CN116700173A - Dynamic scheduling method of production line based on graph representation learning - Google Patents
Dynamic scheduling method of production line based on graph representation learning Download PDFInfo
- Publication number
- CN116700173A CN116700173A CN202310704917.9A CN202310704917A CN116700173A CN 116700173 A CN116700173 A CN 116700173A CN 202310704917 A CN202310704917 A CN 202310704917A CN 116700173 A CN116700173 A CN 116700173A
- Authority
- CN
- China
- Prior art keywords
- order
- production line
- scheduling
- task
- state
- 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.)
- Pending
Links
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 56
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000010586 diagram Methods 0.000 claims abstract description 17
- 230000009471 action Effects 0.000 claims abstract description 7
- 239000013598 vector Substances 0.000 claims description 24
- 230000008569 process Effects 0.000 claims description 10
- 230000004044 response Effects 0.000 claims description 9
- 238000009826 distribution Methods 0.000 claims description 6
- 230000008901 benefit Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 6
- 238000012545 processing Methods 0.000 description 5
- 230000006872 improvement Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 229910000831 Steel Inorganic materials 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000010485 coping Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000035484 reaction time Effects 0.000 description 1
- 239000010959 steel Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM]
- G05B19/41865—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM] characterised by job scheduling, process planning, material flow
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/30—Nc systems
- G05B2219/32—Operator till task planning
- G05B2219/32252—Scheduling production, machining, job shop
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Manufacturing & Machinery (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a dynamic scheduling method of a production line based on graph representation learning, which constructs a dynamic scheduling model of the production line according to the quantity of resource types of the production line, the resource requirement of tasks corresponding to orders, the duration time of the tasks and the arrival time; extracting hidden features required by the production line scheduling through state coding and feature embedding representation to form corresponding state diagram representation; forming an initial solution, namely an initial state of scheduling according to a first-in first-out rule; selecting an area needing rescheduling each time in the state diagram through an area selector module; selecting proper exchange action to reschedule the region output by the region selector through the rule selector module; repeatedly inputting the scheduling solutions into the scheduling selector and rule selector module in sequence, and obtaining a final solution after the termination condition is met; the invention can obtain a production line dispatching scheme with excellent effect.
Description
Technical Field
The invention belongs to the technical field of production line scheduling, and particularly relates to a dynamic scheduling method of a production line based on graph representation learning.
Background
The problem of dynamic scheduling of a production line is common in the field of production and manufacturing, for example, in the production process of various industries such as automobile, electronics, steel, household appliances, aerospace product manufacturing and the like, reasonable scheduling is required under the conditions of conflicting production targets and limited resources. In most of the previous studies, these problems were solved by means of well-designed heuristics, and a close reading of the related studies in this field found that a typical procedure for solving the problem using heuristics is: a smart heuristic method is provided for the simplified model of the problem, and then the test and adjustment are performed with great effort, so that the method can obtain good performance in practice. Once certain aspects of the problem (e.g., workload or other metric values) change, the above procedure must generally be repeated to cope with these changes. The prior art solves the problem of dynamic scheduling of a production line by adopting a reactive priority rule or a meta heuristic algorithm, and has the defects of insufficient intelligence (incapability of intelligently coping with different working conditions) or the need of a large amount of iterative optimization (efficiency is reduced so as not to meet the requirement of dynamic scheduling on reaction time) and the like.
Disclosure of Invention
In view of the above, the invention provides a dynamic scheduling method of a production line based on graph representation learning, which can obtain a production line scheduling scheme with excellent effect.
The technical scheme for realizing the invention is as follows:
a dynamic scheduling method of a production line based on graph representation learning constructs a dynamic scheduling model of the production line according to the quantity of resource types of the production line, the resource requirements of tasks corresponding to orders, the duration time of the tasks and the arrival time; extracting hidden features required by the production line scheduling through state coding and feature embedding representation to form corresponding state diagram representation; forming an initial solution, namely an initial state of scheduling according to a first-in first-out rule; selecting an area needing rescheduling each time in the state diagram through an area selector module; selecting proper exchange action to reschedule the region output by the region selector through the rule selector module; and repeatedly inputting the scheduling solutions into the scheduling selector and rule selector modules in sequence, and obtaining a final solution after the termination condition is met.
