Controlling chaos with an artificial neural network

A novel method for controlling chaos is proposed by resorting to an artificial neural network (ANN). It is widely applicable where the behavior of a chaotic system is well described by the next return map. The method appears to be superior to other alternatives since a single ANN trained with time-series data is capable of controlling chaos, i.e., rendering stable multiple unstable periodic orbits embedding a strange attractor, and extensive computation or analysis is unnecessary. Its efficacy is illustrated with an example of a logistic map