Sciact
  • EN
  • RU

Fitness landscapes of buffer allocation problem for production lines with unreliable machines Full article

Journal Computers and Operations Research
ISSN: 0305-0548 , E-ISSN: 1873-765X
Output data Year: 2024, Volume: 172, Article number : 106819, Pages count : 12 DOI: 10.1016/j.cor.2024.106819
Tags Production line, Series–parallel structure, Unreliable machines, Buffer allocation, Genetic algorithms, Local optima
Authors Dolgui Alexandre 1 , Eremeev Anton V. 2 , Sigaev Vyatcheslav S. 3
Affiliations
1 IMT Atlantique
2 Sobolev Institute of mathematics, Omsk Department
3 Avtomatika-Servis LLC

Funding (1)

1 Russian Science Foundation 21-41-09017

Abstract: We study the structural properties of the buffer allocation problem from the fitness landscape perspective. We consider manufacturing flow lines with series–parallel network structure. The machines are supposed to be unreliable, their time to failure and repair time are exponentially distributed. Tentative solutions are evaluated by means of an approximate method based on the Markov models aggregation. We carry out computational experiments with local search and genetic algorithms in order to evaluate the fitness landscape properties of previously published instances and their modifications. It turns out that the so-called ‘massif central’ or ‘big valley’ structure of the fitness landscape is present but only partially: The fitness of local optima is negatively correlated with the distance to the best found solution, yet the set of local optima cannot be encompassed by a ball of relatively small size with respect to the size of solution space. Moreover, we show that in many problem instances, several clusters of local optima can be identified. The performance of genetic algorithms is discussed with respect to population clustering and the permanent usage of crossover is recommended.
Cite: Dolgui A. , Eremeev A.V. , Sigaev V.S.
Fitness landscapes of buffer allocation problem for production lines with unreliable machines
Computers and Operations Research. 2024. V.172. 106819 :1-12. DOI: 10.1016/j.cor.2024.106819 WOS Scopus РИНЦ OpenAlex
Dates:
Submitted: Nov 6, 2023
Accepted: Aug 19, 2024
Published online: Aug 22, 2024
Published print: Aug 27, 2024
Identifiers:
Web of science: WOS:001402992100001
Scopus: 2-s2.0-85202206028
Elibrary: 73909324
OpenAlex: W4401816025
Citing:
DB Citing
Scopus 5
OpenAlex 4
Elibrary 2
Web of science 4
Altmetrics: