Next: Outline analysis of complexity
Up: Inference
Previous: From model space to
aDCSP + preferences = aDPCSP
The aDCSP produced as above formalises the hard requirements imposed
upon the scenario models. Among the scenario models that meet these
requirements, some may be better than others, because the underlying
model design decisions may be deemed more appropriate by the user.
Preferences that express this (relative) level of appropriateness are
attached to the assumptions that describe the model design decisions,
and by extension, to the attribute-value pairs in the aDCSP. As
discussed in section 2, such an extension to the
aDCSP constitutes an aDPCSP.
More specifically, it is worth recalling that in section
2.2 an order-of-magnitude preference calculus is presented
that enables representation and reasoning with subjective user
preferences for different relevance and modelling assumption. Next,
section 2.3 introduces a solution algorithm for
aDPCSPs that include an aDCSP, such as the ones constructed with the
approach of section 3.3.2, and are extended
with subjective user preferences for alternative design decisions.
Jeroen Keppens
2004-03-01