In this book we have studied learning automata which are adaptive decision making devices operating in unknown random environments. They learn by updating their action probability distributions in accordance with the reinforcement obtained from the environment. In Chapter 1 we have considered a variety of LA to suit different situations. Chapters 2 and 3 described how collections of LA could be formed into teams and networks to solve more complex learning problems. One example of an application, namely, pattern recognition was considered in detail in Chapter 4. The problem of speed of learning, which becomes crucial with larger assemblage of LA, was handled in Chapter 5 by forming modules which enable parallel operation.
Unable to display preview. Download preview PDF.