Parallel transfer evolution algorithm

Yuanjun Laili, Lin Zhang, Yun Li

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Parallelization of an evolutionary algorithm takes the advantage of modular population division and information exchange among multiple processors. However, existing parallel evolutionary algorithms are rather ad hoc and lack a capability of adapting to diverse problems. To accommodate a wider range of problems and to reduce algorithm design costs, this paper develops a parallel transfer evolution algorithm. It is based on the island-model of parallel evolutionary algorithm and, for improving performance, transfers both the connections and the evolutionary operators from one sub-population pair to another adaptively. Needing no extra upper selection strategy, each sub-population is able to select autonomously evolutionary operators and local search operators as subroutines according to both the sub-population's own and the connected neighbor's ranking boards. The parallel transfer evolution is tested on two typical combinatorial optimization problems in comparison with six existing ad-hoc evolutionary algorithms, and is also applied to a real-world case study in comparison with five typical parallel evolutionary algorithms. The tests show that the proposed scheme and the resultant PEA offer high flexibility in dealing with a wider range of combinatorial optimization problems without algorithmic modification or redesign. Both the topological transfer and the algorithmic transfer are seen applicable not only to combinatorial optimization problems, but also to non-permutated complex problems.

LanguageEnglish
Pages686-701
Number of pages16
JournalApplied Soft Computing Journal
Volume75
Early online date5 Dec 2018
DOIs
Publication statusPublished - 28 Feb 2019

Fingerprint

Evolutionary algorithms
Combinatorial optimization
Parallel algorithms
Subroutines
Costs

Keywords

  • algorithmic adaptation
  • evolutionary computation
  • parallel algorithm
  • topological design

Cite this

Laili, Yuanjun ; Zhang, Lin ; Li, Yun. / Parallel transfer evolution algorithm. In: Applied Soft Computing Journal. 2019 ; Vol. 75. pp. 686-701.
@article{895e3d464c8347b4a102c3cdf06a6675,
title = "Parallel transfer evolution algorithm",
abstract = "Parallelization of an evolutionary algorithm takes the advantage of modular population division and information exchange among multiple processors. However, existing parallel evolutionary algorithms are rather ad hoc and lack a capability of adapting to diverse problems. To accommodate a wider range of problems and to reduce algorithm design costs, this paper develops a parallel transfer evolution algorithm. It is based on the island-model of parallel evolutionary algorithm and, for improving performance, transfers both the connections and the evolutionary operators from one sub-population pair to another adaptively. Needing no extra upper selection strategy, each sub-population is able to select autonomously evolutionary operators and local search operators as subroutines according to both the sub-population's own and the connected neighbor's ranking boards. The parallel transfer evolution is tested on two typical combinatorial optimization problems in comparison with six existing ad-hoc evolutionary algorithms, and is also applied to a real-world case study in comparison with five typical parallel evolutionary algorithms. The tests show that the proposed scheme and the resultant PEA offer high flexibility in dealing with a wider range of combinatorial optimization problems without algorithmic modification or redesign. Both the topological transfer and the algorithmic transfer are seen applicable not only to combinatorial optimization problems, but also to non-permutated complex problems.",
keywords = "algorithmic adaptation, evolutionary computation, parallel algorithm, topological design",
author = "Yuanjun Laili and Lin Zhang and Yun Li",
year = "2019",
month = "2",
day = "28",
doi = "10.1016/j.asoc.2018.11.044",
language = "English",
volume = "75",
pages = "686--701",
journal = "Applied Soft Computing",
issn = "1568-4946",

}

Parallel transfer evolution algorithm. / Laili, Yuanjun; Zhang, Lin; Li, Yun.

In: Applied Soft Computing Journal, Vol. 75, 28.02.2019, p. 686-701.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Parallel transfer evolution algorithm

AU - Laili, Yuanjun

AU - Zhang, Lin

AU - Li, Yun

PY - 2019/2/28

Y1 - 2019/2/28

N2 - Parallelization of an evolutionary algorithm takes the advantage of modular population division and information exchange among multiple processors. However, existing parallel evolutionary algorithms are rather ad hoc and lack a capability of adapting to diverse problems. To accommodate a wider range of problems and to reduce algorithm design costs, this paper develops a parallel transfer evolution algorithm. It is based on the island-model of parallel evolutionary algorithm and, for improving performance, transfers both the connections and the evolutionary operators from one sub-population pair to another adaptively. Needing no extra upper selection strategy, each sub-population is able to select autonomously evolutionary operators and local search operators as subroutines according to both the sub-population's own and the connected neighbor's ranking boards. The parallel transfer evolution is tested on two typical combinatorial optimization problems in comparison with six existing ad-hoc evolutionary algorithms, and is also applied to a real-world case study in comparison with five typical parallel evolutionary algorithms. The tests show that the proposed scheme and the resultant PEA offer high flexibility in dealing with a wider range of combinatorial optimization problems without algorithmic modification or redesign. Both the topological transfer and the algorithmic transfer are seen applicable not only to combinatorial optimization problems, but also to non-permutated complex problems.

AB - Parallelization of an evolutionary algorithm takes the advantage of modular population division and information exchange among multiple processors. However, existing parallel evolutionary algorithms are rather ad hoc and lack a capability of adapting to diverse problems. To accommodate a wider range of problems and to reduce algorithm design costs, this paper develops a parallel transfer evolution algorithm. It is based on the island-model of parallel evolutionary algorithm and, for improving performance, transfers both the connections and the evolutionary operators from one sub-population pair to another adaptively. Needing no extra upper selection strategy, each sub-population is able to select autonomously evolutionary operators and local search operators as subroutines according to both the sub-population's own and the connected neighbor's ranking boards. The parallel transfer evolution is tested on two typical combinatorial optimization problems in comparison with six existing ad-hoc evolutionary algorithms, and is also applied to a real-world case study in comparison with five typical parallel evolutionary algorithms. The tests show that the proposed scheme and the resultant PEA offer high flexibility in dealing with a wider range of combinatorial optimization problems without algorithmic modification or redesign. Both the topological transfer and the algorithmic transfer are seen applicable not only to combinatorial optimization problems, but also to non-permutated complex problems.

KW - algorithmic adaptation

KW - evolutionary computation

KW - parallel algorithm

KW - topological design

UR - http://www.scopus.com/inward/record.url?scp=85058468365&partnerID=8YFLogxK

U2 - 10.1016/j.asoc.2018.11.044

DO - 10.1016/j.asoc.2018.11.044

M3 - Article

VL - 75

SP - 686

EP - 701

JO - Applied Soft Computing

T2 - Applied Soft Computing

JF - Applied Soft Computing

SN - 1568-4946

ER -