作者: Lofti Zadeh
关键词: Local consistency 、 Predicate logic 、 Fuzzy set 、 Theoretical computer science 、 Fuzzy number 、 Artificial intelligence 、 Computing with words and perceptions 、 Rule of inference 、 Computer science 、 Natural language 、 Constraint satisfaction problem
摘要: Discusses a methodology for reasoning and computing with perceptions rather than measurements. An outline of such methodology-referred to as computational theory is presented in this paper. The perceptions, or CTP short, based on the CW. In CTP, words play role labels and, more generally, are expressed propositions natural language. CW-based techniques employed translate language into what called Generalized Constraint Language (GCL). language, meaning proposition generalized constraint, N R, where constrained variable, R constraining relation isr variable copula which r whose value defines way constrains S. Among basic types constraints are: possibilistic, veristic, probabilistic, random set, Pawlak fuzzy graph usuality. wide variety GCL makes much expressive predicate logic. CW, initial terminal data sets, IDS TDS, assumed consist These translated, respectively, antecedent consequent constraints. Consequent derived from through use rules constraint propagation. principal propagation rule extension principle. retranslated yielding set (TDS). CW coincide inference A problem that explicitation N, X represents proposition, p,