Further, the dynamic scheduling model of the production line is specifically:
setting D resource types in the production line, wherein m orders in the production line arrive in an online mode in discrete time steps, and the task corresponding to any order j has p j ＝(p j,1 ,……,p j,D ) Resource requirement (p is 0.ltoreq.p) j,n Less than or equal to 1, n=1, …, D) and the arrival time is A j The duration of the task is T j The method comprises the steps of carrying out a first treatment on the surface of the And makes the following assumptions:
1) During the whole production scheduling process, the resource requirement of each order corresponding to the task is fixedly known when arriving;
2) Each order corresponding task must run continuously until completion, and preemption is not allowed;
for each order j, define a task start time B j The task end time is C j Task response ratio H j ＝C j -A j /T j 。
Further, order node v j Corresponding order attribute p j ,A j ,T j ,B j Is embedded in (D× (T) max +1) dimension vector e j Wherein T is max Is the task maximum duration; the vector encodes task attribute and resource occupation state in the whole dynamic dispatching process of the production line, and the front D-dimension vector is p j ＝(p j,1 ,p j,2 ,……,p j,D )，D×T j The dimension vector describes the amount of resources that order j occupies for all order tasks during their task execution, inExpressed in terms of the amount of resources occupied by all tasks at each time step t, expressed as p' t ＝(p′ t,1 ,p′ t,2 ,……,p′ t,D ) There is usually T j <T max Remaining D× (T max -T j ) The dimension vectors are all 0, vector e j The last dimension of (1) represents the task response ratio H corresponding to the order j placed by the current schedule j ；
Calculate each order node v j Implicit feature h of (2) j For each order node v j C (j) is the set of all its child nodes, (h) 1 ,c 1 ),(h 2 ,c 2 ),……,(h k ,c k ) For its child node's LSTM state, its LSTM state is:
wherein h is j For order node v j Implicit features of c j For order node v j Cell state of h k Is child node v k Implicit features of c k Is child node v k Is a cell state of (2);
finally, a scheduling state diagram and each order node v in the diagram are obtained j Corresponding embedded vector e j And implicit feature h j 。
Further, given state s t ，s t Including the current arbitrary order node v j Corresponding embedded vector e j And implicit feature h j The region selector is for each region ω t ∈Ω(s t ) Calculate a score Q(s) t ,ω t )，Q(s t ,ω t ) Measuring the current state s t Lower pair of regions omega t The benefit of rescheduling, a high score indicates that for s t [ω t ]It is desirable to reschedule, Ω (s t ) Is a set of regions associated with dynamic scheduling of a production line that covers all order nodes that can be used for rescheduling and is based on a score Q (s t ,ω t ) Outputting a probability distribution pi ω (ω t ∣s t )
Selecting an area omega for rescheduling t0 Outputting to a rule selector; region omega t Covered by order node v j0 The front and back W order nodes are centered (in the order of arrival), i.e. each order node can exchange positions with the front and back W order nodes at most in each rescheduling process.
Further, given a state and region s t [ω t0 ]Requiring rescheduling, the rule selector predicts a probability distribution pi over the entire rule set u u (u t ∣s t [ω t0 ]) And selects a rule u t0 Application of E u to s t [ω t ]The method comprises the steps of carrying out a first treatment on the surface of the Rescheduling action corresponds to the current order node v j0 Moving to region omega in the scheduling state diagram t0 Another order node v j0′ Or v 0 I.e. after completion of order j0' or at arrival time a of order j0 j0 Immediately assigning the order j0; a new scheduling solution (scheduling state) s can be obtained t+1 。
The beneficial effects are that:
1. the invention provides a novel dynamic scheduling method of a production line based on graph representation learning, which has excellent effect of scheduling the production line. The method adopts a directed acyclic graph to carry out state coding and feature embedding representation, and obtains deeper features of the state of the production line; on the basis, the method carries out iterative improvement on the current scheduling solution through two components of the region selector and the rule selector, continuously improves the quality of the scheduling solution, and finally can obtain the suboptimal solution of the production line scheduling.
