Title: Completely Uncoupled Algorithms for Network Utility Maximization
Roll No: EE12D036
Guide: Dr. VenkateshRamaiyan
The recent advances in wireless systems demand addressing the following resource allocation problems,viz channel selection, user association and power control. The solution to these problems should address the following objectives: (i) Network throughput optimality is ensured (ii) Users get a fair share of the network throughput. Also in a heterogeneous network, where multiple radio technologies coexist, a distributed solution is preferable.
In the previous talk (Seminar I), we presented two fully distributed algorithms which provide solutions to the above problems.We assumed that the users’ decisions are based only
on their past actions and payoffs which is popularly known as completely uncoupled. The
objective of the algorithms is to maximize the sum utility of the network.
In this talk, we extend our work to a state based model, where we consider a finite set, S of system states. The state could represent the channel in the case of a wireless network or a background algorithm, which the users do not have control over but affects their performance.We assume that users’ rate is a function of the system state and the action profile of the users inthe current time slot. We consider three cases depending on users’ knowledge of the state andstate transition model. In the first case, we assume that the state transitions are deterministicand unknown to the users. Secondly, we assume that the state is known to the users. In thiscase, we allow for random state transitions. Finally, we assume that the state is unknown to theusers and the state transitions are . In this work, we present distributed utility maximizationalgorithms for the above mentioned scenarios. We show that our algorithms are indeed optimal.