CN104636871B - A kind of control method of the single phase multi-product batch processing based on data - Google Patents
A kind of control method of the single phase multi-product batch processing based on data Download PDFInfo
- Publication number
- CN104636871B CN104636871B CN201510070318.1A CN201510070318A CN104636871B CN 104636871 B CN104636871 B CN 104636871B CN 201510070318 A CN201510070318 A CN 201510070318A CN 104636871 B CN104636871 B CN 104636871B
- Authority
- CN
- China
- Prior art keywords
- order
- processing
- scheduling
- time
- production
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
- 238000012545 processing Methods 0.000 title claims abstract description 163
- 238000000034 method Methods 0.000 title claims abstract description 54
- 238000004519 manufacturing process Methods 0.000 claims abstract description 76
- 238000012549 training Methods 0.000 claims abstract description 30
- 230000002068 genetic effect Effects 0.000 claims abstract description 24
- 238000009826 distribution Methods 0.000 claims description 14
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 3
- 238000004364 calculation method Methods 0.000 description 10
- 238000005457 optimization Methods 0.000 description 9
- 238000012360 testing method Methods 0.000 description 7
- 238000003754 machining Methods 0.000 description 6
- 238000010923 batch production Methods 0.000 description 5
- 210000000349 chromosome Anatomy 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000035772 mutation Effects 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 241000135164 Timea Species 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 239000012634 fragment Substances 0.000 description 1
- 238000010353 genetic engineering Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 108090000623 proteins and genes Proteins 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- General Factory Administration (AREA)
Abstract
The invention discloses a kind of control method of the single phase multi-product batch processing based on data, comprise the following steps:Scene is handled for different target batch, takes out corresponding history operation and dispatching information, according to the difference of production scheduling task in the production scheduling cycle, establishes SMSP scheduling cases respectively;A kind of heuristic process equipment allocation rule is given, the optimal Order Processing sequence of SMSP scheduling cases is then solved using genetic algorithm;Case and optimal Order Processing sequence are dispatched using SMSP, builds training dataset, training obtains Order Processing sequential forecasting models;Batch processing order is handled using Order Processing sequential forecasting models, Order Processing is obtained and arranges sequence;Sequence is arranged using heuristic process equipment allocation rule and Order Processing, by batch processing Order splitting to different equipment, is produced and processed.The present invention can be in the rapid development difference production cycle production scheduling scheme, and there is preferable optimizing scheduling effect.
Description
Technical Field
The invention relates to the field of process industrial control, in particular to a control method for single-stage multi-product batch processing based on data.
Background
To better satisfy customer requirements, the production and processing tasks of the process enterprises are usually performed in batches, and therefore, the problems of determining the size and quantity of the batches, the release date and the delivery date are faced.
The batch processing process is an important component in the production process of the process enterprise, and has the advantages of small product batch, multiple varieties, high added value, small equipment investment, quick production updating and the like, so the batch processing process is widely applied to the process enterprises of petrochemical industry, pharmacy, food and the like, the batch processing process is effectively scheduled, the flexibility of the production process can be enhanced, resources can be saved, the cost is reduced, and the enterprises can obtain better economic benefits.
A typical batch process often consists of a series of manufacturing operations, each of which may be performed on one or more manufacturing facilities, each of which may have a different processing capacity, and may process different types and batches of products. At present, the batch production scheduling process in a process enterprise mostly belongs to a single-stage multi-product scheduling (SMSP) process.
The existing SMSP scheduling optimization solving method usually adopts a mathematical programming method or a genetic algorithm and other complex optimization methods which are often complex to realize, a large amount of parameter setting work needs to be carried out, the optimization solving time of the method is explosively increased along with the increase of the scheduling problem scale, and when the scheduling problem scale is large, the methods often cannot meet the real-time requirement of the production processing process due to overlong solving time.
In order to solve the defect that the existing single-stage multi-product batch processing scheduling optimization solving method has large limitation in the practical application process, a scheduling scheme of multi-product batch processing needs to be optimized.
Disclosure of Invention
The invention provides a control method for single-stage multi-product batch processing based on data, aiming at a single-stage multi-product batch processing production processing scene, the production scheduling schemes in different production periods can be rapidly formulated, and a better scheduling optimization effect is achieved.
