Publication Date:
2013-02-28
Description:
We consider multi-antenna base stations using orthogonalfrequency-division multiple access and space division multipleaccess techniques to serve single-antenna users. Some users, calledreal-time users, have minimum rate requirements and must be servedin the current time slot while others, called non real-time users,do not have strict timing constraints and are served on abest-effort basis. The resource allocation (RA) problem is to findthe assignment of users to subcarriers and the transmit beamformingvectors that maximize the total user rates subject to power andminimum rate constraints. In general, this is a nonlinear andnon-convex program and the zero-forcing technique used here makes itinteger as well, exact optimal solutions cannot be computed inreasonable time for realistic cases. For this reason, we present atechnique to compute both upper and lower bounds and show that theseare quite close for some realistic cases. First, we formulate thedual problem whose optimum provides an upper bound to all feasiblesolutions. We then use a simple method to get a primal-feasiblepoint starting from the dual optimal solution, which is a lowerbound on the primal optimal solution. Numerical results for severalcases show that the two bounds are close so that the dual method canbe used to benchmark any heuristic used to solve this problem. As anexample, we provide numerical results showing the performance gap ofthe well-known weight adjustment method and show that there isconsiderable room for improvement.
Print ISSN:
1687-1472
Electronic ISSN:
1687-1499
Topics:
Electrical Engineering, Measurement and Control Technology
,
Computer Science
Permalink