Line-prioritized environmental selection and normalization scheme for many-objective optimization using reference-lines-based frameworkCitation formats

Standard

Line-prioritized environmental selection and normalization scheme for many-objective optimization using reference-lines-based framework. / Sharma, Deepak; Shukla, Pradyumn Kumar.

In: Swarm and Evolutionary Computation, Vol. 51, 100592, 12.2019.

Research output: Contribution to journalArticlepeer-review

Harvard

APA

Vancouver

Author

Bibtex

@article{d788cc9f263345d0a03a06ec6547c32e,
title = "Line-prioritized environmental selection and normalization scheme for many-objective optimization using reference-lines-based framework",
abstract = "The Pareto-dominance-based multi-objective evolutionary algorithms (MOEAs) have been successful in solving many test problems and other engineering optimization problems. However, their performance gets affected when solving more than 3-objective optimization problems due to lack of sufficient selection pressure. Many attempts have been made by the researchers toward improving the environmental selection of those MOEAs. One such attempt is selecting solutions using the reference-lines-based framework. In this paper, an efficient environmental selection and normalization scheme are proposed for this framework. The environmental selection operator is developed to equally prioritize solutions associated with different lines drawn from the origin and the reference points. A normalization scheme is also suggested in which the extreme point is used which gets updated on the designed rules. The framework is referred to as LEAF, and it is tested on 3-, 5-, 10-, and 15-objective DTLZ and WFG test instances. LEAF demonstrates its outperformance on almost all DTLZ instances and shows better performance on most of WFG instances over six MOEAs from the literature.",
keywords = "Environmental selection, Normalization, Many-objective optimization, Reference lines",
author = "Deepak Sharma and Shukla, {Pradyumn Kumar}",
year = "2019",
month = dec,
doi = "10.1016/j.swevo.2019.100592",
language = "English",
volume = "51",
journal = "Swarm and Evolutionary Computation",
issn = "2210-6502",
publisher = "Elsevier BV",

}

RIS

TY - JOUR

T1 - Line-prioritized environmental selection and normalization scheme for many-objective optimization using reference-lines-based framework

AU - Sharma, Deepak

AU - Shukla, Pradyumn Kumar

PY - 2019/12

Y1 - 2019/12

N2 - The Pareto-dominance-based multi-objective evolutionary algorithms (MOEAs) have been successful in solving many test problems and other engineering optimization problems. However, their performance gets affected when solving more than 3-objective optimization problems due to lack of sufficient selection pressure. Many attempts have been made by the researchers toward improving the environmental selection of those MOEAs. One such attempt is selecting solutions using the reference-lines-based framework. In this paper, an efficient environmental selection and normalization scheme are proposed for this framework. The environmental selection operator is developed to equally prioritize solutions associated with different lines drawn from the origin and the reference points. A normalization scheme is also suggested in which the extreme point is used which gets updated on the designed rules. The framework is referred to as LEAF, and it is tested on 3-, 5-, 10-, and 15-objective DTLZ and WFG test instances. LEAF demonstrates its outperformance on almost all DTLZ instances and shows better performance on most of WFG instances over six MOEAs from the literature.

AB - The Pareto-dominance-based multi-objective evolutionary algorithms (MOEAs) have been successful in solving many test problems and other engineering optimization problems. However, their performance gets affected when solving more than 3-objective optimization problems due to lack of sufficient selection pressure. Many attempts have been made by the researchers toward improving the environmental selection of those MOEAs. One such attempt is selecting solutions using the reference-lines-based framework. In this paper, an efficient environmental selection and normalization scheme are proposed for this framework. The environmental selection operator is developed to equally prioritize solutions associated with different lines drawn from the origin and the reference points. A normalization scheme is also suggested in which the extreme point is used which gets updated on the designed rules. The framework is referred to as LEAF, and it is tested on 3-, 5-, 10-, and 15-objective DTLZ and WFG test instances. LEAF demonstrates its outperformance on almost all DTLZ instances and shows better performance on most of WFG instances over six MOEAs from the literature.

KW - Environmental selection

KW - Normalization

KW - Many-objective optimization

KW - Reference lines

U2 - 10.1016/j.swevo.2019.100592

DO - 10.1016/j.swevo.2019.100592

M3 - Article

VL - 51

JO - Swarm and Evolutionary Computation

JF - Swarm and Evolutionary Computation

SN - 2210-6502

M1 - 100592

ER -