A control method of data-based single-stage multi-product batch processing comprises the following steps:
(1) and aiming at different target batch processing scenes, corresponding historical production scheduling data is taken out from the historical production data, and SMSP scheduling cases are respectively established according to different production scheduling tasks in a production scheduling period.
The related information of each order and the corresponding scheduling environment information are described by four attributes, which are respectively: actual processing time, order completion time, order processing time ratio and order completion ratio.
(2) And (3) giving a heuristic processing equipment distribution rule, and solving the optimal order processing sequence of the SMSP scheduling case in the step (1) by using a genetic algorithm.
The heuristic processing equipment distribution rule is obtained through manual setting, and all processing orders can be distributed to different equipment for processing by using the selected heuristic processing equipment distribution rule.
The termination conditions of the genetic algorithm are as follows: if the currently obtained optimal solution is not improved after the iteration for T times, the algorithm is terminated, and T is half of all orders.
(3) And constructing a training data set by using the SMSP scheduling case and the corresponding optimal order processing sequence, and training to obtain an order processing sequence prediction model.
The more the number of the SMSP scheduling cases for constructing the training data set is, the more reliable the order processing sequence prediction model obtained by training is.
The order processing sequence prediction model is expressed by adopting a single hidden layer feedforward neural network. And training an order processing sequence prediction model by adopting an extreme learning machine algorithm.
The input features of each training sample in the training data set include five types, which are respectively: the ratio of actual processing time, the ratio of completion time, the ratio of order processing time, the ratio of delivery date and the ratio of the number of orders scheduled in the current production scheduling stage to the total number of orders.
(4) Processing a batch processing order to be processed by using an order processing sequence prediction model to obtain an order processing arrangement sequence;
(5) and (3) distributing the batch processing orders to be processed to different equipment by using the heuristic processing equipment distribution rule in the step (2) and the order processing arrangement sequence in the step (4), determining the processing start time and the processing end time, and performing production processing.
The control method for single-stage multi-product batch processing based on data mainly has the following beneficial effects:
(1) the production scheduling scheme aiming at the single-stage multi-product batch processing production process can be rapidly formulated, and the real-time requirement of the actual production and processing process is met;
(2) in the on-line application stage, complex modeling and parameter setting work is not required, and the implementation is convenient;
(3) the method has better scheduling optimization capability.
Drawings
FIG. 1 is a flow chart of a method of controlling a data-based single-stage multi-product batch process of the present invention;
FIG. 2 is a flow chart of a scheduling scheme construction based on heuristic processing equipment allocation rules;
FIG. 3 is a block diagram of a genetic algorithm;
FIG. 4 is a schematic diagram of a model for predicting an order processing sequence;
FIG. 5 is a diagram of SMSP production scheduling scheme (time on abscissa) constructed based on heuristic processing equipment allocation rules;
fig. 6 is a schematic diagram of an optimal production scheduling scheme constructed based on a genetic algorithm (time is shown on the abscissa).
Detailed Description
The control method of the data-based single-stage multi-product batch processing of the present invention is described in detail below with reference to the accompanying drawings.
As shown in fig. 1, a control method of data-based single-stage multi-product batch processing includes the following steps:
(1) and aiming at different target batch processing scenes, corresponding historical production scheduling data is taken out from the historical production data, and SMSP scheduling cases are respectively established according to different production scheduling tasks in a production scheduling period.
(2) And (3) giving a heuristic processing equipment distribution rule, and solving the optimal order processing sequence of the SMSP scheduling case in the step (1) by using a genetic algorithm.
The heuristic machining equipment allocation rule is selected manually to allocate all orders to different equipment for production machining, for example, in a production scheduling period, assuming that N machining orders O ═ O1, O2, …, oN are in total, N machining orders are formed into an order machining sequence pi ═ (pi ═ N) in the production scheduling period1,π1,…,πN),πiE { o1, o2, …, oN }, i e {1,2, …, N }, and according to the order processing sequence, each of the N orders is allocated to a specific processing equipment for processingDetermining the starting time and the ending time of processing, and if an undistributed order still exists, continuing to distribute; if all orders are distributed, calculating the performance index f (pi) of the scheduling scheme to obtain a complete SMSP scheduling scheme, wherein the distribution process is shown in FIG. 2.
