Aumann-Shapley and Shapley value-based priority pricing: an approach to congestion control in data networks (extended summary)

Common problems in telecommunications networks are usually related to optimal network design and to establishing reasonable cost allocation criteria. Such cost allocation methods do not have to do only with infrastructure costs but also with the effects that a set of users cause on other users' quality of service by using shared network resources. This is precisely the problem of congestion. Although it is well known that this problem is not new and has been studied sufficiently, there has not been a common agreement on how to face it and it remains a challenging issue. We take as a starting point the results originally brought up by McLean and Sharkey (1994, 1997), who used the AumannShapley and Shapley value pricing rules to distribute congestion-related costs among the network users. In this paper we extend those results to the priority pricing framework. Specifically, we intend to quantify congestion, representing a data network by means of a non preemptive queue system with a single server, with an infinite buffer. It is also our purpose to explore the quality of the cost allocation mechanisms and to establish a comparative framework for them.