Optimal Transmission over Decentralized Gaussian MAC with Large Users

Abstract

In this project, we investigate a recently proposed alternating maximization (AM) method to numerically solve the optimal decentralized powers and ergodic capacity of wireless Gaussian MACs. This method is suited well for MAC when there are small number of transmitters. For moderate to large MACs, the numerical solutions are possible but at very high expense of computational cost. We seek to investigate this AM algorithm for the decentralized MAC to improve its computational efficiency. In this direction, we notice a key observation that central limit theorem can be invoked for the MAC with large users to solve the partial optimization quickly for the identical-users MAC settings. We also investigate the possibility of increasing the computational efficiency of the AM algorithm by making use of the fast Fourier transform (FFT) algorithms.