According to the related information of the orders and the corresponding scheduling environment information, all processing orders are sequenced through pairwise comparison of the orders, for example, any two orders are given, namely an order 1 and an order 2, a class attribute value is 0 or 1, a class attribute value is 1, and the fact that the order 1 is preferentially arranged for processing is indicated; the class attribute value is 0, and the order 2 should be preferentially arranged for processing.
For the kth order, the kth order can be preferentially arranged to be processed relative to the (k + 1) -n orders, and if the kth order is taken as 'order 1', the class attribute value is 1; if the kth order is taken as order 2, the value of the class attribute is 0.
The flow of the genetic algorithm is shown in fig. 3, and the chromosome is represented as pi ═ by (pi) using a permutation-based coding scheme1,π2,…,πN) In which piiE { o1, o2, …, oN }, i e {1,2, …, N }, wherein the decoding process is to process orders in the orders in sequence by using a given heuristic processing equipment distribution rule so as to determine a complete production scheduling scheme, and evaluate the performance of each scheduling scheme to obtain an adaptive value.
The main genetic manipulations used include:
selecting operation: selecting operation is finished by adopting a roulette selecting method;
and (3) cross operation: completing the cross operation by adopting a partial mapping cross method;
mutation operation: the mutation operation is carried out by adopting a reverse sequence method, namely, the fragments between two different random positions of the chromosome are reversed.
The genetic algorithm termination conditions are as follows: if the currently obtained optimal solution is not improved after the T iterations, the algorithm is terminated. T is half of the number of all processing orders.
(3) And constructing a training data set by using the SMSP scheduling case and the corresponding optimal order processing sequence, and training to obtain an order processing sequence prediction model.
For the k order, the k order can be preferentially arranged to be processed relative to the (k + 1) -n orders, therefore, the k order and any subsequent order are used for constructing a training sample, and the relevant information of the corresponding order and the real-time scheduling environment information are used for constructing the input characteristics.
The following four attributes are used to describe the relevant information and corresponding scheduling context information for each order:
i, actual processing time (Real _ Proc): in an order which is not scheduled to be processed, if an order i is selected to preferentially perform production processing, determining corresponding processing equipment according to a given heuristic processing equipment distribution rule so as to determine equipment switching time and processing time required by the processing of the order, wherein a calculation formula of actual processing time is as follows:
Real_Proci=PTi(mach)+CTji(1)
wherein, the mach represents the processing equipment distributed for the processing equipment according to the heuristic processing equipment distribution rule; j indicates the last order processed on the equipment. If the order i is the first order to be processed on the equipment, the switching time is 0.
II, Order completion time (Order _ Comp): if the order i is selected to preferentially carry out production processing in the order which is not scheduled to be processed, the calculation formula of the completion time of the order is as follows:
Order_Compi=Order_Compj+Real_Proci(2)
where j is the last order processed on the corresponding equipment. If order i is the firstOrder for processing on the equipment, Order _ CompjIs 0.
III, order processing time Ratio (Proc _ Ratio): the processing time required for processing the same order on different processing equipment varies. And at the current production scheduling moment, if the order i is arranged for processing, determining the corresponding processing equipment, the required switching time and the processing time according to the selected heuristic processing equipment distribution rule.
The comparison value of the processing time required for the production processing of the order i at the current time and the processing time required for the production processing of the order i at other times is calculated by adopting the following formula:
wherein,the average value of the processing time required by the order i to be processed on different equipment is represented;
representing the average of the switching times required to switch from different orders to order i.
Proc_RatioiThe smaller the processing time, the shorter the processing time required for scheduling the production processing for the order i in the current production scheduling stage, and vice versa, the longer the processing time required.
IV, Order completion ratio (Order _ Num _ R): the order completion ratio represents the ratio of the number of orders scheduled to be processed at the current production scheduling stage to the total number of orders, and this attribute is used to describe the real-time status of the current production processing environment.
In addition, the Delivery Date (DD) of the order may also be used as a static attribute to describe the order.
The definitions of the relevant symbols used in the four attributes i to iv are shown in table 1.
