Fighting for Routes: Resource Allocation among Competing Planners in Transportation Networks

In transportation networks, incomplete information is ubiquitous, and users often delegate their route choice to distributed planners. To model study these systems, we introduce network control games, consisting of multiple actors seeking optimise the social welfare assigned subpopulations through resource allocation in an underlying nonatomic congestion game. We first analyse inefficiency routing equilibria by calculating Price Anarchy for polynomial cost functions, then, using Asynchronous Advantage Actor–Critic algorithm implementation, show that reinforcement learning agents are vulnerable choosing suboptimal as predicted theory. Finally, extend analysis allow vehicles choose planner associated equilibria. Our results can be applied mitigate issues arising large transport networks with controlled autonomous vehicles.