On nonlinear reinforcement schemes

This paper deals with the analysis of nonlinear reinforcement schemes for learning automata. The learning automaton is connected in feedback loop to a random environment. The correction term of the action probability vector depends on a nonlinear function /spl phi/(x). Results concerning the convergence, the convergence rate, and the effect of the function /spl phi/(x) are stated. A comparison between the convergence rate of nonlinear and linear reinforcement schemes is presented.