Title
On performance optimization for multi-carrier MIMO ad hoc networks
Abstract
Broadband multi-carrier MIMO (MC-MIMO) is a promising technology that could provide significant capacity gain for wireless ad hoc networks. For MC-MIMO networks, since the capacity is affected by potential mutual interference on subcarriers, scheduling for subcarriers and algorithms for power control/allocation become key problems to harness their potential. However, due to non-convexity and large size of the underlying problem, there are few results on this important problem. In this paper, we first show that the non-convex problem for MC-MIMO networks satisfies the so-called concave perturbation condition, which gives a zero duality gap for the problem. This important result allows us to tackle the problem in the dual domain. The dual approach has the highly desirable benefit of reducing the complexity of the underlying problem, which allows us to design a near-optimal off-line algorithm. In addition to the off-line algorithm, we also devise an online adaptive algorithm (OAA) without the need of channel distribution information (CDI). We show that OAA is able to achieve the same result as the off-line algorithm.
Year
DOI
Venue
2009
10.1145/1530748.1530756
MobiHoc
Keywords
Field
DocType
dual domain,key problem,important problem,non-convex problem,off-line algorithm,underlying problem,performance optimization,online adaptive algorithm,dual approach,mc-mimo network,near-optimal off-line algorithm,ad hoc networks,non linear programming,wireless communication,mimo,numerical analysis,power control,wireless ad hoc network,stochastic programming,ad hoc network
Duality gap,Scheduling (computing),Computer science,Power control,MIMO,Computer network,Communication channel,Interference (wave propagation),Wireless ad hoc network,Adaptive algorithm,Distributed computing
Conference
Citations 
PageRank 
References 
4
0.44
23
Authors
4
Name
Order
Citations
PageRank
Jia Liu120218.09
Y. Thomas Hou22040186.12
Yi Shi32110120.50
Hanif D. Sherali43403318.40