Return to search


The first purpose of this thesis is to provide a framework for
the inclusion of data from a secondary source in Bayesian decision
analysis as an aid in decision making under uncertainty. A second purpose
is to show that the Bayesian procedures can be implemented on a
computer to obtain accurate results at little expense in computing time.
The state variables of a bridge design example problem are the
unknown parameters of the probability distribution of the primary data.
The primary source is the annual peak flow data for the stream being
spanned. Information pertinent to the choice of bridge design is contained
in rainfall data from gauges on the watershed but the distribution
of this secondary data cannot be directly expressed in terms of
the state variables. This study shows that a linear regression equation
relating the primary and secondary data provides a means of using
secondary data for finding the Bayes risk and expected opportunity loss
associated with any particular bridge design and single new rainfall
The numerical results for the example problem indicate that the
information gained from the rainfall data reduces the Bayes risk and
expected opportunity loss and allows for a more economical structural
design. Furthermore, the careful choice of the numerical methods employed
reduces the computation time for these quantities to a level
acceptable to any budget.
Date10 1900
CreatorsGray, Howard Axtell
ContributorsDepartment of Hydrology & Water Resources, The University of Arizona
PublisherDepartment of Hydrology and Water Resources, University of Arizona (Tucson, AZ)
Source SetsUniversity of Arizona
Detected LanguageEnglish
Typetext, Technical Report
SourceProvided by the Department of Hydrology and Water Resources.
RightsCopyright © Arizona Board of Regents
RelationTechnical Reports on Hydrology and Water Resources, No. 14

Page generated in 0.0017 seconds