Abstract - In this paper, the discrete link capacity assignment problem for multiservice networks is considered. The discreteness of the link capacities is forced by the telecommunications equipment technology and vendors. Here, the designer is required to find the cheapest capacity assignment while satisfying the perfor-mance constraints for different classes of data originated by multiple services offered on the network. In order to solve this difficult optimization problem, two algorithms based on artificial intelligence search techniques, simulated annealing and genetic algorithms, are proposed. Performance of these methods are compared to a successful heuristic by Maruyama and Tang. The distinction between competing techniques becomes clearer as the network gets larger.
 

Paper in Adobe Acrobat  (73 KBytes)

Back to List of Papers
Back to NETLAB home page
Back to Cem Ersoy's home page