Distributed simulation execution on a high-performance cluster using heuristiclab hive

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

Abstract

This paper presents a generic way of distributing simulation executions. The proposed method can be applied on high performance clusters in order to execute many simulation runs simultaneously and gather respective results. These results can then be collected into datasets, which one can use for data analysis and examination of different properties of the implemented simulation model and the underlying simulated process, e.g. production plants or logistics systems. The approach is showcased using a concrete simulation of a production process controlled by different disposition parameters. After executing 30 000 simulation runs, we are able to create respective datasets in order to further analyse the properties of the production process and build surrogate models of the simulation model, which in turn can be used in surrogate-assisted parameter optimization.

Original languageEnglish
Title of host publication30th European Modeling and Simulation Symposium, EMSS 2018
EditorsYuri Merkuryev, Miquel Angel Piera, Francesco Longo, Agostino G. Bruzzone, Michael Affenzeller, Emilio Jimenez
PublisherDIME UNIVERSITY OF GENOA
Pages107-114
Number of pages8
ISBN (Electronic)9788885741065
ISBN (Print)978-88-85741-03-4
Publication statusPublished - 2018
Event30th European Modeling and Simulation Symposium, EMSS 2018 - Budapest, Hungary
Duration: 17 Sept 201819 Sept 2018

Publication series

Name30th European Modeling and Simulation Symposium, EMSS 2018

Conference

Conference30th European Modeling and Simulation Symposium, EMSS 2018
Country/TerritoryHungary
CityBudapest
Period17.09.201819.09.2018

Keywords

  • Distributed simulation execution
  • HeuristicLab
  • High-performance cluster
  • Software architecture

Fingerprint

Dive into the research topics of 'Distributed simulation execution on a high-performance cluster using heuristiclab hive'. Together they form a unique fingerprint.

Cite this