Classifier-assisted optimization

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


Engineering design optimization often uses computer simulations to evaluate candidate designs. Such numerical simulations may consistently fail for some designs, but the failure reason being unknown. If such failures are frequent than the effectiveness of the optimization process can severely degrade. To address this issue this study describes the integration of classifiers, borrowed from the domain of machine learning, into the optimization search. The classifiers attempt to predict if a candidate design will cause a simulation crash, and this prediction is then used to bias the search. The effectiveness of the approach is demonstrated through several numerical experiments.

Original languageEnglish
Title of host publication9th Hellenic Conference on Artificial Intelligence, SETN 2016
EditorsAntonis Bikakis, Dimitrios Vrakas, Nick Bassiliades, Ioannis Vlahavas, George Vouros
ISBN (Electronic)9781450337342
StatePublished - 18 May 2016
Event9th Hellenic Conference on Artificial Intelligence, SETN 2016 - Thessaloniki, Greece
Duration: 18 May 201620 May 2016

Publication series

NameACM International Conference Proceeding Series


Conference9th Hellenic Conference on Artificial Intelligence, SETN 2016


  • Black-box optimization
  • Machine learning
  • Metamodels


Dive into the research topics of 'Classifier-assisted optimization'. Together they form a unique fingerprint.

Cite this