Dear SIPTA list,

After having lain dormant, research on the CONEstrip algorithms [*] has awakened again. These algorithms are essentially meant for checking coherence of general sets of desirable gambles. This is very closely related to checking coherence of imprecise or conditional probability assessments and probabilistic satisfiability (PSAT). The classical CONEstrip algorithm assumes gambles are directly specified (as vectors); this corresponds to having an explicit description of the possibility space. The propositional CONEstrip algorithm assumes gambles are expressed as linear combinations of propositional sentences; this corresponds to specifying (imprecise/conditional) probabilities for events and relationships between events, but leaving the underlying possibility space implicit.

A collaborator is working with my support on creating implementations and evaluating these. For the evaluation, we want to use realistic test cases. These are nontrivial to construct, certainly for the propositional context. Therefore, we are looking for such test cases. So I would like to ask you to send me any information you may have about existing test cases of this kind (or ones that are ‘close enough’). This can be in the form of pointers to papers with a description of one or more problems that may function as test cases or as pointers to files containing test cases encoded in some documented format.

I am grateful for any input.

Best regards,

Erik

[*] https://ac.erikquaeghebeur.name/?search=conestrip

-- https://ac.erikquaeghebeur.name