Article ID: | iaor2001487 |
Country: | United States |
Volume: | 30 |
Issue: | 1 |
Start Page Number: | 26 |
End Page Number: | 44 |
Publication Date: | Jan 2000 |
Journal: | Interfaces |
Authors: | Ambs Ken, Cwilich Sebastian, Deng Mei, Houck David J., Lynch David F., Yan Dicky |
Keywords: | communication |
To ensure high network reliability, AT&T employs two basic approaches: preventing failures and responding quickly when failures occur. For AT&T to quickly reroute traffic in the event of a network failure, the network must contain sufficient restoration capacity to carry the displaced demand. A team of AT&T OR experts, network planners, and managers developed a method for determining the appropriate quantity and location of restoration capacity required to restore the demand during any single link failure. The approach centers on a linear programming model to minimize the cost of the restoration network and uses column generation to generate new restoration paths as needed. In about 10 months, the team converted the methodology into a tool to optimize the allocation of restoration capacity. This tool was then extended to plan for the recovery of a switching-center disaster and to reoptimize the entire restoration network. It has contributed to AT&T's achieving high-quality service, while saving valuable resources. It resulted in hundreds of millions of dollars in cost savings and increased revenues.