Real-time game theoretic coordination of competitive mobility-on-demand systems

This paper considers competitive mobility-on-demand systems where a group of vehicle sharing companies, on one hand, want to collectively regulate the traffic of the user queueing network, and on the other hand, maximize their own profits at each time instant. We formulate the strategic interconnection among the companies as a real-time game theoretic coordination problem. We propose an algorithm to achieve vehicle balance and practical regulation of the user queueing network. We quantify the relation between the regulation error and the system parameters (e.g., the maximum variation of the user arrival rates).