Parallelized bayesian optimization for expensive robot controller evolution

Margarita Rebolledo*, Frederik Rehbach, A. E. Eiben, Thomas Bartz-Beielstein

*Corresponding author for this work

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

4 Downloads (Pure)

Abstract

An important class of black-box optimization problems relies on using simulations to assess the quality of a given candidate solution. Solving such problems can be computationally expensive because each simulation is very time-consuming. We present an approach to mitigate this problem by distinguishing two factors of computational cost: the number of trials and the time needed to execute the trials. Our approach tries to keep down the number of trials by using Bayesian optimization (BO) –known to be sample efficient– and reducing wall-clock times by parallel execution of trials. We compare the performance of four parallelization methods and two model-free alternatives. Each method is evaluated on all 24 objective functions of the Black-Box-Optimization-Benchmarking (BBOB) test suite in their five, ten, and 20-dimensional versions. Additionally, their performance is investigated on six test cases in robot learning. The results show that parallelized BO outperforms the state-of-the-art CMA-ES on the BBOB test functions, especially for higher dimensions. On the robot learning tasks, the differences are less clear, but the data do support parallelized BO as the ‘best guess’, winning on some cases and never losing.

Original languageEnglish
Title of host publicationParallel Problem Solving from Nature – PPSN XVI
Subtitle of host publication16th International Conference, PPSN 2020, Leiden, The Netherlands, September 5-9, 2020, Proceedings, Part I
EditorsThomas Bäck, Mike Preuss, André Deutz, Michael Emmerich, Hao Wang, Carola Doerr, Heike Trautmann
PublisherSpringer Science and Business Media Deutschland GmbH
Pages243-256
Number of pages14
Volume1
ISBN (Electronic)9783030581121
ISBN (Print)9783030581114
DOIs
Publication statusPublished - 2020
Event16th International Conference on Parallel Problem Solving from Nature, PPSN 2020 - Leiden, Netherlands
Duration: 5 Sep 20209 Sep 2020

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12269 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference16th International Conference on Parallel Problem Solving from Nature, PPSN 2020
CountryNetherlands
CityLeiden
Period5/09/209/09/20

Keywords

  • Bayesian optimization
  • BBOB benchmarking
  • CMA-ES
  • Parallelization
  • Robotics

Fingerprint

Dive into the research topics of 'Parallelized bayesian optimization for expensive robot controller evolution'. Together they form a unique fingerprint.

Cite this