Applications of Optimization

Modeling Tools

Integer Programming and Combinatorial Optimization

Stochastic and Dynamic Optimization

Game Theory and Equillibrium Models

Network Planning and Mulitpath Routing for Achieving a Target Quality of Service
The next generation of Internet is evolving from a "best effort" approach to one that is engineered for better performance in terms of Quality of Service (QoS). In this paper we show that by integrating QoS constraints within a planning model, we are able to provide routes for multi-path routing protocol. As a by-product of our resesearch, we also propose a multi-path routing algorithm for improving a variety of performance measures such as end-to-end delays, and packet losses. A linear programming model of the network is solved to obtain the optimal bandwidth values on the links for the given quality of service requirements. Since the LP model is merely an approximation, the actual delays and latencies differ during actual network operation. A better approximation of the actual operation is obtained by running a simulation in Network Simulator (NS). If the observed end-end delays and latencies do not satisfy the QoS requirements, the LP model is updated and a new set of link capacities and flows are computed. We also demonstrate the benefits of our multipath routing policy by comparing the results with those of conventional routing.


 

contact us * the department of systems & industrial engineering
the college of engineering & mines
* the university of arizona


© 2002-06 University of Arizona, Department of Systems & Industrial Engineering