Home > Research > Publications & Outputs > Metric learning for simulation analytics

Electronic data

  • WSCPaper2020_Final

    Accepted author manuscript, 452 KB, PDF document

    Available under license: CC BY-NC: Creative Commons Attribution-NonCommercial 4.0 International License

Links

View graph of relations

Metric learning for simulation analytics

Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSNConference contribution/Paperpeer-review

Published

Standard

Metric learning for simulation analytics. / Laidler, Graham; Morgan, Lucy; Nelson, Barry; Pavlidis, Nicos.

Proceedings of the 2020 Winter Simulation Conference. IEEE, 2020. p. 349-360.

Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSNConference contribution/Paperpeer-review

Harvard

Laidler, G, Morgan, L, Nelson, B & Pavlidis, N 2020, Metric learning for simulation analytics. in Proceedings of the 2020 Winter Simulation Conference. IEEE, pp. 349-360. <https://informs-sim.org/wsc20papers/029.pdf>

APA

Vancouver

Laidler G, Morgan L, Nelson B, Pavlidis N. Metric learning for simulation analytics. In Proceedings of the 2020 Winter Simulation Conference. IEEE. 2020. p. 349-360

Author

Laidler, Graham ; Morgan, Lucy ; Nelson, Barry ; Pavlidis, Nicos. / Metric learning for simulation analytics. Proceedings of the 2020 Winter Simulation Conference. IEEE, 2020. pp. 349-360

Bibtex

@inproceedings{58166e4cd8ff4617b48036d455df07ab,
title = "Metric learning for simulation analytics",
abstract = "The sample path generated by a stochastic simulation often exhibits significant variability within each replication, revealing periods of good and poor performance alike. As such, traditional summaries of aggregate performance measures overlook the more fine-grained insights into the operational system behavior. In this paper, we take a simulation analytics view of output analysis, turning to machine learning methods to uncover key insights from the dynamic sample path. We present a k nearest neighbors model on system state information to facilitate real-time predictions of a stochastic performance measure. This model is built on the premise of a system-specific measure of similarity between observations of the state, which we inform via metric learning. An evaluation of our approach is provided on a stochastic activity network and a wafer fabrication facility, both of which give us confidence in the ability of metric learning to provide interpretation and improved predictive performance.",
author = "Graham Laidler and Lucy Morgan and Barry Nelson and Nicos Pavlidis",
year = "2020",
month = dec,
day = "18",
language = "English",
pages = "349--360",
booktitle = "Proceedings of the 2020 Winter Simulation Conference",
publisher = "IEEE",

}

RIS

TY - GEN

T1 - Metric learning for simulation analytics

AU - Laidler, Graham

AU - Morgan, Lucy

AU - Nelson, Barry

AU - Pavlidis, Nicos

PY - 2020/12/18

Y1 - 2020/12/18

N2 - The sample path generated by a stochastic simulation often exhibits significant variability within each replication, revealing periods of good and poor performance alike. As such, traditional summaries of aggregate performance measures overlook the more fine-grained insights into the operational system behavior. In this paper, we take a simulation analytics view of output analysis, turning to machine learning methods to uncover key insights from the dynamic sample path. We present a k nearest neighbors model on system state information to facilitate real-time predictions of a stochastic performance measure. This model is built on the premise of a system-specific measure of similarity between observations of the state, which we inform via metric learning. An evaluation of our approach is provided on a stochastic activity network and a wafer fabrication facility, both of which give us confidence in the ability of metric learning to provide interpretation and improved predictive performance.

AB - The sample path generated by a stochastic simulation often exhibits significant variability within each replication, revealing periods of good and poor performance alike. As such, traditional summaries of aggregate performance measures overlook the more fine-grained insights into the operational system behavior. In this paper, we take a simulation analytics view of output analysis, turning to machine learning methods to uncover key insights from the dynamic sample path. We present a k nearest neighbors model on system state information to facilitate real-time predictions of a stochastic performance measure. This model is built on the premise of a system-specific measure of similarity between observations of the state, which we inform via metric learning. An evaluation of our approach is provided on a stochastic activity network and a wafer fabrication facility, both of which give us confidence in the ability of metric learning to provide interpretation and improved predictive performance.

M3 - Conference contribution/Paper

SP - 349

EP - 360

BT - Proceedings of the 2020 Winter Simulation Conference

PB - IEEE

ER -