in placing a material order that have used the material; the area layer purity is the total area of the part using the material related to placing the material order/the total area of the part related to placing the material order;

the material numbers are taken as columns, and the following steps are taken: the order quantity, the part quantity, the order grade material purity, the part grade material purity and the area grade material purity of the related materials are listed, the material numbers are traversed, and the materials are stored in sequence.

Further, the selecting of the first order according to the order selection decision matrix includes:

selecting the material with the largest number of related parts through a formulated rule strategy, searching order information according to the material information, and calculating the purity of the searched order information to construct an order-purity matrix; and (5) comparing the purity of the materials, and selecting the order with the highest purity at the current stage.

Further, the order-purity matrix construction method comprises the following steps:

according to the order number, the material number of the order and the number of the parts using the same material correspondingly are retrieved from the order-part decision array, and the order purity is calculated

The order purity calculation method comprises the following steps: counting the number of parts using the materials of the distributed orders in the current batch by retrieving the material numbers of the orders and the number of parts corresponding to the same materials from the order-part decision array; the order purity is the part quantity of the current order using the material of the allocated order in the current batch/the total part quantity of the current order; and (4) finishing the construction of the order purity matrix by taking the order number, the order purity and the total number of parts placed in the order as columns.

Further, when the selection of the order is completed according to the order-purity matrix, the following strategies are adopted:

firstly, selecting an order with highest correlation purity; second, when there are multiple orders of the same highest purity, the order with the largest number of parts is selected.

Compared with the prior art, the invention has the following technical characteristics:

aiming at large-scale personalized plate-type furniture enterprises with capacity constraint, the utilization rate of plates is improved, and the production efficiency is improved; the orders are batched, and a batching rule is constructed, so that the production is optimal, and the method has important significance for the furniture industry.

Drawings

FIG. 1 is a schematic flow diagram of the process of the present invention;

FIG. 2 is a schematic diagram of the number of orders placed by the prior art solution in the example of the 6.2 production date lot number;

FIG. 3 is a schematic diagram of the order quantity of the present embodiment at the production date lot number of 6.3;

FIG. 4 is a schematic view of the number of parts of the present embodiment under the production date lot number of 6.4 in the example;

FIG. 5 is a schematic view of the number of parts of the present embodiment under a production date lot number of 6.5 in one embodiment;

FIG. 6 is a schematic diagram showing the number of used sheets in the prior art on the production date lot number of 6.6 in the example;

FIG. 7 is a schematic diagram showing the number of used sheets in the embodiment of the present invention under the production date lot number of 6.7;

FIG. 8 is a graph showing the outturn rate of the prior art protocol in the example under the batch number of 6.8 production dates;

FIG. 9 is a schematic diagram of the output rate of the present embodiment under the production date lot number of 6.8.

Detailed Description

Referring to the attached drawings, the invention provides a batch optimization method based on a hierarchical aggregation strategy, which is used for solving the problem of batch optimization of orders; in one embodiment of the invention, the application problem is described as follows:

in the manufacturing process of an enterprise, orders needing to be arranged are preferentially selected, for example, 200 to 500 orders are formed into one batch according to 25 to 35 orders in each batch or the number of parts is used as batch basis, and then professional typesetting optimization software is adopted for calculating typesetting in each batch. At present, the batching of orders of a plurality of enterprises is completed by planning personnel, which may take several days and several hours at the fastest, and the emergency orders cannot be put into production immediately, thus consuming manpower; the present invention is designed to solve this problem.

The method comprises the following steps:

step 1, inputting an order information table and a component information table; the upper and lower limits of the number of parts in the batch, the number of orders in the batch and the number of lots are input (items are freely configured by the planning personnel).

Wherein: order information table information: 1. numbering the order; 2. the number of orders; 3. number of parts category placed by order; 4. the number of parts to be merged; 5. total number of parts placed in the order;

component information table information: 1. numbering the order; 2. a part number; 3. part material numbering; 4. part size-height; 5. part size-width; 6. part size-area.

And 2, starting the batch algorithm, acquiring the quantity of the parts and the quantity of the orders input into the algorithm interface, and calculating the maximum part quantity standard in the batch and the maximum order quantity standard in the batch.

