Environmental Engineering Reference
In-Depth Information
assessing the value of information
to design site investigation and
construction quality assurance programs
Robert B. Gilbert and Mahdi Habibi
13.1 IntroDuCtIon
The potential to improve the types and quantities of data that are collected in the practice of
geotechnical engineering is enormous. There are many situations where additional informa-
tion would be valuable in developing designs and making decisions. Likewise, there are also
many situations where data that are obtained have little or no value in developing designs
or making decisions.
The goal of this chapter is to describe a framework and provide practical tools and
insights for assessing the value of information to design site investigation and construction
quality assurance programs. A decision framework for assessing the value of information
is presented, the role and practical application of Bayes' Theorem is explored, and illustra-
tive examples and case histories are provided throughout to demonstrate and elucidate the
theor y.
13.2 Value oF InForMatIon FraMeWork
Assessing the value of information involves considering how the information might be used
to guide decision-making. This section describes the basic framework for decision analysis
and provides illustrative examples of applying this framework.
13.2.1 Decision analysis
The generic decision tree in Figure 13.1 shows the role of information in making a decision;
see Benjamin and Cornell (1970), Ang and Tang (1984), and Gilbert et al. (2008) for details
on decision trees. The decision here is between two alternatives, Plan A versus Plan B. The
uncertainty in the decision is the consequence that will be realized when one of the plans
is implemented, which is represented by a probability distribution, p C ( C ), for the differ-
ent possibilities. The preferred alternative is that with the greatest expected consequence, *
EC
()
=
cpc
C
×
()
where the expected consequence once the decision is made is the
allc
maximum of the possible expected consequence values
EC forDecision
(
)
=
maxEC
[(
), (
C
)]
(13.1)
AlternativeA
AlternativeB
* Consequence is used here assuming that larger values are preferred to smaller values, such as net profit with
positive values being a gain and negative values being a loss or nondimensional utility with a range between the
least—preferred consequence of zero and the most preferred consequence of one.
491
 
Search WWH ::




Custom Search