Accession Number : ADP008650

Title :   Adaptive Fuzzy Systems,

Corporate Author : UNIVERSITY OF SOUTHERN CALIFORNIA LOS ANGELES SIGNAL AND IMAGE PROCESSING INS T

Personal Author(s) : Kosko, Bart

Report Date : 22 MAY 1992

Pagination or Media Count : 4

Abstract : Fuzziness is multivaluedness. Truth, set membership, and subset containment take on degrees in (0,1) instead of taking on only the limiting bivalent extremes in (0,1). The degrees in (0,1) define fuzzy units, or fits. Statements are true to some degree. An element belongs to, or fits in, a fuzzy set to some degree. One fuzzy set contains another fuzzy set to some degree. Fuzzy systems estimate functions without a mathematical model of how outputs depend on inputs. Mathematically fuzzy systems define mappings between unit hypercubes. These fuzzy associative memories (FAMs) associate output fuzzy-set descriptions. Expert advice and engineering judgement generate fuzzy systems. Sample data generates adaptive fuzzy systems, time-varying mappings between fuzzy cubes.

Descriptors :   *FUZZY SETS, *ADAPTIVE SYSTEMS, *ARTIFICIAL INTELLIGENCE, SYSTEMS ENGINEERING, NEURAL NETS, EXPERT SYSTEMS.

Subject Categories : Theoretical Mathematics
      Cybernetics

Distribution Statement : APPROVED FOR PUBLIC RELEASE