Target-Driven and Incentive-Aligned Power Control for Wireless Networks
Benjamin Yolken, Nicholas Bambos · 2007
In this paper, we examine the wireless network power control problem. We first consider two approaches that have evolved in parallel, catering to distinct concerns: (a) a distributed approach with convergence to 'hard' SIR targets, introduced by Foschini and Miljanic [1] and (b) an incentive- based, game theoretic approach, investigated by Saraydar, Man- dayam, and Goodman [2] among others. We then seek to reconcile these two approaches and explore the rich space in between by formulating a utility-based model in which users have 'soft' SIR targets. We prove that, under certain cost conditions, a Nash Equilibrium of the resulting game is identical to the convergence point of the Foschini-Miljanic (FM) algorithm. Thus, one can use the latter in an incentive-compatible way. If these conditions are not met, however, then the system necessarily will operate at a non-FM point, i.e. one in which some SIR targets are not attained. We propose an algorithm for this case and show by simulation that the resulting Nash points may have desirable power efficiency properties. Thus, under our model, the network can be either aligned or not aligned with the FM scheme, each of which potentially has its advantages.