A fast algorithm for estimating Lyapunov exponents from time series

摘要

We propose a fast algorithm for estimating Lyapunov exponents from time series. We divide the phase space into boxes, which can move to adapt to the geometry of the reconstructed attractor. Lyapunov exponents are calculated from the average Jacobians estimated for each box. Results for known maps and fluxes are in agreement with values calculated through usual time average methods. The application of the procedure requires a small number of adjustable parameters. (C) 1998 Elsevier Science B.V. [References: 21]