With the deployment of fiber optics,the graph of a telecommunications transmission facility network is becoming sparse.An important step towards traffic (logical) network survivability is to provide diversity at the physical level by splitting requirements over disjoint paths so that if a transmission link fails,another path is still available.In this paper,we address the problem of obtaining optimal pairs of diverse routes to minimize cost for given demand,capacity and bound requirements.The mathematical formulation of this problem yields a mixedinteger,multi-commodity flow model which we solve by a Lagrangean relaxation based algorithm exploiting the structure of the problem.A preliminary study shows improvement in worst case network survivability with diverse routing.
展开▼