2. Experiments prove that the scheduling effect of the method is superior to various heuristic scheduling rules and a general operation planning optimization tool Google OR-tools, and the effectiveness of the method is proved.
Drawings
FIG. 1 is a diagram illustrating an example of a scheduling scheme and its counterparts;
FIG. 2 is an embedded representation of an order task;
FIG. 3 is an example of a representation of a region selector module;
fig. 4 shows the comparative experimental results of the present method and various scheduling methods.
Detailed Description
The invention will now be described in detail by way of example with reference to the accompanying drawings.
The invention provides a production line dynamic scheduling method based on graph representation learning, which comprises the following specific embodiments:
1. construction of dynamic scheduling model of production line
Considering a production line with D resource types, m orders of the production line arrive in an online manner in discrete time steps, and the task corresponding to any order j has p j ＝(p j,1 ,……,p j,D ) Resource requirement (p is 0.ltoreq.p) j,n Less than or equal to 1, n=1, …, D) and the arrival time is A j The duration of the task is T j . And makes the following assumptions:
1) During the whole production scheduling process, the resource requirement of each order corresponding to the task is fixedly known when arriving;
2) Each order correspondence task must continue running until completion, not allowing preemption.
For each order j, define a task start time B j The task end time is C j Task response ratio H j ＝C j -A j /T j . The invention determines a scheduling schedule for all orders by a production line dynamic scheduling method based on graph representation learning, so that the average task response ratio H is minimum, and the average task response ratio is defined
2. State encoding and feature embedding representation
As shown in FIG. 1, the scheduling state is represented at each time step as a directed acyclic graph that describes the dependency between different order scheduling times (orders). Specifically, each order j is represented as an order node v in the directed acyclic graph j At the same time add a v 0 The node acts as the starting node if order j is at its arrival time A j Is scheduled immediately (i.e. B j ＝A j ) Then add a directed edge to the graph<v 0 ,v j >Otherwise, there must be at least one order j' such that C j′ ＝B j (i.e., order j begins immediately after order j' ends), adding an edge to such order<v j′ ,v j >Into the figure.
As shown in FIG. 2, order node v j Corresponding order attribute p j ,A j ,T j ,B j Is embedded in (D× (T) max +1) dimension vector e j Wherein T is max Is the task maximum duration. The vector encodes task attribute and resource occupation state in the whole dynamic dispatching process of the production line, and the front D-dimension vector is p j ＝(p j,1 ,p j,2 ,……,p j,D )，D×T j The dimension vector describes the amount of resources that order j occupies for all order tasks during their task execution, inExpressed in terms of the amount of resources occupied by all tasks at each time step t, expressed as p' t ＝(p′ t,1 ,p′ t,2 ,……,p′ t,D ) There is usually T j <T max Remaining D× (T max -T j ) The dimension vectors are all 0, vector e j The last dimension of (1) represents the task response ratio H corresponding to the order j placed by the current schedule j 。
Each order node v is then calculated j Implicit feature h of (2) j For each order node v j C (j) is the set of all its child nodes, (h) 1 ,c 1 ),(h 2 ,c 2 ),……,(h k ,c k ) For its child node's LSTM state, its LSTM state is:
wherein h is j For order node v j Implicit features of c j For order node v j Cell state of h k Is child node v k Implicit features of c k Is child node v k Is a cell state of (a) a cell.
