Meta-Regression: A Framework for Robust Reactive Optimization

Dan McClary, Violet R. Syrotiuk, Murat Kulahci

    Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

    Abstract

    Maintaining optimal performance as the conditions of a system change is a challenging problem. To solve this problem, we present meta-regression, a general methodology for alleviating traditional difficulties in nonlinear regression modelling. Meta-regression allows for reactive optimization, in which system components self-organize to changing conditions in a manner that is robust, or affected minimally by other sources of variability. Meta-regression extends profiling, providing a methodology for model-building when there is incomplete knowledge of the mechanisms and interactions of a nonlinear system.
    Original languageEnglish
    Title of host publicationFirst IEEE International Conference on Self-Adaptive and Self-Organizing Systems
    PublisherIEEE Computer Society Press
    Publication date2007
    Pages375-378
    ISBN (Print)978-0-7695-2906-6
    DOIs
    Publication statusPublished - 2007
    EventFirst IEEE International Conference on Self-Adaptive and Self-Organizing Systems (SASO 2007) - Boston, MA, United States
    Duration: 9 Jul 200711 Jul 2007
    Conference number: 1

    Conference

    ConferenceFirst IEEE International Conference on Self-Adaptive and Self-Organizing Systems (SASO 2007)
    Number1
    CountryUnited States
    CityBoston, MA
    Period09/07/200711/07/2007

    Fingerprint Dive into the research topics of 'Meta-Regression: A Framework for Robust Reactive Optimization'. Together they form a unique fingerprint.

    Cite this