Dispatching rule selection with Gaussian processes

Research output: Journal contributionsJournal articlesResearchpeer-review

Standard

Dispatching rule selection with Gaussian processes. / Heger, Jens; Hildebrandt, Torsten; Scholz-Reiter, Bernd.
In: Central European Journal of Operations Research, Vol. 23, No. 1, 03.2015, p. 235-249.

Research output: Journal contributionsJournal articlesResearchpeer-review

Harvard

APA

Vancouver

Heger J, Hildebrandt T, Scholz-Reiter B. Dispatching rule selection with Gaussian processes. Central European Journal of Operations Research. 2015 Mar;23(1):235-249. doi: 10.1007/s10100-013-0322-7

Bibtex

@article{016051ed68b2451ba5ca60a8cfb7f077,
title = "Dispatching rule selection with Gaussian processes",
abstract = "Decentralized scheduling with dispatching rules is applied in many fields of logistics and production, especially in highly complex and dynamic scenarios, such as semiconductor manufacturing. Many dispatching rules are proposed in the literature, which perform well on specific scenarios. No rule is known, however, consistently outperforming all other rules. One approach to meet this challenge and improve scheduling performance is to select and switch dispatching rules depending on current system conditions. For this task machine learning techniques (e.g., Artificial Neural Networks) are frequently used. In this paper we investigate the use of a machine learning technique not applied to this task before: Gaussian process regression. Our analysis shows that Gaussian processes predict dispatching rule performance better than Neural Networks in most settings. Additionally, already a single Gaussian Process model can easily provide a measure of prediction quality. This is in contrast to many other machine learning techniques. We show how to use this measure to dynamically add additional training data and incrementally improve the model where necessary. Results therefore suggest, Gaussian processes are a very promising technique, which can lead to better scheduling performance (e.g., reduced mean tardiness) compared to other techniques.",
keywords = "Dispatching rules, Gaussian processes, Machine learning, Planning and scheduling, Production management and logistics, Engineering",
author = "Jens Heger and Torsten Hildebrandt and Bernd Scholz-Reiter",
year = "2015",
month = mar,
doi = "10.1007/s10100-013-0322-7",
language = "English",
volume = "23",
pages = "235--249",
journal = "Central European Journal of Operations Research",
issn = "1435-246X",
publisher = "Springer",
number = "1",

}

RIS

TY - JOUR

T1 - Dispatching rule selection with Gaussian processes

AU - Heger, Jens

AU - Hildebrandt, Torsten

AU - Scholz-Reiter, Bernd

PY - 2015/3

Y1 - 2015/3

N2 - Decentralized scheduling with dispatching rules is applied in many fields of logistics and production, especially in highly complex and dynamic scenarios, such as semiconductor manufacturing. Many dispatching rules are proposed in the literature, which perform well on specific scenarios. No rule is known, however, consistently outperforming all other rules. One approach to meet this challenge and improve scheduling performance is to select and switch dispatching rules depending on current system conditions. For this task machine learning techniques (e.g., Artificial Neural Networks) are frequently used. In this paper we investigate the use of a machine learning technique not applied to this task before: Gaussian process regression. Our analysis shows that Gaussian processes predict dispatching rule performance better than Neural Networks in most settings. Additionally, already a single Gaussian Process model can easily provide a measure of prediction quality. This is in contrast to many other machine learning techniques. We show how to use this measure to dynamically add additional training data and incrementally improve the model where necessary. Results therefore suggest, Gaussian processes are a very promising technique, which can lead to better scheduling performance (e.g., reduced mean tardiness) compared to other techniques.

AB - Decentralized scheduling with dispatching rules is applied in many fields of logistics and production, especially in highly complex and dynamic scenarios, such as semiconductor manufacturing. Many dispatching rules are proposed in the literature, which perform well on specific scenarios. No rule is known, however, consistently outperforming all other rules. One approach to meet this challenge and improve scheduling performance is to select and switch dispatching rules depending on current system conditions. For this task machine learning techniques (e.g., Artificial Neural Networks) are frequently used. In this paper we investigate the use of a machine learning technique not applied to this task before: Gaussian process regression. Our analysis shows that Gaussian processes predict dispatching rule performance better than Neural Networks in most settings. Additionally, already a single Gaussian Process model can easily provide a measure of prediction quality. This is in contrast to many other machine learning techniques. We show how to use this measure to dynamically add additional training data and incrementally improve the model where necessary. Results therefore suggest, Gaussian processes are a very promising technique, which can lead to better scheduling performance (e.g., reduced mean tardiness) compared to other techniques.

KW - Dispatching rules

KW - Gaussian processes

KW - Machine learning

KW - Planning and scheduling

KW - Production management and logistics

KW - Engineering

UR - http://www.scopus.com/inward/record.url?scp=84881500606&partnerID=8YFLogxK

U2 - 10.1007/s10100-013-0322-7

DO - 10.1007/s10100-013-0322-7

M3 - Journal articles

AN - SCOPUS:84881500606

VL - 23

SP - 235

EP - 249

JO - Central European Journal of Operations Research

JF - Central European Journal of Operations Research

SN - 1435-246X

IS - 1

ER -