Low complexity stable link scheduling for maximizing throughput in wireless networks

This paper presents novel distributed algorithms for scheduling transmissions in multi-hop wireless networks. Our algorithms generate new schedules in a distributed manner via simple local changes to existing schedules. Two classes of algorithms are designed: one assumes that the location information of all wireless nodes are known, and the other does not. Both classes of algorithms are parameterized by an integer k (called algorithm-k). We show that algorithm-k that uses geometry location achieves (1 - 2/k) ² of the capacity region, for every k ≤ 3; algorithm-k which does not use geometry location achieves 1/? of the capacity region, for every k ≤ 3 and a constant ? depending on k. Our algorithms have small worst-case overheads. Both classes of algorithms can generate a new schedule by requiring communications within θ(k) hops for every node, which can be implemented by letting each node transmit at most O(k) messages. The parameter k explicitly captures the tradeoff between control overhead and the throughput performance of any scheduler. Additionally, the class of algorithms with known geometry location of nodes can find a new schedule in time θ(k²Δ), where Δ is the minimum mini-time-slots such that each of the n nodes can communicate with its neighbors once, which is the minimum time-slots required by any scheduling algorithm.