Mastering the Processing of Preferences by Using Symbolic Priorities in Possibilistic Logic

摘要: The paper proposes a new approach to the handling of preferences expressed in compact way under form conditional statements. These statements are translated into classical logic formulas associated with symbolic levels. Ranking two alternatives then leads compare their respective amount violation respect set expressing preferences. amounts, which can be computed possibilistic manner, partially ordered lexicographically once put vector form. This is compared ceteris paribus-based CP-net approach, main existing artificial intelligence processing It shown that partial order obtained fully agrees one proposed but generally includes further strict between (considered as being not comparable by level logic-based approach). additional fact debatable, since they reflection explicit user's result application paribus principle implicitly, and quite arbitrarily, favors father node graphical structure Adding constraints levels for nodes less allowed than children nodes, it possible recover CP-net-induced order. Due results levels, computationally tractable. Key words: preference, priority, order, CP-net, logic.