An adaptive lagrangian algorithm for solving the capacity and flow assignment problem in self-healing ATM networks

This paper addresses the problem of ATM networks survivability under two path restoration strategies, namely, the Global Reconfiguration and the Failure-Oriented Reconfiguration. For each restoration strategy, the problem is formulated as non-convex, non-linear multicommodity flow problem, where the objective function is given by the capacity installation cost and the routing cost. These problems are presented as special cases of the capacity and flow assignment (CFA) problem, and are solved by means of a special augmented Lagrangian approach (socalled the Separable Augmented Lagrangian Algorithm (SALA)). Numerical results are provided, comparing both restoration strategies in terms of network costs.