On mathematical models of adaptive systems
Hirotomo Aso, Masayuki Kimura · International Journal of Systems Science · 1974
The aim of this paper is to present a mathematical model for a system to control an object favourably which is unknown for the system—an adaptive system. We make the concept of adaptiveness precise first and consider systems as stochastic sequential machines (SSM's) to answer the question : what kind of logical structure is necessary and sufficient for a SSM to be adaptive. Some necessary conditions for a SSM to be adaptive are shown, and for a specified class of SSM's, the necessary and sufficient conditions to be adaptive are given. Furthermore, some examples of adaptive SSM's are shown.