Lagrangian-relaxation based mesh traffic grooming for network profit optimization

The optimization of grooming static traffic is formulated in this paper as a constrained integer linear programming problem and the objective is to minimize the summation of the routing cost and the rejection penalty of the traffic demands. The routing cost takes into account both the cost in the electronic and the optical domains. Such optimization objective represents the network profit optimization and is of practical significance in network planning. Through the optimization process, traffic demands are selectively accepted based on the profit they generate, which is the difference between the revenue and the network cost. We propose a decomposition approach using Lagrangian-relaxation, which can also obtain the theoretical bounds for traffic grooming problem. The comparison with some previous research shows the great advantage of our solution framework. The optimization results for some network examples also indicate that the proposed algorithm achieves good near-optimal solutions.