Power optimization on a random wireless network: A statistical physics approach
Aris Moustakas (National and Capodistrian University of Athens, Dept of Physics)
The efficient use of transmitted power is important in modern day wireless networks. Not only is the battery life of transmitting devices thus extended, but also the interference caused to neighboring transmitters is minimized. There are iterative methods to obtain the optimal power per node in a network, however it is known that depending on the strength of the interference, the problem may be infeasible. However, the properties of the optimal power vector in a random network, as well as the convergence to it are not well understood. In this work we study power optimization in a simple network with randomness. We show that the problem can be mapped to an Anderson model and we analyze its properties, such as the average power per transmitter, its probability of failure, the tails of its power distribution and dynamics using ideas from the theory of disordered metals.
café/thé seront servis à 14h00