Stage-oe-small.jpg

Inproceedings3185: Unterschied zwischen den Versionen

Aus Aifbportal
Wechseln zu:Navigation, Suche
 
(Eine dazwischenliegende Version von einem anderen Benutzer wird nicht angezeigt)
Zeile 17: Zeile 17:
 
|Year=2011
 
|Year=2011
 
|Month=Juni
 
|Month=Juni
|Booktitle=Proceedings of the Internantional Workshop on Bio-inspired Approaches for Distributed Computing (BADS)
+
|Booktitle=Proceedings of the International Workshop on Bio-inspired Approaches for Distributed Computing (BADS)
 
|Pages=49-56
 
|Pages=49-56
 
|Organization=International Conference on Autonomic Computing (ICAC 2011)
 
|Organization=International Conference on Autonomic Computing (ICAC 2011)
Zeile 29: Zeile 29:
 
resources. The novelty here is that there is no need to specify the number of parallel computing resources in the beginning of the optimization. This number is estimated during the optimization process by the resources. The proposed new framework of SIPO is described in this paper with respect to multi-objective optimization problems but it has a much larger scope. A comparative evaluation of using SIPO in multi-objective optimization problems shows that this adaptive approach can obtain equally good or sometimes even
 
resources. The novelty here is that there is no need to specify the number of parallel computing resources in the beginning of the optimization. This number is estimated during the optimization process by the resources. The proposed new framework of SIPO is described in this paper with respect to multi-objective optimization problems but it has a much larger scope. A comparative evaluation of using SIPO in multi-objective optimization problems shows that this adaptive approach can obtain equally good or sometimes even
 
better solutions than other parallel and non-parallel methods which are not self-organized.
 
better solutions than other parallel and non-parallel methods which are not self-organized.
 +
|Download=Bads04a-mostaghim.pdf,
 
|Forschungsgruppe=Effiziente Algorithmen
 
|Forschungsgruppe=Effiziente Algorithmen
 
}}
 
}}

Aktuelle Version vom 14. Mai 2012, 06:58 Uhr


Self-organized Invasive Parallel Optimization


Self-organized Invasive Parallel Optimization



Published: 2011 Juni

Buchtitel: Proceedings of the International Workshop on Bio-inspired Approaches for Distributed Computing (BADS)
Seiten: 49-56
Verlag: ACM
Erscheinungsort: Karlsruhe
Organisation: International Conference on Autonomic Computing (ICAC 2011)

Referierte Veröffentlichung

BibTeX

Kurzfassung
Self-organized Invasive Parallel Optimization (SIPO) is a new framework for solving optimization problems on parallel platforms. In contrast to existing approaches, the resources in SIPO are self-organized and represented as a unified resource to the user who specifies the optimization problem and its preferences to the system. SIPO starts working with one resource and automatically divides the optimization task stepwise into smaller tasks which are assigned to more resources. This job assignment is decided on demand by the resources. The novelty here is that there is no need to specify the number of parallel computing resources in the beginning of the optimization. This number is estimated during the optimization process by the resources. The proposed new framework of SIPO is described in this paper with respect to multi-objective optimization problems but it has a much larger scope. A comparative evaluation of using SIPO in multi-objective optimization problems shows that this adaptive approach can obtain equally good or sometimes even better solutions than other parallel and non-parallel methods which are not self-organized.

Download: Media:Bads04a-mostaghim.pdf



Forschungsgruppe

Effiziente Algorithmen


Forschungsgebiet

Organic Computing, Parallele Algorithmen, Multikriterielle Optimierung, Optimierung