Reinforcement learning neural network used in control of nonlinear systems

A new category of methods used in managing the problems that appears in systems control is inspired from intelligent computation area. In this paper is presented a method of designing a controller for nonlinear systems based on a recurrent neural network which is training in real time using a reinforcement learning (RL) procedure. The advantage of this method is by-passing of the difficulties implied by the direct solving of the differential models, which are necessary in a classical approach. Moreover this new technique using a real-time training is better then the MLP network controller and also than the RBF network implementation which both need a preliminary training process, based on a set of input-output data that has to be a-priori experimentally determined.