Risk-Sensitive Optimal Control of Queues
Abstract
We consider the problem of designing risk-sensitive optimal control policies for scheduling packet transmissions in a stochastic wireless network. A single client is connected to an access point (AP) through a wireless channel. Packet transmission incurs a cost , while packet delivery yields a reward of units. The client maintains a finite buffer of size , and a penalty of units is imposed upon packet loss which occurs due to finite queueing buffer. We show that the risk-sensitive optimal control policy for such a simple set-up is of threshold type, i.e., it is optimal to carry out packet transmissions only when , i.e., the queue length at time exceeds a certain threshold . It is also shown that the value of threshold increases upon increasing the cost per unit packet transmission . Furthermore, it is also shown that a threshold policy with threshold equal to is optimal for a set of problems in which cost lies within an interval . Equations that need to be solved in order to obtain are also provided.
Keywords
Cite
@article{arxiv.1708.08178,
title = {Risk-Sensitive Optimal Control of Queues},
author = {Rahul Singh and Xueying Guo and Eytan Modiano},
journal= {arXiv preprint arXiv:1708.08178},
year = {2017}
}
Comments
accepted for publication in IEEE Conference on Decision and Control 2017