ARCHIVES - Telecommunuications...

 

Telecommunications Network Design

 

There have been a number of applications of optimization in the rapidly growing telecommunications industry. Two general types of applications are: a) the design of network structure, in particular, the extent to which hubs are used; and b) the design of fail-soft redundancy into a network.

For example, A. Dutta of George Mason University and Peter Kubat of GTE Laboratories, in a paper entitled "Design of Partially Survivable Networks for Cellular Telecommunication Systems" (European Journal of Operational Research, 1999), developed a model for deciding how cell sites in a cellular phone system should be connected to a land based network. The objective was to minimize cost while achieving a certain specified level of redundancy in the network, and not exceeding installed capacity. They used the LINGO optimization system to find optimal network designs.

An early application of LINDO in telecommunication network design was finding a least cost set of links to use in a network, so, for every pair of nodes, there are at least two paths between them with each path containing at most two links. An interesting feature of solutions to this problem is that, as the number of nodes become large, the optimal design tends to be a two hub network with each node connected directly to each of the two hubs. A generic model of this problem is the model DUPLEX that can be found in the Model Library on this website.

Recently a large telecommunications company used LINGO to look at the problem of how many T1 and T3 lines to install between pairs of nodes and from each node to a single given hub. Two nodes that have a lot of traffic between them will tend to get a direct line; whereas, a node with no exceptionally high traffic to any specific other hub will tend to get only a high capacity connection to the hub. A generic version of a model for this problem appears as model TELEHUB3 in the LINDO model library.

Another generic application is the call or traffic routing problem. Given an existing network structure and demand volumes between every pair of nodes, the traffic routing problem is to select a route for each call pair, so the traffic over each link in the network does not exceed that link's capacity. A generic model for this problem appears in the LINDO library under the name NETUND.

For more information on LINDO and LINGO, please see the product pages. You can also download a demo version from our download page or order a full blown version directly from our order page.