STRUCTURE ADAPTATION OF STOCHASTIC NEURAL NETS USING LEARNING AUTOMATA TECHNIQUE

The selection of a number of nodes in artificial neural nets containin g stochastic noise perturbations in the outputs and inputs of each nod e is examined. The suggested approach is based on a reinforcement lear ning technique. To solve this optimization problem we introduce a spec ial performance index in such a way that the best number of nodes corr esponds to the minimum point of the suggested criterion. This criterio n presents a linear combination of a residual minimization functional and some generalized variance' of the involved disturbances of random nature. A large value of the noise variance leads to a different optim al number of neurons in a neural networks because of the 'interference ' effect. The optimal point is obtained by the learning procedure base d on the Bush-Mosteller reinforcement scheme. This numerical method is commonly used in Intelligent Control Theory. Simulation modelling res ults are presented to illustrate the effectiveness of the suggested ap proach.