A LAGRANGEAN AND SUBGRADIENT METHOD FOR THE OPTIMIZATION OF WAVELENGTH ROUTING NETWORK DESIGN

摘要

In this paper, we study the logical topology design and reconfiguration problem in a DWDM network based on oar newest observation and former research. We formulated the problem into a constrained mixed-integer linear problem (MILP), and the problem is solved with Lagrangean relaxation (LR) technique and snbgradient method that can provide a lower bound on the objective function value. We provided a proof to the convergence the subgradient method with more than one set of Lagrange multipliers. We have also proposed a heuristics that would allow us to construct a feasible routing scheme. The effectiveness of our optimization approach is tested numerically based on the relative duality-gap measure. Based on the characteristics of the LR approach, we studied the reconfiguration issue of the network when there are minor changes of resources in the network.