10.1 Introduction
Bayesian probabilistic networks (BPN) or Bayesian belief networks (BBN) were developed
during the last two decades, as a decision support tool originally targeted for purposes of
artificial intelligence engineering. Until then artificial intelligence systems were mostly based
on “rule based” systems, which besides many merits also have some problems in dealing with
uncertainties, especially in the context of introducing new knowledge.
Bayesian probabilistic networks are in contrast to the rule based decision support systems so-
called normative expert systems meaning that:
Instead of modelling the expert they model the domain of uncertainty
Instead of using a non-coherent probability calculus tailored for rules, they are based on
classical probability calculus and decision theory
Instead of replacing the expert they support her/him.
The developments of the theory and application areas for Bayesian probabilistic nets have
been and are still evolving rapidly. It is at present possible to utilise the techniques for almost
any aspect of probabilistic modelling and decision making, ranging from inference problems,
model building and data mining over to pre-posterior decision analysis.
In the following some of the most basic aspects of Bayesian probabilistic networks will be
discussed following Jensen (1996). The following text is far from complete and should be
seen as a very first introduction to the concepts of Bayesian nets.
10.2 Causality and Reasoning
Causal networks are graphical representations of causally interrelated events. In the following
some of the most important aspects of causal networks will be illustrated through a number of
examples.
Example 10.1– Reasoning on the quality of concrete structures
Imagine that you are the happy new owner of two concrete bridges erected almost at the same
location and built using concrete produced from a small concrete production plant installed
near the construction site.
After the erection of the structures a routine inspection of one of the structures (bridge 1)
clearly shows that the concrete quality in regard to the durability aspects is far less than
originally prescribed in the design basis for the structure.
The first thought that goes through your mind might be: Bad luck - there is a good chance that
something went wrong with the concrete production – probably also the second bridge has the
same problems.
You call in the site engineers and the responsible for the concrete production and after some
discussions and review of production records two new facts are established. The first being
that all production records for both bridges clearly show that the production was in