Unbiased black-box complexity of parallel search

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Authors

Colleges, School and Institutes

External organisations

  • Sheffield University
  • NOTTINGHAM UNIVERSITY

Abstract

We propose a new black-box complexity model for search algorithms evaluating λ search points in parallel. The parallel unbiased black-box complexity gives lower bounds on the number of function evaluations every parallel unbiased black-box algorithm needs to optimise a given problem. It captures the inertia caused by offspring populations in evolutionary algorithms and the total computational effort in parallel metaheuristics. Our model applies to all unary variation operators such as mutation or local search. We present lower bounds for the Leading-Ones function and general lower bound for all functions with a unique optimum that depend on the problem size and the degree of parallelism, λ. The latter is tight for OneMax; we prove that a (1+λ) EA with adaptive mutation rates is an optimal parallel unbiased black-box algorithm.

Details

Original languageEnglish
Title of host publicationParallel Problem Solving from Nature – PPSN XIII
Subtitle of host publication13th International Conference Ljubljana, Slovenia, September 13-17, 2014 Proceedings
EditorsThomas Bartz-Beielstein, Jürgen Branke, Bogdan Filipic, Jim Smith
Publication statusPublished - 24 Sep 2014
Event13th International Conference on Parallel Problem Solving from Nature – PPSN XIII - Ljubljana, Slovenia
Duration: 13 Sep 201417 Sep 2014

Publication series

NameLecture Notes in Computer Science
Volume8672
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference13th International Conference on Parallel Problem Solving from Nature – PPSN XIII
CountrySlovenia
City Ljubljana
Period13/09/1417/09/14

Keywords

  • Linear Speedup, Search Point, Island Model, Parallel Time, Clonal Selection Algorithm