Virtual Network Operators lease bandwidth from different data carriers to offer well managed Virtual Private Networks. By using proprietary algorithms and leased resource diversity they can offer Quality of Service, at competitive prices, which is difficult to attain from traditional data network operators. In this paper, we describe a novel distributed resource management approach, based on an economic model, that allows continuous optimizing of the network profit, while keeping network blocking constraints. The approach integrates leased link capacity adaptations, based on measurements, with connection admission control and routing policies. This is done in the framework that applies decomposition of Markov Decision Process. Our numerical analysis validates the approach and shows the convergence and stability of the adaptation as functions of measurement parameters.