The number of allocated orders is set to 0, and the maximum number of parts in a batch standard and the maximum number of orders in a temporary batch standard are calculated as follows:

Step 3, in order to establish a better data mapping relationship and facilitate data input/output during data reading, data preprocessing needs to be performed on the input order information table and the component information table, therefore, the phenomenon that the algorithm spends a large amount of time in data storage to generate meaningless running time in the actual execution process is avoided (for example, if all workpieces under a certain order number are searched each time based on original order information, the searching of the set scale of the whole workpieces needs to be firstly found, but only the retrieval of the order number index position needs to be carried out after primary processing, and the time spent by the program in a searching scene can be reduced to a certain extent);

step 3.1, order-component information matrix construction

The order-component information matrix construction method comprises the following steps:

TABLE 1 order-parts information matrix example

1	2
		[11；12]	[1；3]
[52；53]	[51；53]
		[330；633]	[330；633]
[766；765]	[766；765]
		[252780；484245]	[252780；484245]
1	1

The examples illustrate that: the order with the number of 1; 11 and 12 parts need to be produced; wherein the material of the component 11 is 52, the height is 330, the width 766 and the area 252780, and the number of the combined components is 1; where part 11 had a mass of 53, height of 633, width 765, area 484245, and part merge number of 1.

Step 3.2, order-component information decision array

And indexing according to the order number: order quantity, order placed material information, order placed part size information (height, width, and area), and part quantity; searching a component set placed in the order according to the material type and counting the quantity of the components using the current material type and the total area of the components according to the material information placed in the order; generating an order-component decision matrix containing information including: 1. numbering the order; 2. the number of orders; 3. numbering materials; 4. the number of parts using the same material is placed for this order; 5. an area corresponding to 4; 6. the variety and quantity of materials; 7. total number of parts ordered.

TABLE 2 order-component information decision matrix example

1	2	3
			1	1	1
[52；53]	52	52
			[1,1]	15	9
[252780；484245]	3682668	299020
			2	1	1
2	15	9

The examples illustrate that: order number 1: the order quantity is 1, using both 52 and 53 materials, where the order places part quantity using 52 material as 1, corresponding to area 252780; wherein the number of parts using 53 materials for this order is 1, and the corresponding area is 484245; the order used 2 materials, 2 parts.

Step 3.3, Material-order information decision array

The method for constructing the material-order information decision matrix comprises the following steps:

constructing a matrix, wherein information in the matrix comprises 1 material number, 2 current material order number set, 3 part number under corresponding order number, 4 part number using the material under order number, 5 order number (no repeated order), 6 order number using the current material as unique material, 7 total area of parts using the material under order, 8 total area of parts using the current material as unique material,

the material numbers are used as rows in the matrix, and the order number set, the number of components, the number of orders (not counting repeated orders), the number of orders, the total area of the components and the total area of the components are used as columns;

TABLE 3 Material-order information decision matrix example

Material numbering	5	6	7
				Current material order number set	[248；249]	252	[354；356]
Number of parts under corresponding order number	[86,166]	92	[50,39]
				Number of parts using the material under order number	[1,4]	8	[5,3]
Quantity of orders (not counting duplicate orders)	2	1	2
				Order quantity with current material as unique material	0	0	0
Total area of parts using the material when placing an order	1404750	486432	11974698
				Component use with current material as sole materialTotal area of	0	0	0

Step 3.4, order selection decision matrix

calculating the material order layer purity, and the method comprises the following steps: order quantity of an order that uses the material and is the only material/order quantity of an order that uses the material.

Calculating the purity of the component layer of the material by the following steps: retrieving from the material-order information decision matrix a total number of parts related to placing the material order, a total number of parts related to placing the material order using the material; the part layer purity is the total number of parts that have used the material in relation to the material order/the total number of parts in relation to the material order.

The method for calculating the area layer purity of the material comprises the following steps: retrieving from the material-order information decision matrix a total area of components involved in placing a material order, the total area of components involved in placing a material order that have used the material; area layer purity is the total area of the part that uses the material in relation to placing a material order/the total area of the part in relation to placing a material order.

Taking the material number as a column (the column of the matrix represents the material number), and taking 1 as the order number related to the material; 2. number of parts related to the material; 3. purity of the order-level material; 4. part-level material purity; 5. and (4) traversing the material numbers and storing the materials in sequence, wherein the material purities of area grades are rows.

TABLE 4 order selection decision matrix example

2	10	11
			90	9630	1442
1	0	0
			1	0.093043	0.074896
1	0	0

Material number 2 (in the second column) relates to the order quantity 10 of the material; number of parts related to material 9630; the purity of the order-level material is 0; 4. part-grade material purity 0.093043; 5. the purity of the material in area level was 0.

Step 4, judging whether the current batch has the batched orders/components, if not, selecting the first order according to an order selection decision matrix (selecting the materials with the most related components through a formulated rule strategy, retrieving order information according to the material information, performing purity calculation on the retrieved order information, constructing an order-purity matrix, performing material purity comparison, and selecting the order with the highest purity at the current stage); otherwise, go to step 5.

Construction of the order-purity matrix:

and searching the material number of the order and the quantity of the parts using the same material correspondingly from the order-part information decision array according to the order number, and calculating the order purity.

