Institutional Repository [SANDBOX]
Technical University of Crete
EN  |  EL

Search

Browse

My Space

An asynchronous parallel differential evolution algorithm

Nikolos Ioannis, Marina S. Ntipteni, Ioannis M. Valakos

Full record


URI: http://purl.tuc.gr/dl/dias/10290966-6BA3-415C-9B79-CB201BC3A134
Year 2006
Type of Item Conference Paper Abstract
License
Details
Bibliographic Citation M. S. Ntipteni, I. M. Valakos, I. K. Nikolos. An Asynchronous Parallel Differential Evolution Algorithm. Presented at Conference on Design Optimization: Methods & Applications. [Online]. Available: http://velos0.ltt.mech.ntua.gr/ERCOFTAC/PROC06-GC/resources/fullpapers/paper_Nikolos20060317130050.pdf
Appears in Collections

Summary

A Parallel Differential Evolution algorithm is presented in this work, developed for a cluster ofcomputers in Windows environment. The parallelization is realized using an asynchronousapproach, utilizing a Master-Slave architecture. A separate executable program is used to evolveeach member of the population. The current population is stored in a folder accessible by allexecutables; each current member of the population, along with its fitness, is stored in a separatetext file contained in this common folder. Each slave program uses the information stored in thecommon folder to evolve the corresponding member of the population and to update theinformation stored in the corresponding text file, independently from the rest executables. Morethan one executables may be assigned to each computer. The procedure is tested in two airfoiloptimization problems and the parallel code is compared to a serial one, with respect to theconvergence behaviour, the quality of the optimum solution and the total computation time.

Services

Statistics