Symbolic dynamics analysis of nonstationary data from a model of a magnetic system with solitons

The probability distribution of a complexity measure is used to characterize chaotic states: an estimator of the algorithmic complexity of a time series of symbolic words is calculated within a fixed length time window, which sweeps through the time series analyzed. The words are derived through a symbolic dynamics scheme applied in an m-dimensional delay coordinate space. Time:intervals instead of the variables of the system are used. The chaotic states of a model of a magnetic domain wall are characterized better by the methods presented than with the use of fractal dimensions and new-intermittent states of the system were easily identified. Using an artificial nonstationary signal composed of different chaotic states of the Bloch wall as a test for chaos-chaos intermittency we demonstrate that the method developed is suitable for the detection and characterization of intermittency. It is also shown that nonstationarity in the form of a slow monotonic drift in the control parameter may extend the stability range of periodic states of the spatially extended system studied-a trackinglike phenomenon. [S1063-651X(99)11510-1].