Distributed Strategic Learning for Wireless Engineers

9

Learning in Risk-Sensitive Games

9.1    Introduction

Uncertainty, noise, and time delays are frequently met in wireless networks and communications in multiuser systems. Under such configuration, it is crucial to have adaptive and learning schemes based on the perceived measurements.

This chapter proposes a theoretical framework for learning in risk-sensitive dynamic robust games. The risk-sensitive model began with two key observations: first, that a player chooses between lotteries that the other players generate, and second, that payoffs are not the same as utilities, as models of learning in games often implicitly assume. It then becomes important to study the consequences of risk-aversion, or, more generally, utility functions that ...