Abstract
This paper presents a parallel search parallel move approach to parallelise neighbourhood search algorithms on many-core platforms. In this approach, a large number of searches are run concurrently and coordinated periodically. Iteratively, each search generates and evaluates multiple moves in parallel. The proposed approach can fully utilise the computing capability of many-core platforms under various platform specific constraints. A parallel simulated annealing algorithm for solving the travelling salesman problem is developed using the parallel search parallel move scheme and implemented on an NVIDIA Tesla C2050 GPU platform. We evaluate the performance of our approach against a multi-threaded CPU implementation on a server containing two Intel Xeon X5650 CPUs 12 cores in total. The experimental results of 20 benchmark problems show that the GPU implementation achieves 99 times speedup on average in solution space exploration speed. In terms of effectiveness, the GPU implementation is capable of finding good solutions 39.5 times faster or with 21.7% solution quality improvement given the same searching time.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: International Journal of Computational Science and Engineering
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
