Abstract
Cuckoo Search is one of the recent swarm itelligence metaheuritics. It has been succesfuly applied to a number of optimization problems, but is stil not very well researched. In this paper we present a parallelized version of the Cuckoo Search algorithm. The parallelization is implemented using CUDA architecture. The algorithm is significantly changed
compared to the sequential version. Changes are partialy done to exploit the power of mass parallelization by the graphical processing unit and partialy as a consequence of the memory access restrictions that exist in CUDA. Tests on standard benchmark functions show that our proposed parallized algorithm greatly decreases the execution time and achieves similar or slightly better quality of the results compared to the sequential algorithm.
compared to the sequential version. Changes are partialy done to exploit the power of mass parallelization by the graphical processing unit and partialy as a consequence of the memory access restrictions that exist in CUDA. Tests on standard benchmark functions show that our proposed parallized algorithm greatly decreases the execution time and achieves similar or slightly better quality of the results compared to the sequential algorithm.
| Original language | English |
|---|---|
| Number of pages | 6 |
| Publication status | Published - 2013 |
| Externally published | Yes |