ISSN:
1432-0770
Source:
Springer Online Journal Archives 1860-2000
Topics:
Biology
,
Computer Science
,
Physics
Notes:
Summary An algorithm is given for an adaptive and learning system which controls any given unknown objective system while learning its stochastic behaviour by observing its reactions to control inputs. The algorithm consists of two subalgorithms; one for estimating the stochastic transition structure of the objective system with regard to some kinds of a priori information, and the other for determining the optimal control input to each state of the objective system on the basis of its estimated transition structure. This combined algorithm has shown to serve as an adaptive and learning system. This paper concerns with the decision algorithm of control inputs, and the subsequent with the estimation algorithm of transition structure.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1007/BF00289407
Permalink