Improved Functional Identification Using Genetic Simulation

摘要

Identification is used for logical design, test or sequential learning. It consists of deriving an automaton model of a given system from an external description of its behavior. We present a new identification approach based on genetic simulation. The first section offers a unified classification of the various identification methods according to three basic criteria. In the second section, we stress the special case of functional identification. In the third one, the potential and interest of genetic simulation for identification are analyzed, and a new genetic approach is introduced. We describe a computational experiment that validates our approach.