University of Hertfordshire

From the same journal

By the same authors

Standard

Harvard

APA

Vancouver

Author

Bibtex

@article{e19ee2430fa24a688c08fc3f198d7685,
title = "How to exploit fitness landscape properties of timetabling problem: A newoperator for quantum evolutionary algorithm",
abstract = "The fitness landscape of the timetabling problems is analyzed in this paper to provide some insight into theproperties of the problem. The analyses suggest that the good solutions are clustered in the search space andthere is a correlation between the fitness of a local optimum and its distance to the best solution. Inspiredby these findings, a new operator for Quantum Evolutionary Algorithms is proposed which, during the searchprocess, collects information about the fitness landscape and tried to capture the backbone structure of thelandscape. The knowledge it has collected is used to guide the search process towards a better region in thesearch space. The proposed algorithm consists of two phases. The first phase uses a tabu mechanism to collectinformation about the fitness landscape. In the second phase, the collected data are processed to guide thealgorithm towards better regions in the search space. The algorithm clusters the good solutions it has foundin its previous search process. Then when the population is converged and trapped in a local optimum, itis divided into sub-populations and each sub-population is designated to a cluster. The information in thedatabase is then used to reinitialize the q-individuals, so they represent better regions in the search space.This way the population maintains diversity and by capturing the fitness landscape structure, the algorithmis guided towards better regions in the search space. The algorithm is compared with some state-of-the-artalgorithms from PATAT competition conferences and experimental results are presented.",
keywords = "Fitness landscape analysis, Quantum evolutionary algorithms, Timetabling-coloring",
author = "Mohammad Tayarani",
note = "{\textcopyright} 2020 Elsevier Ltd. All rights reserved. This is the accepted manuscript version of an article which has been published in final form at https://doi.org/10.1016/j.eswa.2020.114211",
year = "2020",
month = nov,
day = "4",
doi = "10.1016/j.eswa.2020.114211",
language = "English",
volume = "168",
journal = "Expert Systems with Applications",
issn = "0957-4174",
publisher = "Elsevier Limited",

}

RIS

TY - JOUR

T1 - How to exploit fitness landscape properties of timetabling problem: A newoperator for quantum evolutionary algorithm

AU - Tayarani, Mohammad

N1 - © 2020 Elsevier Ltd. All rights reserved. This is the accepted manuscript version of an article which has been published in final form at https://doi.org/10.1016/j.eswa.2020.114211

PY - 2020/11/4

Y1 - 2020/11/4

N2 - The fitness landscape of the timetabling problems is analyzed in this paper to provide some insight into theproperties of the problem. The analyses suggest that the good solutions are clustered in the search space andthere is a correlation between the fitness of a local optimum and its distance to the best solution. Inspiredby these findings, a new operator for Quantum Evolutionary Algorithms is proposed which, during the searchprocess, collects information about the fitness landscape and tried to capture the backbone structure of thelandscape. The knowledge it has collected is used to guide the search process towards a better region in thesearch space. The proposed algorithm consists of two phases. The first phase uses a tabu mechanism to collectinformation about the fitness landscape. In the second phase, the collected data are processed to guide thealgorithm towards better regions in the search space. The algorithm clusters the good solutions it has foundin its previous search process. Then when the population is converged and trapped in a local optimum, itis divided into sub-populations and each sub-population is designated to a cluster. The information in thedatabase is then used to reinitialize the q-individuals, so they represent better regions in the search space.This way the population maintains diversity and by capturing the fitness landscape structure, the algorithmis guided towards better regions in the search space. The algorithm is compared with some state-of-the-artalgorithms from PATAT competition conferences and experimental results are presented.

AB - The fitness landscape of the timetabling problems is analyzed in this paper to provide some insight into theproperties of the problem. The analyses suggest that the good solutions are clustered in the search space andthere is a correlation between the fitness of a local optimum and its distance to the best solution. Inspiredby these findings, a new operator for Quantum Evolutionary Algorithms is proposed which, during the searchprocess, collects information about the fitness landscape and tried to capture the backbone structure of thelandscape. The knowledge it has collected is used to guide the search process towards a better region in thesearch space. The proposed algorithm consists of two phases. The first phase uses a tabu mechanism to collectinformation about the fitness landscape. In the second phase, the collected data are processed to guide thealgorithm towards better regions in the search space. The algorithm clusters the good solutions it has foundin its previous search process. Then when the population is converged and trapped in a local optimum, itis divided into sub-populations and each sub-population is designated to a cluster. The information in thedatabase is then used to reinitialize the q-individuals, so they represent better regions in the search space.This way the population maintains diversity and by capturing the fitness landscape structure, the algorithmis guided towards better regions in the search space. The algorithm is compared with some state-of-the-artalgorithms from PATAT competition conferences and experimental results are presented.

KW - Fitness landscape analysis

KW - Quantum evolutionary algorithms

KW - Timetabling-coloring

UR - http://www.scopus.com/inward/record.url?scp=85095834054&partnerID=8YFLogxK

U2 - 10.1016/j.eswa.2020.114211

DO - 10.1016/j.eswa.2020.114211

M3 - Article

VL - 168

JO - Expert Systems with Applications

JF - Expert Systems with Applications

SN - 0957-4174

M1 - 114211

ER -