Finally, a scheduling state diagram and each order node v in the diagram are obtained j Corresponding embedded vector e j And implicit feature h j 。
3. An initial solution, i.e., the initial state of the schedule, is formed according to a first-in first-out rule.
Scheduling according to first-in first-out heuristic scheduling rule, i.e. rule that order arrived first is scheduled first to obtain initial scheduling state s 0 。
4. And the region selector module is used for selecting the region which needs to be rescheduled each time in the state diagram.
Given state s t ，s t Including the current arbitrary order node v j Corresponding embedded vector e j And implicit feature h j As shown in fig. 3, the region selector is for each region ω t ∈Ω(s t ) Calculate a score Q(s) t ,ω t ) It measures the current state s t Lower pair of regions omega t The benefit of rescheduling, a high score indicates that for s t [ω t ]It is desirable to reschedule, Ω (s t ) Is a set of regions associated with dynamic scheduling of a production line that covers all order nodes that can be used for rescheduling and is based on a score Q (s t ,ω t ) Outputting a probability distribution pi ω (ω t ∣s t )
And selecting an area omega for rescheduling t0 And outputting to a rule selector. Region omega t Covered by order node v j0 The front and back W order nodes are centered (in the order of arrival), i.e. each order node can exchange positions with the front and back W order nodes at most in each rescheduling process.
5. And the rule selector module selects proper switching actions and reschedules the area output by the area selector.
Given state and region s t [ω t0 ]Requiring rescheduling, the rule selector predicts a probability distribution pi over the entire rule set u u (u t ∣s t [ω t0 ]) And selects a rule u t0 Application of E u to s t [ω t ]. Rescheduling action corresponds to the current order node v j0 Moving to region omega in the scheduling state diagram t0 Another order node v j0′ Or v 0 I.e. after completion of order j0' or at arrival time a of order j0 j0 The order j0 is immediately assigned. A new scheduling solution (scheduling state) s can be obtained t+1 。
6. Repeatedly inputting the scheduling solutions into the region selector and rule selector modules in turn to obtain a rescheduling sequence
(s 0 ,(ω 0 ,u 0 )),(s 1 ,(ω 1 ,u 1 )),…,(s T-1 ,(ω T-1 ,u T-1 )),s T (4) When Q(s) t ,ω t )<E or rule u t When the solution is unavailable, the rescheduling process is immediately terminated, and the E=0 is taken, so that the final solution s of the scheduling can be obtained T 。
The following implementation effects can be obtained by using the method:
the dynamic scheduling method of the production line based on graph representation learning provided by the invention is compared with various heuristic scheduling rules and a general operational preparation optimizing tool Google OR-tools. The heuristic scheduling rules employed are as follows:
(1) First-in first-out FIFO (First In First Out): the first order is scheduled;
(2) Shortest task duration priority SPTF (Shortest Processing Time First): the order with the shortest task duration is scheduled preferentially;
(3) Longest task duration priority LPTF (Longest Processing Time First): the order with the longest task duration is preferentially scheduled;
(4) Minimum resource occupation priority LROF (Least Resource Occupation First): priority processing the order with least occupation of resources;
(5) Maximum resource occupancy priority MROF (Most Resource Occupation First): priority processing orders with most occupied resources;
(6) Minimum remaining latency priority LWTF (Least Waiting Time First): preferentially processing orders with minimum waiting time of the residual tasks;
(7) Random scheduling ramcom: order priority is randomly assigned.
The experimental result is shown in fig. 4, and the result shows that the average task response ratio H obtained by the dynamic scheduling method (shown by GRL, graph Representation Learning in the figure) of the production line based on graph representation learning is minimum, the scheduling effect is better than that of other methods, and the effectiveness of the method is proved.
