Combining Greedy Hyperbolic Routing with Back – pressure Scheduling for Better Network Performance
Authors :
Rabieekenari, Ladan
Conference : 22nd International Symposium on Mathematical Theory of Networks and Systems pp. 698-705 , Minneapolis
Date: July 11 - July 15, 2016
The back-pressure algorithm (BP) has received much attention for jointly routing and scheduling over multihop networks. The back-pressure algorithm is throughput optimal, but it has poor delay performance. This is due to selection of unnecessarily long paths and routing loops, as the algorithm operates without accounting for the network topology. In this paper, we propose a throughput-optimal routing and scheduling algorithm that improves delay performance by greedy embedding of the network in hyperbolic space. We improve delay performance by solving an optimization problem,
which aims to send packets mostly to greedy neighbors, subject to throughput-optimality constraints. The algorithm that solves this optimization problem has a design parameter M. We study the effect of M on delay performance analytically. We validate our theoretical results via simulations and demonstrate that the proposed algorithm improves the delay performance.