Network routing games are instrumental in understanding traffic patterns and improving congestion in networks. They have a direct application in transportation and telecommunication networks and extend to many other applications such as task planning, etc. Theoretically, network games were one of the central examples in the development of algorithmic game theory.
In these games, multiple users need to route between different source-destination pairs and links are congestible, namely, each linkdelay is a non-decreasing function of the flow on the link. Many of the fundamental game theoretic questions are now well understood for these games, for example, does equilibrium exist, is it unique, can itbe computed efficiently, does it have a compact representation; the same questions can be asked of the socially optimal solution that minimizes the total user delay.
So far, most research has focused on the classical models in which thelink delays are deterministic. In contrast, real world applications contain a lot of uncertainty, which may stem from exogenous factors such as weather, time of day, weekday versus weekend, etc. or endogenous factors such as the network traffic. Furthermore, many users are risk-averse in the presence of uncertainty, so that they do not simply want to minimize expected delays and instead may need to add a buffer to ensure a guaranteed arrival time to a destination.
I present my recent work on a new stochastic network game model with risk-averse users. Risk-aversion poses a computational challenge and it often fundamentally alters the mathematical structure of the game compared to its deterministic counterpart, requiring new tools for analysis. The talk will discuss best response and equilibrium analysis, as well as equilibrium efficiency (price of anarchy) and a new concept, the price of risk.