Modeling the reachable sets for positive linear systems using self-regulating adaptive perceptron type neural networks

The paper presents a technique for modeling reachable states of positive linear discrete-time systems (PLDS) using static feed-forward neural networks. The proposed method is based on design of self-regulating two layer perceptron type neural network for the modeling of reachable sets of PLDS systems represented by polyhedral cones using a pattern recognition method.