Please use this identifier to cite or link to this item:
|Scopus||Web of Science®||Altmetric|
|Title:||On optimising the capacity and power efficiency of a wireless network|
|Citation:||2012 Australasian Telecommunication Networks and Applications Conference (ATNAC): pp.1-7|
|Conference Name:||Australasian Telecommunication Networks and Applications Conference (2012 : Brisbane, Queensland)|
|Robert Hunjet, Andrew Coyle|
|Abstract:||With wireless networking becoming more prevalent, the capacity and power efficiency of these networks is of concern to both users and network designers. Wireless network capacity is difficult to define, with current research providing upper bounds and throughput results for given traffic profiles. However, these measures are not suitable as optimisation metrics. This paper first outlines a novel method to measure the average capacity metric of a network. Then we compare the optimisation of power efficiency to that of average capacity and investigate the effect each metric has on the other. Finally, a new approach to network optimisation is presented in the form of a multi-objective cross entropy optimisation. This optimisation simultaneously improves the capacity and power efficiency of the networks indicating that superior solutions can be found when considering the power and capacity problem in unison. When conducting the aforementioned optimisations, the impact that the insertion of additional nodes has upon the metrics is also investigated. The current literature suggests that the additional nodes would reduce the network capacity. Our results, however, show that this is not the case when appropriate transmission powers are utilised in conjunction with the node addition.|
|Keywords:||topology control; wireless networks; MANET; optimisation; cross entropy; multi-objective optimisation; power control; network capacity; power efficiency|
|Appears in Collections:||Electrical and Electronic Engineering publications|
Files in This Item:
There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.