Optimal Paths Design for a GMPLS Network using the Lagrangian Relaxation Method

Abstract

We describe an optimal path design for a GMPLS network that uses the Lagrangian relaxation method, which can estimate the lower bounds of the solution to a problem. This feature helps the designer of the problem to take the accuracy of the solution obtained by the calculation into consideration when he makes a decision to assign the solution to a real network in critical situations. A formulation of the problem and how to solve it using the Lagrangian relaxation method is described, and the results obtained by a prototype and considerations are shown in this paper.