| In this dissertation, we consider approaches for pricing multiple services offered over a single telecommunications network. Each service has quality of service (QoS) requirements that are guaranteed to users. Service classes may be defined by the type of service, such as voice, video or data, as well as the origin and destination of the connection provided to the user. We formulate our problems to address two issues, flow control and efficient allocation of resources. We proceed by first considering the ability to estimate demand quantities and the demand function on-line. We identify this problem as a critical issue in the design of pricing schemes, where the calculated price depends on a real-time estimate of demand. On-line estimation of demand, under a pricing policy that uses the estimated functions, is found to be prone to bias in the estimates. This serves to motivate the design of a novel flow control scheme for connection requests. We limit the number of connection requests by offering discounts in exchange for delayed use of connections. This allows us to regulate the proportion of blocked requests. Furthermore, the scheme can be implemented under uncertainty, since all the problem parameters can be estimated on-line. In the latter part of the dissertation, we address efficient allocation of resources. We formulate and solve problems that calculate optimal prices by service class. We simultaneously solve for the resource allocations necessary to provide connections with guaranteed QoS, to serve the demand resulting from the prices. We first consider single switch problems and study the impact of service class parameters on the optimal solutions. We derive optimality conditions and a solution method for this class of problems. We extend both the formulation and solution method for network problems. The optimality properties derived earlier still apply, in addition to properties dependent on the routing of traffic within the network. |
