“Beat-your-rival” routing games

To date, game theoretic models have been employed in virtually all networking contexts. These include control tasks at the network layer, such as flow control and routing (e.g., [14, 36, 161, 191, 202, 224] and references therein), as well as numerous studies on control tasks at the link and MAC layers. A fundamental assumption in all of these referenced studies is that the selfish agents compete over resources in the network and aim to optimize their own performance; agents do not care (either way) about the performance of their competitors. However, and typically in the context of routing, scenarios exist in which this assumption is not warranted.