TABLE 1
Parameter(s) | Definition of |
m | Number of processing facilities |
n | Processing of orders |
PTi(mach) | Processing time of order i on processing equipment machh |
CTij | Switching time required for switching from processing of order i to processing of order j |
DDi | Delivery date of order i |
Ci | Completion time of order i |
The input features for each training sample in the training dataset include five, as listed in table 2:
TABLE 2
Feature(s) | Definition of |
Real_Proc_R | Ratio of actual processing time of order 1 and order 2 |
Order_Comp_R | Ratio of completion time of order 1 and order 2 |
Proc_Ratio_R | Ratio of order processing time ratio of order 1 and order 2 |
DD_R | Ratio of delivery dates of order 1 and order 2 |
Order_Num_R | Ratio of number of orders scheduled in current production scheduling stage to total number of orders |
As shown in fig. 4, a single-hidden layer feed forward neural network (SLFN) is used to represent the target scheduling pattern. The SLFN model has 5 input nodes which respectively correspond to 5 input characteristics of the training sample and 2 output nodes, wherein if the value of the node 1 is higher than that of the node 0, the order 1 is preferentially arranged for production and processing; conversely, "order 2" is prioritized for production processing. An extreme learning machine algorithm is used for the training of the SLFN model.
(4) Processing a batch processing order to be processed by using an order processing sequence prediction model to obtain an order processing arrangement sequence;
(5) and (3) distributing the batch processing orders to be processed to different equipment by using the heuristic processing equipment distribution rule in the step (2) and the order processing arrangement sequence in the step (4), determining the processing start time and the processing end time, and performing production processing.
The present invention is further described below with reference to a single-stage multi-product batch production scenario with 3 parallel processing production facilities.
The production scenario to minimize total post-pull completion timeAs a performance index for scheduling optimization, the total delay completion time calculation formula is as follows:
the method comprises the steps of firstly, taking out historical production scheduling data under a target batch production scene from historical production data, and respectively establishing SMSP scheduling cases aiming at production scheduling tasks in different production scheduling periods.
For example, 10 process orders O ═ O1, O2, …, O10 are shared in a historical production scheduling period, and the relevant production scheduling data is shown in table 3, and the production scheduling task in this production scheduling period is established as an SMSP scheduling case.
TABLE 3
And (3) distributing the order to different equipment for processing by using an earliest finish time rule (namely a heuristic processing equipment distribution rule), and determining the start time and the end time of order processing.
Taking the SMSP case shown in table 3 as an example, assuming that the scheduled order addition process column is pi ═ (7,6,8,10,4,3,9,1,5,2), the production scheduling scheme shown in fig. 5 can be constructed according to the earliest completion time rule.
And aiming at the SMSP scheduling cases constructed by the scheduling data of different production scheduling periods, optimizing and solving by utilizing a genetic algorithm to obtain an optimal order processing sequence.
The chromosomes of the genetic algorithm adopt an arrangement-based coding method, and are expressed as pi ═ pi (pi ═ pi)1,π2,…,π10) In which piiBelongs to the field of { o1, o2, …, o10}, and i belongs to the field of {1,2, …,10}, wherein the decoding process comprises the steps of utilizing a given earliest completion time rule to arrange and process orders corresponding to genes of the chromosome from left to right in sequence, so that a complete production scheduling scheme is determined, and corresponding total post-trawling completion time is obtained to evaluate the chromosome.
The population size of the genetic algorithm is set to 200, the initial population is randomly generated, the cross probability is set to 0.6, and the mutation probability is set to 0.1. The algorithm termination condition is as follows: after T consecutive iterations, if the optimal solution is still not improved, the algorithm is terminated, and T takes half of all orders.
Taking the SMSP scheduling case shown in table 3 as an example, an optimal order adding procedure obtained by using a genetic algorithm is listed as (1,9,10,2,8,3,7,6,5,4), the corresponding production scheduling scheme is shown in fig. 6, and the total post-pull completion time is 30.76.
And constructing a training data set according to related scheduling data in the historical production data and the optimal order processing sequence obtained by the genetic algorithm optimization solving unit so as to train an order processing sequence prediction model.
Taking the SMSP scheduling case shown in table 3 as an example, for the optimal order process sequence pi ═ (1,9,10,2,8,3,7,6,5,4), where the kth order (k ∈ {1,2, …,9}) can be prioritized for process production over the kth + 1-10 orders. Therefore, the order and any subsequent order can be used to construct a training sample. If the kth order is taken as 'order 1', the class attribute value of the training sample is 1; if the kth order is taken as the order 2, the class attribute value of the training sample is 0.