The order purity calculation method comprises the following steps: counting the number of parts using the materials of the distributed orders in the current batch by retrieving the material numbers of the orders and the number of parts corresponding to the same materials from the order-part decision array; the order purity is the part quantity of the current order using the material of the allocated order in the current batch/the total part quantity of the current order;

numbering with 1, order; 2. the purity of the order; 3. and the total number of the parts placed in the order is columns, and the construction of the order purity matrix is completed.

Step 5, selecting materials of the distributed orders in the current batch, finding order numbers related to the materials from the non-batch orders according to the order-component information decision matrix, calculating the order purity of the order numbers, and constructing an order-purity matrix; completing the selection of the order according to the order-purity matrix (and through a formulated rule strategy, firstly, selecting the order with the highest correlation purity, and secondly, selecting the order with the largest number of parts when a plurality of orders with the same highest purity exist);

step 6, deleting the batched orders from the order-component information array, and correspondingly finishing the updating of the order selection decision matrix, the order-component information decision array and the material-order decision array

And 7, counting the order quantity and the part quantity of the current batch, judging whether the order quantity standard in the batch and the maximum order quantity standard in the temporary batch are exceeded, and if the order quantity standard in the temporary batch is exceeded, taking another batch.

Example (b):

the experimental data relates to the actual 6 days of the factory, the total number of the orders is 2387, 116 plates are related, and the total number of the parts is 102006.

TABLE 5 comparison of the utilization rates of the manual batch scheme and the present scheme

Date of production	Raw batch protocol	Batch optimization algorithm	CPU operation time(s)
				0604	0.8814	0.8932	3.5199
0605	0.8556	0.8643	2.3432
				0606	0.8427	0.8500	2.98704
0607	0.8354	0.8391	2.7054
				0608	0.8710	0.8965	3.0937
0609	0.8474	0.8462	2.7185

By comparing the total outturn rate of the scheme, the result obtained by the rule algorithm is superior to the batch scheme given by the case, and when the target value of the batch scheme is calculated, the operation time of a CPU (Central processing Unit) of the outturn rate is only 2-4 s.

And counting the number of the components of the original batch scheme, wherein 16 batches of the original batch scheme No. 6/month and No. 4 are calculated, and 20 batches are obtained by calculating a batch optimization algorithm.

The proposed batch decision method can control the product quantity of each batch more uniformly, and the scene that the product quantity of a certain batch is excessive can not occur; the effect of the decision-making method in balancing the number of batch parts is less pronounced than controlling the number of products-because the decision-making method releases more weight to the control of the number of products.

Compared with a case scheme, the batch material variety distribution of the scheme of the batch decision method is more discretized, and meanwhile, the number of the used material varieties is not positively correlated with the number of the plates used in the current batch, because the batch decision method has a certain 'material emptying' principle, namely when the material is in the current batch, all orders using the material are preferentially selected to be aggregated and subjected to priority operation, and the medium-scale and small-scale material frequency clusters account for a large proportion.

The number of batches was significantly increased over the case data: the number of batches was greater than the number of batches given the case at 6 production days, according to the batch decision method. Under a single production day of batch, the usage amount of the plates obtained according to the batch decision method is more balanced: compared with a case scheme, most of the decision scheme is concentrated in a certain area range except a few data of the obvious overflow interval range.

The above embodiments are only used to illustrate the technical solutions of the present application, and not to limit the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present application and are intended to be included within the scope of the present application.

Claims

1. A batch optimization method based on a hierarchical polymerization strategy is characterized by comprising the following steps:

2. The batch optimization method based on the hierarchical aggregation strategy according to claim 1, wherein the calculating of the maximum number of components in the batch standard and the maximum number of orders in the batch standard comprises:

3. The batch optimization method based on the hierarchical aggregation strategy as claimed in claim 1, wherein the order-component information matrix construction method is as follows:

4. The batch optimization method based on the hierarchical aggregation strategy as claimed in claim 1, wherein the order-component information decision matrix is constructed by:

5. The batch optimization method based on the hierarchical aggregation strategy as claimed in claim 1, wherein the material-order information decision matrix construction method is as follows:

6. The batch optimization method based on the hierarchical aggregation strategy according to claim 1, wherein in the material-order information decision matrix, a material number is used as a row, and an order number set, a component number, an order number without counting repeated orders, an order number, a total area of components, and a total area of component use are used as columns;

7. The batch optimization method based on the hierarchical aggregation strategy according to claim 1, wherein the order selection decision matrix is constructed by:

8. The batch optimization method based on the hierarchical aggregation strategy according to claim 1, wherein the selecting of the first order according to the order selection decision matrix comprises:

9. The batch optimization method based on the hierarchical aggregation strategy according to claim 1, wherein the order-purity matrix is constructed by the following steps:

10. The batch optimization method based on the hierarchical aggregation strategy according to claim 1, wherein when the selection of the order is completed according to the order-purity matrix, the following strategies are adopted: