Document Type
Thesis - University Access Only
Award Date
1996
Degree Name
Master of Science (MS)
Department / School
Civil and Environmental Engineering
Abstract
The field of network design is a growing one. Sampling sites, sampling pattern and approaches are taken into consideration while designing a network. But sometimes modification of an existing network becomes essential, and requires addition or removal of some gages. In the present study, an optimal way is identified to solve this problem. Network optimization is an optimization problem. A Variance-based approach in which variance of estimation error i.e. , expected value of the squared difference between the true and the estimated value of a random variable, e.g., rainfall, is considered as an improving criteria. A conventional algorithm, Newton's method, is applied to find the optimal location for a new site. Results show that this algorithm fails to find the global optimum. An optimal search algorithm called branch and bound with underestimates, has been successfully applied here to find the optimal gages for removal. Overall, the proposed network design approach will provide a guideline to the future network designers.
Library of Congress Subject Headings
Network analysis (Planning) Mathematical optimization Analysis of variance Rain and rainfall -- Measurement
Format
application/pdf
Number of Pages
89
Publisher
South Dakota State University
Recommended Citation
Moktader, MD. Abdul, "Network Optimization by Variance Reduction Analysis" (1996). Electronic Theses and Dissertations. 535.
https://openprairie.sdstate.edu/etd2/535