Taking orders "o 3" and "o 4" as examples, the construction steps of the input features of the training samples are explained:
in the current production scheduling phase, there are 5 orders (o1, o9, o10, o2, o8) scheduled for production processing. Both orders "o 3" and "o 4" will be scheduled to process plant u3 for processing according to the earliest completion time rule. Accordingly, attribute values for describing the relevant information and the real-time production scheduling environment information for each order are calculated:
(1) the actual machining time attributes of "o 3" and "o 4" are calculated using formula (1). The actual processing time for "o 3" was 18.93, and the actual processing time for "o 4" was 20.82.
(2) The order completion time attributes of "o 3" and "o 4" are calculated using equation (2). The order completion time of "o 3" was 29.63, and the order completion time of "o 4" was 31.52.
(3) The order processing time ratio attributes of "o 3" and "o 4" are calculated using formula (3). The order processing time ratio of "o 3" was 1.06, and the order processing time ratio of "o 4" was 1.10.
(4) Prior to orders "o 3" and "o 4", there are 5 orders scheduled for manufacturing, so the current order completion ratio is 0.5.
(5) The delivery date attribute of "o 3" is 31; the delivery date attribute of "o 4" is 38.
To predict which of the two orders should be prioritized for production processing, input features are further constructed based on the existing attributes to describe the comparison of the two. The input characteristics are as follows:
(1) the characteristic Real _ Proc _ R is the ratio of the actual processing times of the two orders.
(2) The characteristic Order _ Comp _ R is the ratio of the Order completion times of two orders.
(3) The Proc _ Ratio _ R is the Ratio of the order processing time ratios for the two orders.
(4) The characteristic DD _ R is the ratio of the delivery dates of the two orders.
(5) The characteristic Order _ Num _ R is the current completion ratio of the Order.
Two training samples of the final construct are shown in Table 4, class is a class attribute.
TABLE 4
No. | Real_Proc_R | Order_Comp_R | Proc_Ratio_R | DD_R | Order_Num_R | class |
1 | 0.91 | 0.94 | 0.96 | 0.82 | 0.5 | 1 |
2 | 1.10 | 1.06 | 1.04 | 1.23 | 0.5 | 0 |
Wherein training sample 1 has "o 3" as "order 1"; training sample 2 has "o 4" as "order 1".
The obtained order processing sequence prediction model can be regarded as an order sorting Rule aiming at the SMSP problem, and is compared with a genetic algorithm, an Earliest delivery Date Rule (EDD) and a Random order sorting Rule (RR) in order to verify the scheduling performance of the model.
For the SMSP problem of minimizing the total pull-out completion time, the earliest lead time rule can be used to generate the order scheduling sequence and has better scheduling effect. The SMSP case corresponding to 10 production scheduling periods was selected as the test case. Each SMSP case contains 10 orders for production processing, and the scheduling results of the four methods are shown in table 5.
TABLE 5
Test case | Genetic algorithm | SLFN rule | EDD rules | Random rule |
1 | 8.94 | 16.7 | 30.1 | 31.6 |
2 | 6.31 | 8.47 | 10.6 | 25.8 |
3 | 33.9 | 41.4 | 46.7 | 68.9 |
4 | 9.32 | 12.3 | 12.2 | 17.3 |
5 | 17.2 | 20 | 37.5 | 44.3 |
6 | 5.53 | 9.48 | 18.7 | 26 |
7 | 13.1 | 18.3 | 39.9 | 43.3 |
8 | 18.4 | 25.3 | 29 | 75.6 |
9 | 10 | 12.1 | 14.1 | 44.5 |
10 | 14 | 25.7 | 32 | 44.8 |
As can be seen from table 5, the SLFN rule can better reproduce the optimal scheduling capability of the genetic algorithm in all the test cases, and is significantly better than the EDD rule and the random rule.
The scheduling performance of the SLFN model is further verified using the following performance indicators:
wherein,representing the average performance index value of the genetic algorithm on the test case;
representing an average performance index value of an order sort rule (SLFN rule, EDD rule, random rule) on the test cases;
η is used to represent the optimized performance metric of such order ordering rules with respect to a complex optimization algorithm, the genetic algorithm.
the η calculation results of the three order sorting rules are shown in table 6, and the η value of the SLFN rule is much smaller than those of the EDD rule and the random rule, thereby further proving the scheduling performance of the constructed SLFN prediction model.
