This class implements the contension inconsistency measure, cf. [Grant, Hunter, 2011].
This measure is defined on paraconsistent models (three-valued models with truth values F,T,B)
of a knowledge base via taking the minimal number of B-valued propositions needed in a model
of the knowledge base.
The computation of the inconsistency value is done as follows. First, the knowledge base
is transformed into CNF. If the knowledge base is consistent the inconsistency value is 0.
Otherwise, for every proposition "p" we create a new knowledge base by removing all clauses
where "p" appears (either positive or negative) and test the new knowledge base for consistency.
This is equivalent to checking whether there is a paraconsistent model that sets the truth value
of p to B. If one of these knowledge bases is consistent we return "1" as inconsistency value. Otherwise,
the process is repeated with all pairs "p1,p2" of propositions, then triples, etc.