Grammatical inference as a tool for constructing self-learning syntactic pattern recognition-based agents

Abstract

Syntactic pattern recognition-based agents have been proven to be a useful tool for constructing real-time process control intelligent systems. In the paper the problem of self-learning schemes in the agents is discussed. Learning capabilities are very important if practical applications of the agents are considered, since the agents should be able to accumulate knowledge about the environment and flexible react to the changes in the environment. As it is shown in the paper, the learning scheme in the agents can be based on a suitable grammatical inference algorithm. © 2008 Springer-Verlag Berlin Heidelberg.