TABLE 6
(SLFN) | (EDD) | (RR) |
36.8 | 110.3 | 309.8 |
In 10 test cases, the calculation time required for constructing the order arrangement sequence by using three methods, namely the SLFN rule, the EDD rule and the random rule, is negligible, and the calculation time of the genetic algorithm is shown in table 7, where the first behavior in table 7 is the serial number of the test case and the second behavior is the calculation time(s).
TABLE 7
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |
1.2168 | 1.6848 | 1.7316 | 3.3072 | 1.1700 | 2.3088 | 1.2012 | 2.6208 | 1.6068 | 1.8408 |
The calculation time of the genetic algorithm is far longer than that of the SLFN rule, when the scale of the scheduling problem is large, the genetic algorithm often cannot meet the real-time requirement of the production and processing process due to too long solving time, and the constructed prediction model can obtain a better scheduling result only by extremely little calculation time and calculation amount, so that the method is more practical.
Claims (1)
1. A control method for data-based single-stage multi-product batch processing is characterized by comprising the following steps:
(1) aiming at different target batch processing scenes, corresponding historical production scheduling data are taken out from the historical production data, and SMSP scheduling cases are respectively established according to different production scheduling tasks in a production scheduling period;
(2) giving a heuristic processing equipment distribution rule, and solving the optimal order processing sequence of the SMSP scheduling case in the step (1) by using a genetic algorithm;
the termination conditions of the genetic algorithm are as follows: if the currently obtained optimal solution is not improved after continuous T iterations, the algorithm is terminated, and T is half of all orders;
(3) constructing a training data set by using the SMSP scheduling case and the corresponding optimal order processing sequence, training an order processing sequence prediction model by using an extreme learning machine algorithm, and training to obtain an order processing sequence prediction model;
the order processing sequence prediction model is expressed by adopting a single hidden layer feedforward neural network;
for the kth order, the kth order is preferentially arranged to be processed relative to the (k + 1) -n orders, so that the kth order and any subsequent order are used for constructing a training sample, and relevant information and real-time scheduling environment information of the corresponding order are used for constructing input characteristics;
the related information of each order and the corresponding scheduling environment information are described by four attributes, which are respectively: actual processing time, order completion time, order processing time ratio and order completion ratio;
the input features of each training sample in the training data set include five types, which are respectively: the ratio of actual processing time, the ratio of completion time, the ratio of order processing time, the ratio of delivery date and the ratio of the number of orders scheduled in the current production scheduling stage to the total number of orders;
(4) processing a batch processing order to be processed by using an order processing sequence prediction model to obtain an order processing arrangement sequence;
(5) and (3) distributing the batch processing orders to be processed to different equipment by using the heuristic processing equipment distribution rule in the step (2) and the order processing arrangement sequence in the step (4), determining the processing start time and the processing end time, and performing production processing.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510070318.1A CN104636871B (en) | 2015-02-10 | 2015-02-10 | A kind of control method of the single phase multi-product batch processing based on data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510070318.1A CN104636871B (en) | 2015-02-10 | 2015-02-10 | A kind of control method of the single phase multi-product batch processing based on data |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104636871A CN104636871A (en) | 2015-05-20 |
CN104636871B true CN104636871B (en) | 2018-02-23 |
Family
ID=53215590
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510070318.1A Expired - Fee Related CN104636871B (en) | 2015-02-10 | 2015-02-10 | A kind of control method of the single phase multi-product batch processing based on data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104636871B (en) |
Families Citing this family (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105321042B (en) * | 2015-10-19 | 2017-02-22 | 金航数码科技有限责任公司 | Genetic algorithm-based advanced plan scheduling system and method |
DE102016204392A1 (en) * | 2016-03-16 | 2017-09-21 | Trumpf Werkzeugmaschinen Gmbh + Co. Kg | System and process for production planning |
