Runtime analysis of evolutionary algorithms on randomly constructed high-density satisfiable 3-CNF formulas

Andrew M. Sutton; Frank Neumann

doi:10.1007/978-3-319-10762-2_93

Runtime analysis of evolutionary algorithms on randomly constructed high-density satisfiable 3-CNF formulas

Andrew M. Sutton, Frank Neumann

Computer Science (Duluth)

Research output: Contribution to journal › Article › peer-review

11 Scopus citations

Abstract

We show that simple mutation-only evolutionary algorithms find a satisfying assignment on two similar models of random planted 3-CNF Boolean formulas in polynomial time with high probability in the high constraint density regime. We extend the analysis to random formulas conditioned on satisfiability (i.e., the so-called filtered distribution) and conclude that most high-density satisfiable formulas are easy for simple evolutionary algorithms. With this paper, we contribute the first rigorous study of randomized search heuristics from the evolutionary computation community on well-studied distributions of random satisfiability problems.

Original language	English (US)
Pages (from-to)	942-951
Number of pages	10
Journal	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	8672
DOIs	https://doi.org/10.1007/978-3-319-10762-2_93
State	Published - 2014

Bibliographical note

Publisher Copyright:
© Springer International Publishing Switzerland 2014.

Access

10.1007/978-3-319-10762-2_93

OpenUrl availability

Full text

Cite this

Runtime analysis of evolutionary algorithms on randomly constructed high-density satisfiable 3-CNF formulas. / Sutton, Andrew M.; Neumann, Frank.
In: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 8672, 2014, p. 942-951.

Research output: Contribution to journal › Article › peer-review

@article{b2f7911d4a7b4eaf84ecad56dca81e5d,

title = "Runtime analysis of evolutionary algorithms on randomly constructed high-density satisfiable 3-CNF formulas",

abstract = "We show that simple mutation-only evolutionary algorithms find a satisfying assignment on two similar models of random planted 3-CNF Boolean formulas in polynomial time with high probability in the high constraint density regime. We extend the analysis to random formulas conditioned on satisfiability (i.e., the so-called filtered distribution) and conclude that most high-density satisfiable formulas are easy for simple evolutionary algorithms. With this paper, we contribute the first rigorous study of randomized search heuristics from the evolutionary computation community on well-studied distributions of random satisfiability problems.",

author = "Sutton, {Andrew M.} and Frank Neumann",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing Switzerland 2014.",

year = "2014",

doi = "10.1007/978-3-319-10762-2_93",

language = "English (US)",

volume = "8672",

pages = "942--951",

journal = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

issn = "0302-9743",

publisher = "Springer Verlag",

}

TY - JOUR

T1 - Runtime analysis of evolutionary algorithms on randomly constructed high-density satisfiable 3-CNF formulas

AU - Sutton, Andrew M.

AU - Neumann, Frank

PY - 2014

Y1 - 2014

N2 - We show that simple mutation-only evolutionary algorithms find a satisfying assignment on two similar models of random planted 3-CNF Boolean formulas in polynomial time with high probability in the high constraint density regime. We extend the analysis to random formulas conditioned on satisfiability (i.e., the so-called filtered distribution) and conclude that most high-density satisfiable formulas are easy for simple evolutionary algorithms. With this paper, we contribute the first rigorous study of randomized search heuristics from the evolutionary computation community on well-studied distributions of random satisfiability problems.

AB - We show that simple mutation-only evolutionary algorithms find a satisfying assignment on two similar models of random planted 3-CNF Boolean formulas in polynomial time with high probability in the high constraint density regime. We extend the analysis to random formulas conditioned on satisfiability (i.e., the so-called filtered distribution) and conclude that most high-density satisfiable formulas are easy for simple evolutionary algorithms. With this paper, we contribute the first rigorous study of randomized search heuristics from the evolutionary computation community on well-studied distributions of random satisfiability problems.

UR - http://www.scopus.com/inward/record.url?scp=84921696278&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84921696278&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-10762-2_93

DO - 10.1007/978-3-319-10762-2_93

M3 - Article

AN - SCOPUS:84921696278

SN - 0302-9743

VL - 8672

SP - 942

EP - 951

JO - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

JF - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -

Runtime analysis of evolutionary algorithms on randomly constructed high-density satisfiable 3-CNF formulas

Abstract

Bibliographical note

Access

OpenUrl availability

Other files and links

Fingerprint

Cite this