In summary, the above embodiments are only preferred embodiments of the present invention, and are not intended to limit the scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (5)
1. A dynamic scheduling method of a production line based on graph representation learning is characterized in that a dynamic scheduling model of the production line is built according to the quantity of resource types of the production line, the resource requirement of a task corresponding to an order, the duration time of the task and the arrival time; extracting hidden features required by the production line scheduling through state coding and feature embedding representation to form corresponding state diagram representation; forming an initial solution, namely an initial state of scheduling according to a first-in first-out rule; selecting an area needing rescheduling each time in the state diagram through an area selector module; selecting proper exchange action to reschedule the region output by the region selector through the rule selector module; and repeatedly inputting the scheduling solutions into the scheduling selector and rule selector modules in sequence, and obtaining a final solution after the termination condition is met.
2. The dynamic scheduling method of a production line based on graph representation learning as claimed in claim 1, wherein the dynamic scheduling model of the production line specifically comprises:
setting D resource types in the production line, wherein m orders in the production line arrive in an online mode in discrete time steps, and the task corresponding to any order j has p j ＝(p j,1 ,……,p j,D ) Resource requirement (p is 0.ltoreq.p) j,n Less than or equal to 1, n=1, …, D) and the arrival time is A j The duration of the task is T j The method comprises the steps of carrying out a first treatment on the surface of the And makes the following assumptions:
1) During the whole production scheduling process, the resource requirement of each order corresponding to the task is fixedly known when arriving;
2) Each order corresponding task must run continuously until completion, and preemption is not allowed;
for each order j, define a task start time B j The task end time is C j Task response ratio H j ＝C j -A j /T j 。
3. A dynamic scheduling method of a production line based on graph representation learning as claimed in claim 2, wherein the order node v is j Corresponding order attribute p j ,A j ,T j ,B j Is embedded in (D× (T) max +1) dimension vector e j Wherein T is max Is the task maximum duration; the vector encodes task attribute and resource occupation state in the whole dynamic dispatching process of the production line, and the front D-dimension vector is p j ＝(p j,1 ,p j,2 ,……,p j,D )，D×T j The dimension vector describes the resources that order j occupies for all order tasks during its task executionThe amount is as followsExpressed in terms of the amount of resources occupied by all tasks at each time step t, expressed as p' t ＝(p′ t,1 ,p′ t,2 ,……,p′ t,D ) There is usually T j <T max Remaining D× (T max -T j ) The dimension vectors are all 0, vector e j The last dimension of (1) represents the task response ratio H corresponding to the order j placed by the current schedule j ；
Calculate each order node v j Implicit feature h of (2) j For each order node v j C (j) is the set of all its child nodes, (h) 1 ,c 1 ),(h 2 ,c 2 ),……,(h k ,c k ) For its child node's LSTM state, its LSTM state is:
wherein h is j For order node v j Implicit features of c j For order node v j Cell state of h k Is child node v k Implicit features of c k Is child node v k Is a cell state of (2);
finally, a scheduling state diagram and each order node v in the diagram are obtained j Corresponding embedded vector e j And implicit feature h j 。
4. A method for dynamic scheduling of a production line based on graph representation learning as claimed in claim 3, wherein the given state s t ，s t Including the current arbitrary order node v j Corresponding embedded vector e j And implicit feature h j The region selector is for each region ω t ∈Ω(s t ) Calculate a score Q(s) t ,ω t )，Q(s t ,ω t ) Measuring the current state s t Lower pair of regions omega t The benefit of rescheduling, a high score indicates that for s t [ω t ]It is desirable to reschedule, Ω (s t ) Is a set of regions associated with dynamic scheduling of a production line that covers all order nodes that can be used for rescheduling and is based on a score Q (s t ,ω t ) Outputting a probability distribution pi ω (ω t ∣s t )
Selecting an area omega for rescheduling t0 Outputting to a rule selector; region omega t Covered by order node v j0 The front and back W order nodes are centered, namely each order node can exchange positions with the front and back W order nodes at most in each rescheduling process.
