Article: Traffic route optimisation based on clouding computing parallel ACS Journal: International Journal of Information and Communication Technology (IJICT) 2019 Vol.14 No.2 pp.204 - 217 Abstract: Intelligent traffic has demand for massive data environment and high performance processing, which needs cloud computing platform to process massive data and applying distributed parallel guidance algorithms to improve system efficiency. Therefore, this paper proposes an improved scheme based on clouding computing ACS algorithm. It first adopts MapReduce to parallelise traditional ACS, to process the solving problem with distributed parallel mode and to improve the defects in ACS. The improved ACS applies map function to parallelise the part which has the most time consuming, that is, the independent solving process of each ant. Then reduce function is used to describe the processes of pheromone updating and obtaining better solutions. Simultaneously, for the defects of ACS on long searching time and premature convergence to a non-optimal solution, we integrate simulated annealing algorithm to ACS and provide corresponding realisation process. The experiments construct Hadoop cloud computing platform and the improved algorithm is operated and tested on this platform. By the analysis on experimental results, we find the parallel ACS designed by us has improved the query efficiency of the shortest path, which also has advantage on the performance of running time and speedup ratio compared to classic algorithms. Inderscience Publishers - linking academia, business and industry through research

Title: Traffic route optimisation based on clouding computing parallel ACS

Authors: Changyu Li; Yang Zhao

Addresses: School of Civil Engineering, Changchun Institute of Technology, Changchun 130000, China ' Department of Mechanical and Electrical Engineering, Guangdong University of Science & Technology, Dongguan 523083, China

Abstract: Intelligent traffic has demand for massive data environment and high performance processing, which needs cloud computing platform to process massive data and applying distributed parallel guidance algorithms to improve system efficiency. Therefore, this paper proposes an improved scheme based on clouding computing ACS algorithm. It first adopts MapReduce to parallelise traditional ACS, to process the solving problem with distributed parallel mode and to improve the defects in ACS. The improved ACS applies map function to parallelise the part which has the most time consuming, that is, the independent solving process of each ant. Then reduce function is used to describe the processes of pheromone updating and obtaining better solutions. Simultaneously, for the defects of ACS on long searching time and premature convergence to a non-optimal solution, we integrate simulated annealing algorithm to ACS and provide corresponding realisation process. The experiments construct Hadoop cloud computing platform and the improved algorithm is operated and tested on this platform. By the analysis on experimental results, we find the parallel ACS designed by us has improved the query efficiency of the shortest path, which also has advantage on the performance of running time and speedup ratio compared to classic algorithms.

Keywords: cloud computing; ant colony system; ACS; MapReduce; traffic network; pheromone.

DOI: 10.1504/IJICT.2019.097693

International Journal of Information and Communication Technology, 2019 Vol.14 No.2, pp.204 - 217

Received: 16 Nov 2016
Accepted: 13 Feb 2017
Published online: 05 Feb 2019 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Traffic route optimisation based on clouding computing parallel ACS

Keep up-to-date