A new emigrant creation strategy based on local best sources for parallel Artificial Bee Colony algorithm

Abstract

Özetçe -Arıların yiyecek arama davranışlarını modelleyen Yapay Arı Koloni algoritması, en önemli sürü zekası temelli optimizasyon algoritmaları arasında yer alır. Diger popülasyon temelli evrimsel hesaplama yaklaşımlarında oldugu gibi, Yapay Arı Koloni algoritması da dagıtık sistemler üzerinde paralel modellenmeye uygundur. Bu çalışmada, alt kolonilere ayrılarak farklı hesaplama dügümleri üzerinde eş zamanlı çalıştırılan Yapay Arı Koloni algoritması için, alt koloniler arasında dagıtılacak çözümün belirlenmesi adına yeni bir yaklaşım sunulmuştur. Farklı sayıda hesaplama dügümü üzerinde paralelleştirilen Yapay Arı Koloni algoritmasının önerilen yaklaşımı, algoritmanın seri eşdegeri ile fonksiyon sonuçları ve yürütme zamanları açısından karşılaştırılmış ve önerilen göçmen oluşturma stratejisinin Yapay Arı Koloni algoritmasının yakınsama performansını önemli ölçüde artırdıgı görülmüştür. Anahtar Kelimeler-Yapay Arı Koloni algoritması, dagıtık sistemler, göçmen seçimi. Abstract-Artificial Bee Colony algorithm, inspired by the foraging behavior of real honey bees, is one of the most important swarm intelligence based optimization algorithms. Like other population based evolutionary computation techniques, ArtificialBee Colony algorithm is suitable for parallelization on distributed architectures. In this paper, we presented a new emigrant creation strategy that is being distributed between subcolonies running simultaneously on the independent compute nodes. The running times and objective function values obtained by the parallelized Artificial Bee Colony algorithm with the proposed model on different number of compute nodes are compared with the sequential counterpart of the algorithm and it is seen that convergence performance of the parallelized Artificial Bee Colony algorithm is significantly improved with the proposed emigrant creation strategy.

Deployment in wireless sensor networks by parallel and cooperative parallel artificial bee colony algorithms

“…However, some steps of the algorithms require sequential operations and a limited set of modifications on the fundamental workflow of them should be made when they are tried to be parallelized. Dividing the whole colony into equally sized small colonies and evaluating them simultaneously on the different computing units are probably the most preferred parallelization approach [24][25][26]. However, number of bees in computing units is usually not enough compared to the serial implementations on single computing unit and parallelization does not go beyond a method that only focusing improvement on the execution times [24][25][26].…”

“…The best solutions in each subcolony are the appropriate emigrant candidates and they usually change with the worst solutions in the neighbor subcolonies. This type of migration schema is the common part of the studies devoted to the parallelization and can be thought as the conventional approach [24][25][26]. However, if the best solutions can not be improved between subsequent migration periods, two or more copies of the same emigrant can be seen in the neighbor subcolony.…”