Optimal transmission policy in energy harvesting wireless communications: A learning approach
Lists
Only you see this
Abstract
We consider an energy harvesting wireless communication link, where arriving data packets have different importance values. The wireless transmitter needs to decide whether each arriving data packet should be transmitted or not, based on the packet's importance value, channel condition, and energy status. Under certain conditions, we show this high dimensional control problem can be transformed to a one dimensional continuous value function estimation problem using the notion of after-state. Then, by analyzing the structure of the value function, we propose a polynomial approximation to effectively compress the continuous function space into a finite weight space. Furthermore, we develop a reinforcement learning algorithm for our after-state setting. Finally, the proposed function approximation and learning algorithm are investigated under various system parameter settings via simulation.
BibTeX
@inproceedings{wu2017optimal,
author = {Keyu Wu and Chintha Tellambura and Hai Jiang},
title = {{Optimal transmission policy in energy harvesting wireless communications: A learning approach}},
booktitle = {ICC},
year = {2017},
doi = {10.1109/icc.2017.7997233},
}Loading related content...