Article Title

Representing dynamics in the eccentric Hill system using a neural network architecture


periapse, Poincaré map, artificial neural networks, eccentric Hill system


This paper demonstrates how artificial neural networks can be used to alleviate commonproblems encountered when creating a large database of Poincaré map responses. A generalarchitecture is developed using a combination of regression and classification feedforwardneural networks. This allows one to predict the response of the Poincaré map, as well asto identify anomalies, such as impact or escape. Furthermore, this paper demonstrateshow an artificial neural network can be used to predict the error between a more complexand a simpler dynamical system. As an example application, the developed architecture isimplemented on the Sun–Mars eccentric Hill system. Error statistics of the entire architectureare computed for both one Poincaré map and for iterated maps. The neural networks arethen applied to study the long-term impact and escape stability of trajectories in this system.


Tsinghua University Press