In this paper we consider a power allocation problem in multi-class wireless systems. We focus on the downlink of the system. Each mobile has a utility function that characterizes its degree of satisfaction for the received service. The objective is to obtain a power allocation that maximizes the total system utility. Typically, natural utility functions for each mobile are nonconcave. Hence, we cannot use existing convex optimization techniques to derive a global optimal solution. We develop a simple (distributed) algorithm to obtain a power allocation that is asymptotically optimal in the number of mobiles. The algorithm is based on dynamic pricing and consists of two stages. At the mobile selection stage, the base station selects mobiles to which power is allocated. At the power allocation stage, the base station allocates power to the selected mobiles. We provide numerical results that illustrate the performance of our scheme. In particular, we show that our algorithm results in system performance that is close to the performance of a global optimal solution in most cases.
Bibliographical noteFunding Information:
Manuscript received September 9, 2002; revised December 17, 2003; approved by IEEE/ACM TRANSACTIONS ON NETWORKING Editor M. Gross-glauser. This work was supported in part by the National Science Foundation under NSF Grants ANI-0073359, ANI-9805441, and ANI-0207728. This research was conducted when J.-W. Lee and R. R. Mazumdar were at Purdue University.
All Science Journal Classification (ASJC) codes
- Computer Science Applications
- Computer Networks and Communications
- Electrical and Electronic Engineering