FUZZY INFERENCE SYSTEM OR ADAPTIVE NEURO-FUZZY INFERENCE SYSTEM, AND INTELLIGENT AGENT FOR THE DYNAMIC GENERATION AND RETRIEVAL OF USER INTERFACE SOFTWARE MODULES

摘要

A method of operating a target systemwherein small changes in input variables producesmall changes in output variables in a mannerpermits system learning on a time dependentbasis. The rate of change of the system output isdirectly dependent upon the product of the rate ofchange of the system input and a matrix consistingof input variables and antecedent and consequentparameters. A preferred performance criterionis obtained through the approximation of the saidmatrix to a weighted-augmented pseudo-inverseJacobian. Off-line, the system undergoes a seriesof iterations using a wide range of input variableswherein the actual outputs are compared to thedesired outputs and optimized values for theantecedent and consequent parameters are obtainedand passed back to the said matrix for use in thesubsequent iteration.