Adding Possibilistic Knowledge To Probabilities Makes Many Problems Algorithmically Decidable

Authors

Olga Kosheleva, The University of Texas at El PasoFollow
Vladik Kreinovich, The University of Texas at El PasoFollow

Publication Date

3-2015

Comments

Technical Report: UTEP-CS-15-13a

To appear in Proceedings of the World Congress of the International Fuzzy Systems Association IFSA'2015, joint with the Annual Conference of the European Society for Fuzzy Logic and Technology EUSFLAT'2015, Gijon, Asturias, Spain, June 30 - July 3, 2015.

Abstract

Many physical theories accurately predict which events are possible and which are not, or -- in situations where probabilistic (e.g., quantum) effects are important -- predict the probabilities of different possible outcomes. At first glance, it may seem that this probabilistic information is all we need. We show, however, that to adequately describe physicists' reasoning, it is important to also take into account additional knowledge -- about what is possible and what is not. We show that this knowledge can be described in terms of possibility theory, and that the presence of this knowledge makes many problems algorithmically decidable.