Propagation of Preference Relations in Qualitative Inference Networks

S. K. Michael Wong, Pawan Lingras, Yiyu Yao

Abstract: Preference relations can provide a more realistic model of random phenomena than quantita­ tive probability or belief functions. In order to use preference relations for reasoning under un­ certainty, it is necessary to perform sequential and parallel combinations of propagated infor­ mation in a qualitative inference network. This paper discusses the rules for such sequential and parallel combinations.