Learning from Prior Designs for Facility Layout Optimization

Hannu Rummukainen (Corresponding author), Jukka K. Nurminen, Timo Syrjänen, Jukka Pekka Numminen

Research output: Chapter in Book/Report/Conference proceedingChapter or book articleScientificpeer-review

Abstract

The problem of facility layout involves not only optimizing the locations of process components on a factory floor, but in real-world applications there are numerous practical constraints and objectives that can be difficult to formulate comprehensively in an explicit optimization model. As an alternative to explicit modelling, we present an optimization approach that learns structural properties from examples of expert-designed layouts of other similar facilities, and considers similarity to the examples as one objective in a multiobjective facility layout optimization problem. We have tested the approach on small-scale artificial test data, and the initial results indicate that a layout objective can be learned from example layouts, even if the process structure in the examples differs from the target case.
Original languageEnglish
Title of host publicationStudies in Computational Intelligence
EditorsFarouk Yalaoui, Lionel Amodeo, El-Ghazali Talbi
PublisherSpringer
Chapter6
Pages87-101
Number of pages15
ISBN (Electronic)978-3-030-58930-1
ISBN (Print)978-3-030-58929-5
DOIs
Publication statusPublished - 2021
MoE publication typeA3 Part of a book or another research book

Publication series

SeriesStudies in Computational Intelligence
Volume906
ISSN1860-949X

Keywords

  • optimisation
  • facility layout
  • machine learning

Fingerprint Dive into the research topics of 'Learning from Prior Designs for Facility Layout Optimization'. Together they form a unique fingerprint.

Cite this