Distributed learning and control for manufacturing systems scheduling

Joonki Hong, Vittaldas V. Prabhu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

A new distributed learning and control (DLC) approach is presented in this paper which integrates part-driven distributed arrival time control (DATC) and machine-driven distributed reinforcement learning control. This approach is suitable for just-in-time (JIT) production for multi-objective scheduling problem in dynamically changing shop floor environment. While part controllers are adjusting their associated parts’ arrival time to minimize due-date deviation, machine controllers equipped with learning components are searching for optimal dispatching policies. The machines’ control problem is modeled as Semi Markov Decision Process (SMDP) and solved using one-step Q-learning. The DLC algorithms are evaluated for minimizing the sum of duedate deviation cost and setup cost using simulation. Results show that DLC algorithms achieve significant performance improvement over usual dispatching rules in complex real-time shop floor control problems for JIT production.

Original languageEnglish (US)
Title of host publicationEngineering of Intelligent Systems - 14th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, IEA/AIE 2001, Proceedings
EditorsLaszlo Monostori, Jozsef Vancza, Moonis Ali
PublisherSpringer Verlag
Pages582-591
Number of pages10
ISBN (Print)3540422196, 9783540422198
StatePublished - Jan 1 2001
Event14th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, IEA/AIE 2001 - Budapest, Hungary
Duration: Jun 4 2001Jun 7 2001

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume2070
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other14th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, IEA/AIE 2001
CountryHungary
CityBudapest
Period6/4/016/7/01

Fingerprint

Due Dates
Arrival Time
Scheduling
Distributed Algorithms
Control Algorithm
Just in time production
Learning Algorithm
Control Problem
Deviation
Semi-Markov Decision Process
Dispatching Rules
Controller
Learning Control
Q-learning
Setup Cost
Dispatching
Parallel algorithms
Learning algorithms
Reinforcement Learning
Scheduling Problem

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Hong, J., & Prabhu, V. V. (2001). Distributed learning and control for manufacturing systems scheduling. In L. Monostori, J. Vancza, & M. Ali (Eds.), Engineering of Intelligent Systems - 14th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, IEA/AIE 2001, Proceedings (pp. 582-591). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2070). Springer Verlag.
Hong, Joonki ; Prabhu, Vittaldas V. / Distributed learning and control for manufacturing systems scheduling. Engineering of Intelligent Systems - 14th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, IEA/AIE 2001, Proceedings. editor / Laszlo Monostori ; Jozsef Vancza ; Moonis Ali. Springer Verlag, 2001. pp. 582-591 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{cc989d7dbcc3417aae72ca521b2ea858,
title = "Distributed learning and control for manufacturing systems scheduling",
abstract = "A new distributed learning and control (DLC) approach is presented in this paper which integrates part-driven distributed arrival time control (DATC) and machine-driven distributed reinforcement learning control. This approach is suitable for just-in-time (JIT) production for multi-objective scheduling problem in dynamically changing shop floor environment. While part controllers are adjusting their associated parts’ arrival time to minimize due-date deviation, machine controllers equipped with learning components are searching for optimal dispatching policies. The machines’ control problem is modeled as Semi Markov Decision Process (SMDP) and solved using one-step Q-learning. The DLC algorithms are evaluated for minimizing the sum of duedate deviation cost and setup cost using simulation. Results show that DLC algorithms achieve significant performance improvement over usual dispatching rules in complex real-time shop floor control problems for JIT production.",
author = "Joonki Hong and Prabhu, {Vittaldas V.}",
year = "2001",
month = "1",
day = "1",
language = "English (US)",
isbn = "3540422196",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "582--591",
editor = "Laszlo Monostori and Jozsef Vancza and Moonis Ali",
booktitle = "Engineering of Intelligent Systems - 14th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, IEA/AIE 2001, Proceedings",
address = "Germany",

}

Hong, J & Prabhu, VV 2001, Distributed learning and control for manufacturing systems scheduling. in L Monostori, J Vancza & M Ali (eds), Engineering of Intelligent Systems - 14th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, IEA/AIE 2001, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 2070, Springer Verlag, pp. 582-591, 14th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, IEA/AIE 2001, Budapest, Hungary, 6/4/01.

Distributed learning and control for manufacturing systems scheduling. / Hong, Joonki; Prabhu, Vittaldas V.

Engineering of Intelligent Systems - 14th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, IEA/AIE 2001, Proceedings. ed. / Laszlo Monostori; Jozsef Vancza; Moonis Ali. Springer Verlag, 2001. p. 582-591 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2070).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Distributed learning and control for manufacturing systems scheduling

AU - Hong, Joonki

AU - Prabhu, Vittaldas V.

PY - 2001/1/1

Y1 - 2001/1/1

N2 - A new distributed learning and control (DLC) approach is presented in this paper which integrates part-driven distributed arrival time control (DATC) and machine-driven distributed reinforcement learning control. This approach is suitable for just-in-time (JIT) production for multi-objective scheduling problem in dynamically changing shop floor environment. While part controllers are adjusting their associated parts’ arrival time to minimize due-date deviation, machine controllers equipped with learning components are searching for optimal dispatching policies. The machines’ control problem is modeled as Semi Markov Decision Process (SMDP) and solved using one-step Q-learning. The DLC algorithms are evaluated for minimizing the sum of duedate deviation cost and setup cost using simulation. Results show that DLC algorithms achieve significant performance improvement over usual dispatching rules in complex real-time shop floor control problems for JIT production.

AB - A new distributed learning and control (DLC) approach is presented in this paper which integrates part-driven distributed arrival time control (DATC) and machine-driven distributed reinforcement learning control. This approach is suitable for just-in-time (JIT) production for multi-objective scheduling problem in dynamically changing shop floor environment. While part controllers are adjusting their associated parts’ arrival time to minimize due-date deviation, machine controllers equipped with learning components are searching for optimal dispatching policies. The machines’ control problem is modeled as Semi Markov Decision Process (SMDP) and solved using one-step Q-learning. The DLC algorithms are evaluated for minimizing the sum of duedate deviation cost and setup cost using simulation. Results show that DLC algorithms achieve significant performance improvement over usual dispatching rules in complex real-time shop floor control problems for JIT production.

UR - http://www.scopus.com/inward/record.url?scp=84947590389&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84947590389&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84947590389

SN - 3540422196

SN - 9783540422198

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 582

EP - 591

BT - Engineering of Intelligent Systems - 14th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, IEA/AIE 2001, Proceedings

A2 - Monostori, Laszlo

A2 - Vancza, Jozsef

A2 - Ali, Moonis

PB - Springer Verlag

ER -

Hong J, Prabhu VV. Distributed learning and control for manufacturing systems scheduling. In Monostori L, Vancza J, Ali M, editors, Engineering of Intelligent Systems - 14th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, IEA/AIE 2001, Proceedings. Springer Verlag. 2001. p. 582-591. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).