Differential evolution with an evolution path: a DEEP evolutionary algorithm

Yuan-Long Li, Zhi-Hui Zhan, Yue-Jiao Gong, Wei-Neng Chen, Jun Zhang, Yun Li

Research output: Contribution to journalArticle

77 Citations (Scopus)

Abstract

Utilizing cumulative correlation information already existing in an evolutionary process, this paper proposes a predictive approach to the reproduction mechanism of new individuals for differential evolution (DE) algorithms. DE uses a distributed model (DM) to generate new individuals, which is relatively explorative, whilst evolution strategy (ES) uses a centralized model (CM) to generate offspring, which through adaptation retains a convergence momentum. This paper adopts a key feature in the CM of a covariance matrix adaptation ES, the cumulatively learned evolution path (EP), to formulate a new evolutionary algorithm (EA) framework, termed DEEP, standing for DE with an EP. Without mechanistically combining two CM and DM based algorithms together, the DEEP framework offers advantages of both a DM and a CM and hence substantially enhances performance. Under this architecture, a self-adaptation mechanism can be built inherently in a DEEP algorithm, easing the task of predetermining algorithm control parameters. Two DEEP variants are developed and illustrated in the paper. Experiments on the CEC'13 test suites and two practical problems demonstrate that the DEEP algorithms offer promising results, compared with the original DEs and other relevant state-of-the-art EAs.

LanguageEnglish
Pages1798-1810
Number of pages13
JournalIEEE Transactions on Cybernetics
Volume45
Issue number9
Early online date9 Oct 2014
DOIs
Publication statusPublished - 30 Sep 2015

Fingerprint

Evolutionary algorithms
Covariance matrix
Momentum
Experiments

Keywords

  • cumulative learning
  • differential evolution
  • evolution path
  • evolutionary computation

Cite this

Li, Yuan-Long ; Zhan, Zhi-Hui ; Gong, Yue-Jiao ; Chen, Wei-Neng ; Zhang, Jun ; Li, Yun. / Differential evolution with an evolution path : a DEEP evolutionary algorithm. In: IEEE Transactions on Cybernetics. 2015 ; Vol. 45, No. 9. pp. 1798-1810.
@article{e505254c08ba4d12b4ae3388b95f6eb2,
title = "Differential evolution with an evolution path: a DEEP evolutionary algorithm",
abstract = "Utilizing cumulative correlation information already existing in an evolutionary process, this paper proposes a predictive approach to the reproduction mechanism of new individuals for differential evolution (DE) algorithms. DE uses a distributed model (DM) to generate new individuals, which is relatively explorative, whilst evolution strategy (ES) uses a centralized model (CM) to generate offspring, which through adaptation retains a convergence momentum. This paper adopts a key feature in the CM of a covariance matrix adaptation ES, the cumulatively learned evolution path (EP), to formulate a new evolutionary algorithm (EA) framework, termed DEEP, standing for DE with an EP. Without mechanistically combining two CM and DM based algorithms together, the DEEP framework offers advantages of both a DM and a CM and hence substantially enhances performance. Under this architecture, a self-adaptation mechanism can be built inherently in a DEEP algorithm, easing the task of predetermining algorithm control parameters. Two DEEP variants are developed and illustrated in the paper. Experiments on the CEC'13 test suites and two practical problems demonstrate that the DEEP algorithms offer promising results, compared with the original DEs and other relevant state-of-the-art EAs.",
keywords = "cumulative learning, differential evolution, evolution path, evolutionary computation",
author = "Yuan-Long Li and Zhi-Hui Zhan and Yue-Jiao Gong and Wei-Neng Chen and Jun Zhang and Yun Li",
note = "{\circledC} 2014 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.",
year = "2015",
month = "9",
day = "30",
doi = "10.1109/TCYB.2014.2360752",
language = "English",
volume = "45",
pages = "1798--1810",
journal = "IEEE Transactions on Cybernetics",
issn = "2168-2267",
number = "9",

}

Li, Y-L, Zhan, Z-H, Gong, Y-J, Chen, W-N, Zhang, J & Li, Y 2015, 'Differential evolution with an evolution path: a DEEP evolutionary algorithm' IEEE Transactions on Cybernetics, vol. 45, no. 9, pp. 1798-1810. https://doi.org/10.1109/TCYB.2014.2360752

Differential evolution with an evolution path : a DEEP evolutionary algorithm. / Li, Yuan-Long; Zhan, Zhi-Hui; Gong, Yue-Jiao; Chen, Wei-Neng; Zhang, Jun; Li, Yun.

In: IEEE Transactions on Cybernetics, Vol. 45, No. 9, 30.09.2015, p. 1798-1810.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Differential evolution with an evolution path

T2 - IEEE Transactions on Cybernetics

AU - Li, Yuan-Long

AU - Zhan, Zhi-Hui

AU - Gong, Yue-Jiao

AU - Chen, Wei-Neng

AU - Zhang, Jun

AU - Li, Yun

N1 - © 2014 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

PY - 2015/9/30

Y1 - 2015/9/30

N2 - Utilizing cumulative correlation information already existing in an evolutionary process, this paper proposes a predictive approach to the reproduction mechanism of new individuals for differential evolution (DE) algorithms. DE uses a distributed model (DM) to generate new individuals, which is relatively explorative, whilst evolution strategy (ES) uses a centralized model (CM) to generate offspring, which through adaptation retains a convergence momentum. This paper adopts a key feature in the CM of a covariance matrix adaptation ES, the cumulatively learned evolution path (EP), to formulate a new evolutionary algorithm (EA) framework, termed DEEP, standing for DE with an EP. Without mechanistically combining two CM and DM based algorithms together, the DEEP framework offers advantages of both a DM and a CM and hence substantially enhances performance. Under this architecture, a self-adaptation mechanism can be built inherently in a DEEP algorithm, easing the task of predetermining algorithm control parameters. Two DEEP variants are developed and illustrated in the paper. Experiments on the CEC'13 test suites and two practical problems demonstrate that the DEEP algorithms offer promising results, compared with the original DEs and other relevant state-of-the-art EAs.

AB - Utilizing cumulative correlation information already existing in an evolutionary process, this paper proposes a predictive approach to the reproduction mechanism of new individuals for differential evolution (DE) algorithms. DE uses a distributed model (DM) to generate new individuals, which is relatively explorative, whilst evolution strategy (ES) uses a centralized model (CM) to generate offspring, which through adaptation retains a convergence momentum. This paper adopts a key feature in the CM of a covariance matrix adaptation ES, the cumulatively learned evolution path (EP), to formulate a new evolutionary algorithm (EA) framework, termed DEEP, standing for DE with an EP. Without mechanistically combining two CM and DM based algorithms together, the DEEP framework offers advantages of both a DM and a CM and hence substantially enhances performance. Under this architecture, a self-adaptation mechanism can be built inherently in a DEEP algorithm, easing the task of predetermining algorithm control parameters. Two DEEP variants are developed and illustrated in the paper. Experiments on the CEC'13 test suites and two practical problems demonstrate that the DEEP algorithms offer promising results, compared with the original DEs and other relevant state-of-the-art EAs.

KW - cumulative learning

KW - differential evolution

KW - evolution path

KW - evolutionary computation

UR - https://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=6221036

U2 - 10.1109/TCYB.2014.2360752

DO - 10.1109/TCYB.2014.2360752

M3 - Article

VL - 45

SP - 1798

EP - 1810

JO - IEEE Transactions on Cybernetics

JF - IEEE Transactions on Cybernetics

SN - 2168-2267

IS - 9

ER -