Computational study and comparisons of LFT reducibility methods

A set of computationally tractable model reduction algorithms are described which may be used to determine minimal realization dimensions for uncertain systems represented by linear fractional transformations on structured uncertainty sets; these computational tools are also applicable to multi-dimensional systems. The methods described utilize linear matrix inequality methods, in addition to straightforward coordinate transformations and truncations. These algorithms are evaluated on a variety of example systems that are constructed to be reducible.