Routing and dimensioning in optical WDM networks for dynamic traffic using post-optimization approach

摘要

Until now, optimization problems have been considered for routing and resource allocation of static traffic. In this paper, we formulate an integer optimization problem which jointly considers routing and dimensioning in WDM networks in the pre-optimization stage. Since the integer optimization problems are computationally difficult to solve, therefore the integer constraints of the integer optimization problem are relaxed using LP relaxation in the pre-optimization stage. The LP relaxation technique computationally makes the optimization problem to give solutions in short duration of time for various networks. The optimal results obtained from the pre-optimization stage are rounded to integer values in the post-optimization stage which give the numbers of wavelength channels. The solutions obtained from the pre-optimization stage are utilized jointly for routing and dimensioning in the post-optimization stage for dynamic traffic. The link blocking probabilities have been considered as a merit of quality of service for dynamic traffic in different WDM networks. It has been shown through simulations that the proposed two stage optimal routing and dimensioning technique outperforms existing routing and undimensioning WDM networks in various network scenarios.