CN108122055B (en) * | 2016-11-28 | 2021-07-30 | 北京理工大学 | Resource scheduling method and device for flow shop |
CN108985617B (en) * | 2018-07-11 | 2021-07-13 | 广东人励智能工程有限公司 | Product production flow scheduling method and system based on intelligent manufacturing |
CN110059886B (en) * | 2019-04-25 | 2023-04-07 | 哈尔滨理工大学 | Comprehensive scheduling method considering simultaneous ending of single group of procedures of equipment batch processing |
CN112446520B (en) * | 2019-08-30 | 2024-06-21 | 京东科技控股股份有限公司 | Method and device for predicting product scheduling time |
CN112907068B (en) * | 2021-02-09 | 2022-12-02 | 刘连英 | Method for the batch production of fastener groups |
CN113031543B (en) * | 2021-02-24 | 2022-10-14 | 同济大学 | Control scheduling method and device for semiconductor production line |
CN116523151B (en) * | 2023-07-05 | 2024-01-09 | 深圳市鑫冠亚科技有限公司 | Electrode production management method, system and storage medium based on artificial intelligence |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101271543A (en) * | 2008-04-23 | 2008-09-24 | 永凯软件技术(上海)有限公司 | Production scheduling system and method using genetic algorithm based on elite solution pool |
EP2610696A1 (en) * | 2010-08-27 | 2013-07-03 | Hitachi, Ltd. | Process design/production planning device |
CN103309316A (en) * | 2013-05-28 | 2013-09-18 | 北京理工大学 | Scheduling method of multi-stage variation hybrid flow shop with batch processor |
-
2015
- 2015-02-10 CN CN201510070318.1A patent/CN104636871B/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101271543A (en) * | 2008-04-23 | 2008-09-24 | 永凯软件技术(上海)有限公司 | Production scheduling system and method using genetic algorithm based on elite solution pool |
EP2610696A1 (en) * | 2010-08-27 | 2013-07-03 | Hitachi, Ltd. | Process design/production planning device |
CN103309316A (en) * | 2013-05-28 | 2013-09-18 | 北京理工大学 | Scheduling method of multi-stage variation hybrid flow shop with batch processor |
Also Published As
Publication number | Publication date |
---|---|
CN104636871A (en) | 2015-05-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104636871B (en) | A kind of control method of the single phase multi-product batch processing based on data | |
Yang et al. | Optimized flowshop scheduling of multiple production lines for precast production | |
CN104268722B (en) | Dynamic flexible job-shop scheduling method based on multi-objective Evolutionary Algorithm | |
Guo et al. | Modeling and Pareto optimization of multi-objective order scheduling problems in production planning | |
Gomes et al. | Reactive scheduling in a make-to-order flexible job shop with re-entrant process and assembly: a mathematical programming approach | |
CN108460463B (en) | High-end equipment assembly line production scheduling method based on improved genetic algorithm | |
CN112882449B (en) | Multi-variety small-batch multi-target flexible job shop energy consumption optimization scheduling method | |
US7292903B2 (en) | Methods for determining tool assignment preference and manufacturing systems using the same | |
CN113379087A (en) | Production, manufacturing and scheduling optimization method based on improved genetic algorithm | |
CN113408771B (en) | Article delivery scheduling method, device and storage medium | |
CN112766659A (en) | Distributed hybrid flow shop scheduling method, medium and system | |
Yue et al. | Hybrid Pareto artificial bee colony algorithm for multi-objective single machine group scheduling problem with sequence-dependent setup times and learning effects | |
CN115933568A (en) | Multi-target distributed hybrid flow shop scheduling method | |
CN109214695B (en) | High-end equipment research, development and manufacturing cooperative scheduling method and system based on improved EDA | |
Chaudhry et al. | No-wait flowshop scheduling using genetic algorithm | |
CN117077975A (en) | Distributed heterogeneous flow shop scheduling method based on mixed initialization modular factor algorithm | |
Kim et al. | Scheduling 3D printers with multiple printing alternatives | |
CN113485278B (en) | Flexible job shop scheduling multi-target distribution estimation method for optimizing two production indexes | |
Lin et al. | Integrated optimization of supplier selection and service scheduling in cloud manufacturing environment | |
Kumar et al. | Integrated scheduling of part, tool and automated guided vehicles in a flexible manufacturing system using modified genetic algorithm | |
CN113112171B (en) | Batch scheduling method based on roulette and genetic algorithm | |
CN112514352A (en) | Method, device, system, storage medium and terminal for updating scheduling rule | |
CN114819558A (en) | Dual-target scheduling optimization method for distributed mixed flow shop | |
Bari et al. | Simulation of job sequencing for stochastic scheduling with a genetic algorithm | |
Caceres-Gelvez et al. | Evaluating the performance of a cellular manufacturing system proposal for the sewing department of a sportswear manufacturing company: A simulation approach |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180223 Termination date: 20210210 |