Emergent Mind

The Price of Local Power Control in Wireless Scheduling

(1502.05279)
Published Feb 18, 2015 in cs.NI and cs.DS

Abstract

We consider the problem of scheduling wireless links in the physical model, where we seek an assignment of power levels and a partition of the given set of links into the minimum number of subsets satisfying the signal-to-interference-and-noise-ratio (SINR) constraints. Specifically, we are interested in the efficiency of local power assignment schemes, or oblivious power schemes, in approximating wireless scheduling. Oblivious power schemes are motivated by networking scenarios when power levels must be decided in advance, and not as part of the scheduling computation. We first show that the known algorithms fail to obtain sub-logarithmic bounds; that is, their approximation ratio are $\tilde\Omega(\log \max(\Delta,n))$, where $n$ is the number of links, $\Delta$ is the ratio of the maximum and minimum link lengths, and $\tilde\Omega$ hides doubly-logarithmic factors. We then present the first $O(\log{\log\Delta})$-approximation algorithm, which is known to be best possible (in terms of $\Delta$) for oblivious power schemes. We achieve this by representing interference by a conflict graph, which allows the application of graph-theoretic results for a variety of related problems, including the weighted capacity problem. We explore further the contours of approximability, and find the choice of power assignment matters; that not all metric spaces are equal; and that the presence of weak links makes the problem harder. Combined, our result resolve the price of oblivious power for wireless scheduling, or the value of allowing unfettered power control.

We're not able to analyze this paper right now due to high demand.

Please check back later (sorry!).

Generate a summary of this paper on our Pro plan:

We ran into a problem analyzing this paper.

Newsletter

Get summaries of trending comp sci papers delivered straight to your inbox:

Unsubscribe anytime.