In this paper we present a new competitive packet routing model with edge priorities. We consider players that route selfishly through a network over time and try to reach their destinations as fast as possible. If the number of players who want to enter an edge at the same time exceeds the inflow capacity of this edge, edge priorities with respect to the preceding edge solve these conflicts. Our edge priorities are well-motivated by applications in traffic. For this class of games, we show the existence of equilibrium solutions for single-source-single-sink games and we analyze structural properties of these solutions. We present an algorithm that computes Nash equilibria and we prove bounds both on the Price of Stability and on the Price of Anarchy. Moreover, we introduce the new concept of a Price of Mistrust. Finally, we also study the relations to earliest arrival flows.
MotivationRouting games over time are widely studied [8,15,16] due to various applications, e.g., road and air traffic control, logistic in production systems, communication networks like the internet, and financial flows. Depending on applications, there is a huge range of models. There are games with non-atomic [16] as well as atomic players [15], further there are games with continuous time [15] as well as games with discrete time steps [14].Usually, players are allowed to enter an edge directly at their arrival time. An essential ingredient of these models is the priority rule for tie-breaking. If two or more players arrive at a node at the same time, one needs to decide which player may enter the subsequent link first. Some models maintain the overall arrival order with a first-in firstout rule [8,15,16], i.e., only players with the exact same arrival time are re-arranged. Other models allow overtaking by higher prioritized players with global priority lists [14,15] or local priority lists [14] of players, that is, priority is a player inherent property and players are linearly ordered.