Sensitivity Analysis, Observing Effect Sensitivity Analysis, Observing Effect 
To observe the effect of the change in the model, we open a probe on the node "Lpci-Sys". The probe (Figure~2) shows the probability that the system works at two points in time. The base probability is the probability before any change (i.e., just after we compiled the model.) The active probability is the probability after the change. Note that the analyst can set the base and active to any two states, not just the initial and current state.
In this case, the change was trivial and we could have predicted it. Adding a 1% chance of system failure even if the subsystem were working added 1% to the chance of system failure. The increase in failure probability is linear in the probability that the system fails according to some unregistered mode.
Fortunately, Graphical-Belief is object-oriented. The four pump train rules ("Pta-Or", "Ptb-Or", "Ptc-Or" and "Ptd-Or") are all instances of the template "Pt-Rule". The template or prototype rules are shown to the left of the graph window in Figure~1 in the rule window. We can edit the (:support-if-false :true) parameter of the template and this will change all four instances. Figure 3 shows the result of this change.
Figure 3. Prob showing change in system failure after modifying
template parameter ("Pt-Rule" :support-if-false :true).
History Continue with this
example and see how to use the probes and the history mechanism to see
the effect of individual changes.
Return to
the main example page.
Back to overview of Graphical-Belief.
View a list
of Graphical-Belief in publications and downloadable technical
reports.
The Graphical-Belief user
interface is implemented in Garnet.
Get more
information about obtaining Graphical-Belief (and why
it is not generally available).
get
the home page for Russell Almond , author
of Graphical-Belief.
Click
here to get to the home page for Insightful (the company that StatSci
has eventually evolved into).