Joint QoS Routing and Dynamic Capacity Dimensioning with Elastic Traffic: A Game Theoretical Perspective

Efficient dynamic resource provisioning algorithms are necessary to the development and automation of Quality of Service (QoS) networks. The main goal of these algorithms is to offer services that satisfy the QoS requirements of individual users while guaranteeing at the same time an efficient utilization of network resources. This paper proposes a novel game theoretical model that solves the joint problem of non-cooperative QoS routing and dynamic capacity allocation in a parallel links network. Two categories of players are introduced: (1) the capacity players that dimension the link capacities to provide QoS guarantees to users, minimizing, at the same time, the links´ congestion, and (2) network users, which are characterized by elastic traffic demands and split their traffic over multiple links, maximizing their objective function. This game is modeled as a multi-leader-follower game, where capacity players are leaders and network users are followers. We derive optimal routing and capacity settings using a round robin greedy algorithm, discussing numerical examples that provide insights into the model´s solution.