PROJECT TITLE :

Relay Placement in Wireless Networks Minimizing Communication Cost - 2016

ABSTRACT:

Given n source nodes and k relay nodes, we have a tendency to model the optimal relay topology downside allowing for simultaneous optimization of the relay node locations and traffic through the network, so that the range of packet retransmissions is minimized. We tend to argue that state-of-the-art models and algorithms for relay placement in wireless networks don't mirror salient characteristics of the optimal relays topology and result in suboptimal solutions. We tend to do not constrain the position of relays to a finite set of discrete points, because the latter may not be feasible in sensible networks. In this case, we have a tendency to show that simply listing a set of feasible sites for the relays is already at least APX-onerous. Exploiting convexity in a very special case of the network communication cost perform, we tend to provide an optimal algorithm for the relay placement problem. However, the algorithm is exponential on the number of nodes in the network. We have a tendency to suggest a sensible heuristic algorithm for relay placement: RePlace. We have a tendency to compare RePlace numerically to the optimal algorithm and show that RePlace achieves the optimal or virtually optimal solutions. We have a tendency to implement RePlace in the full network stack simulator JiST/SWANS. The relay topologies generated by RePlace eliminate overhead communication cost nearly entirely.