Dynamic Wireless Network Control Via Stochastic Approximation PDF Download

Author: Sina Firouzabadi
Publisher:
ISBN:
Category :
Languages : en
Pages :

Get Book Here

Book Description
This thesis investigates different stochastic approximation-based algorithms for performance optimization of wireless networks. Stochastic approximation is used to learn the randomly-varying characteristics of the network conditions and adapt the transmission strategies accordingly. The basic premise of wireless network optimization based on stochastic approximation will be presented, followed by several applications of this technique. The first application optimizes secondary user transmission strategies in cognitive networks with imperfect network state observations. In this setting the secondary user maximizes its revenue while generating a bounded performance loss to the primary users' network. The state of the primary users' network, defined as a collection of variables describing features of the network (e.g., buffer state, packet service state) evolves over time according to a homogeneous Markov process. The statistics of the Markov process are dependent on the strategy of the secondary user and, thus, the instantaneous idleness/transmission action of the secondary user has a long term impact on the temporal evolution of the network. The Markov process generates a sequence of states in the state space of the network that projects onto a sequence of observations in the observation space, that is, the collection of all the observations of the secondary user. Based on the sequence of observations from secondary users, an iterative stochastic approximation based algorithm is proposed that optimizes the strategy of the secondary users with no a priori knowledge of the statistics of the Markov process and of the state-observation probability map. The second application of stochastic approximation theory presented is around the design of green cellular networks through the use of distributed antennas. After presenting an information theoretic analysis of the ergodic capacity of distributed antenna systems in a cellular setting, optimized antenna placement in such systems is investigated. A general framework for this optimization based on stochastic approximation theory, with no constraint on the location of the antennas, will be presented. It will be shown that optimal placement of antennas inside the coverage region can significantly improve the power efficiency of wireless networks. As we will see, our stochastic optimization framework is sufficiently general to incorporate interference as well as general performance metrics. We will also present different numerical studies for illustrating the power efficiency and area spectral efficiency of distributed antenna systems, under different assumptions about availability of channel side information at the transmitter. Finally in the last part of the thesis, we present a distributed algorithm for optimizing the rate-reliability tradeoff in wireless networks with randomly time-varying channels. The stochastic optimization is based on wireless network utility maximization, extended to incorporate dynamics at the physical layer. The proposed algorithm enables a distributed implementation. We also verify the convergence of the proposed algorithm using Stochastic Approximation. The performance of the proposed algorithm and its convergence is illustrated via simulations.