Improving Reliability in a Stochastic Communication Network
Abstract
This research investigated the behavior of increasing the reliability of components in stochastic communication networks. A stochastic communication network is one in which the components, links and nodes, are not 100 percent reliable. In addition the reliability value, each component also has a capacity value that limits the flow through the component. The throughput in a stochastic communication network can be improved by increasing the capacity or the reliability of the individual components. A set of linear and nonlinear models was developed to determine the best investment strategy for increasing the reliability of individual components. Once the models were developed, an experiment was designed to compare the output of the models to another set of existing models that calculated the best investment strategy for increasing the capacity of individual components. The experiment employed a Prolog program to: enumerate the paths, compute the reliability of each path, and formulate the linear and nonlinear models for direct input into mathematical programming packages. In addition, the linear models were reformulated as networks with side constraints and solved using a network flow package. The models and methodology are general and could easily be modified.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1990
- Accession Number
- ADA230442
Entities
People
- William L. Gaught
Organizations
- Air Force Institute of Technology