Risk of extreme events via multiobjective decision trees: application to telecommunications

Simultaneous transmission of high-resolution video, digital audio, and data over fiber optic cables is only one of the many new services that the rapidly changing telecommunications industry will soon be providing to users. Currently, federal regulations are restraining telephone companies from entering the video entertainment business, but with the capabilities that technology can provide, their entrance into the market in the long run is inevitable. Telephone companies are facing the decision of whether or not to invest in fiber optic cables in local telephone loops. They also need a way of determining optimal timing for the installation of fiber optic technology in the network. The authors focus on the risks involved in developing an optimal investment plan for expansion of a telecommunications network. This problem is modeled using a multiobjective decision tree where the objective of monetary return from a particular alternative is traded off against the customer satisfaction which that alternative achieves. At each chance node in the decision tree, a continuous probability distribution represents the possible outcomes. To examine the risk of extreme events, the probability distribution is required, rather than only the expected value. To retain all distribution information, a simulation is used to generate a set of Pareto optimal alternatives for various scenarios considered in the model.