Publication
On asynchronous parallelization of order-based GA over grid-enabled heterogenous commodity hardware
dc.contributor.author | Valente de Oliveira, JOSÉ | |
dc.contributor.author | Baltazar, Sérgio | |
dc.contributor.author | Daniel, Helder | |
dc.date.accessioned | 2019-11-20T15:07:15Z | |
dc.date.available | 2019-11-20T15:07:15Z | |
dc.date.issued | 2017-11 | |
dc.description.abstract | In real-world applications, the runtime of genetic algorithms (GAs) can be computationally demanding, an issue that can be mitigated using parallelization. The study evaluates the parallelization of order-based GAs using the island model in an asynchronous heterogeneous computing environment. The island model allows for a considerable number of migration topologies. The study offers a systematic review of the studies on migration topologies and observes that no study is available yet on the performance of these migration topologies over asynchronous heterogeneous environments. Based on a statistical analysis of a comprehensive set of experiments, using real-world TSPLIB instances, the study researches the question: What is the fastest island model topology for order-based genetic algorithm, in an asynchronous distributed heterogeneous grid-enabled commodity computing environment, without losing significant fitness comparatively to the correspondent sequential panmictic implementation of the same algorithm?. Moreover, a new speedup index, the expected root speedup, is also proposed. A diversity of topology types and characteristics are considered: the single node, star, ring, cartwheel, rooted ordered tree, rooted full binary tree, coordinated tree-ring, and feedforward fully connected layered type. Different number of nodes are also considered. While some of the types of topologies are well known, the coordinated tree-ring topology is a novelty. These types of topologies allow us to assess three notable cases: (i) no migration (isolated island), (ii) migration toward the coordinator only, and (iii) migration flows to, and from, the coordinator. | |
dc.description.sponsorship | FCT-The Portuguese Board of Science and Technology [UID/MULTI/00631/2013 - CEOT] | |
dc.description.version | info:eu-repo/semantics/publishedVersion | |
dc.identifier.doi | 10.1007/s00500-016-2190-2 | |
dc.identifier.issn | 1432-7643 | |
dc.identifier.issn | 1433-7479 | |
dc.identifier.uri | http://hdl.handle.net/10400.1/12949 | |
dc.language.iso | eng | |
dc.peerreviewed | yes | |
dc.publisher | Springer | |
dc.subject | Traveling Salesman Problem | |
dc.subject | Genetic Algorithms | |
dc.subject | Island Model | |
dc.subject | Evolutionary | |
dc.subject | Cluster | |
dc.subject | Implementation | |
dc.title | On asynchronous parallelization of order-based GA over grid-enabled heterogenous commodity hardware | |
dc.type | journal article | |
dspace.entity.type | Publication | |
oaire.awardURI | info:eu-repo/grantAgreement/FCT/5876/UID%2FMulti%2F00631%2F2013/PT | |
oaire.citation.endPage | 6368 | |
oaire.citation.issue | 21 | |
oaire.citation.startPage | 6351 | |
oaire.citation.title | Soft Computing | |
oaire.citation.volume | 21 | |
oaire.fundingStream | 5876 | |
person.familyName | LUÍS VALENTE DE OLIVEIRA | |
person.familyName | Baltazar | |
person.familyName | Daniel | |
person.givenName | JOSÉ | |
person.givenName | Sérgio | |
person.givenName | Helder | |
person.identifier.ciencia-id | 1F12-C1D3-7717 | |
person.identifier.orcid | 0000-0001-5337-5699 | |
person.identifier.orcid | 0000-0003-1270-0234 | |
person.identifier.orcid | 0000-0002-4477-736X | |
project.funder.identifier | http://doi.org/10.13039/501100001871 | |
project.funder.name | Fundação para a Ciência e a Tecnologia | |
rcaap.rights | restrictedAccess | |
rcaap.type | article | |
relation.isAuthorOfPublication | bb726e73-690c-4a33-822e-c47bdac3035b | |
relation.isAuthorOfPublication | 69408591-a742-4ac8-a18d-69c7bb050d5d | |
relation.isAuthorOfPublication | 414fdea2-ad8b-4ee8-8816-094a9802593a | |
relation.isAuthorOfPublication.latestForDiscovery | bb726e73-690c-4a33-822e-c47bdac3035b | |
relation.isProjectOfPublication | 093114eb-daa5-4e3d-98ce-bebbf4dde4b8 | |
relation.isProjectOfPublication.latestForDiscovery | 093114eb-daa5-4e3d-98ce-bebbf4dde4b8 |
Files
Original bundle
1 - 1 of 1
No Thumbnail Available
- Name:
- On asynchronous parallelization - 12949.pdf
- Size:
- 1.81 MB
- Format:
- Adobe Portable Document Format