Permutation schemes

The extendable family of routines for the generation of annealed surrogate data can accomodate different permutation schemes. In this version, the most important module implements the exchange of randomly chosen pairs. An (yet unimplemented) alternative is, at lower temperatures, to select pairs of comparable magnitudes in order to increase the acceptance rate.

For the particulat purpose of generating series of inter-event intervals, it may be computationally better to exchange two adjacent intervals only, since in that case only one event time changes.

Random pairs

randomize_cost_cool_random [-X xfile] [general options] [cost function options] [cooling options] file

-X xfile file containing list of indices excluded from permutation

At each annealing step, one pair of values is exchanged. Here, two disjoint indices are selected at random. It is possible to specify a list of indices which shall be excluded from permutations. This can be useful if the data has artefacts (e.g. zeroed gaps) and if the boundary points are to be kept fixed. The file xfile is expected to contain one (integer) index per line.

Adjacent intervals

randomize_cost_cool_event[general options] [cost function options] [cooling options] file

At each annealing step, one adjacent pair of inter-event intervals is exchanged. That is, one event time selected at random is changed. Note that the first and last events are always ket fixed.

constrained randomization * Table of Contents * TISEAN home