5. The dynamic scheduling method of a production line based on graph representation learning as claimed in claim 4, wherein the given state and region s t [ω t0 ]Requiring rescheduling, rule selector is across rule setUp-prediction of a probability distribution pi u (u t ∣s t [ω t0 ]) And selects a rule +.>Application to s t [ω t ]The method comprises the steps of carrying out a first treatment on the surface of the Rescheduling action corresponds to the current order node v j0 Moving to region omega in the scheduling state diagram t0 Another order node v j0′ Or v 0 I.e. after completion of order j0' or at arrival time a of order j0 j0 Immediately assigning the order j0; a new scheduling solution s can be obtained t+1 。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310704917.9A CN116700173A (en) | 2023-06-14 | 2023-06-14 | Dynamic scheduling method of production line based on graph representation learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310704917.9A CN116700173A (en) | 2023-06-14 | 2023-06-14 | Dynamic scheduling method of production line based on graph representation learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116700173A true CN116700173A (en) | 2023-09-05 |
Family
ID=87840742
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310704917.9A Pending CN116700173A (en) | 2023-06-14 | 2023-06-14 | Dynamic scheduling method of production line based on graph representation learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116700173A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117455194A (en) * | 2023-11-27 | 2024-01-26 | 无锡雪浪数制科技有限公司 | Discrete event simulation-based production scheduling method |
-
2023
- 2023-06-14 CN CN202310704917.9A patent/CN116700173A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117455194A (en) * | 2023-11-27 | 2024-01-26 | 无锡雪浪数制科技有限公司 | Discrete event simulation-based production scheduling method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109561148B (en) | Distributed task scheduling method based on directed acyclic graph in edge computing network | |
| US6278978B1 (en) | Agent scheduling system and method having improved post-processing step | |
| Tavakkoli-Moghaddam et al. | Solving a multi-objective job shop scheduling problem with sequence-dependent setup times by a Pareto archive PSO combined with genetic operators and VNS | |
| CN110458326B (en) | Mixed group intelligent optimization method for distributed blocking type pipeline scheduling | |
| CN109636011A (en) | A kind of multishift operation plan scheduling method based on improved change neighborhood genetic algorithm | |
| CN109582448A (en) | A kind of edge calculations method for scheduling task towards criticality and timeliness | |
| Zhao et al. | An improved Q-learning based rescheduling method for flexible job-shops with machine failures | |
| CN116700173A (en) | Dynamic scheduling method of production line based on graph representation learning | |
| Xiao et al. | A cooperative coevolution hyper-heuristic framework for workflow scheduling problem | |
| CN110187956B (en) | Layered real-time task scheduling method and system for multi-agent platform | |
| CN109445386A (en) | A kind of most short production time dispatching method of the cloud manufacturing operation based on ONBA | |
| CN115310794A (en) | Man-machine cooperation assembly line balancing method and device | |
| CN109871270B (en) | Scheduling scheme generation method and device | |
| CN107958313A (en) | A kind of discrete ripples optimization algorithm | |
| CN110264097A (en) | More scientific workflows based on cloud environment concurrently execute dispatching method | |
| CN104698838B (en) | Based on the fuzzy scheduling rule digging method that domain dynamic is divided and learnt | |
| CN110989538B (en) | Closed-loop scheduling optimization method for complex production process | |
| CN115034615A (en) | Method for improving feature selection efficiency in genetic programming scheduling rule for job shop scheduling | |
| Wei et al. | Composite rules selection using reinforcement learning for dynamic job-shop scheduling | |
| JPH1153006A (en) | Scheduling method | |
| Ramírez-Hernández et al. | A case study in scheduling reentrant manufacturing lines: Optimal and simulation-based approaches | |
| Krömer et al. | Scheduling independent tasks on heterogeneous distributed environments by differential evolution | |
| Mangukia et al. | Improved Variable Round Robin Scheduling Algorithm | |
| Józefowska et al. | Solving discrete-continuous scheduling problems by Tabu Search | |
| CN113064710B (en) | Cloud workflow scheduling method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |