The Matrix Function Toolbox is a MATLAB toolbox connected with functions of matrices. It is associated with the book Functions of Matrices: Theory and Computation and contains implementations of many of the algorithms described in the book. The book is the main documentation for the toolbox.
The toolbox is intended to facilitate understanding of the algorithms through MATLAB experiments, to be useful for research in the subject, and to provide a basis for the development of more sophisticated implementations. The codes are "plain vanilla" versions; they contain the core algorithmic aspects with a minimum of inessential code. In particular, the following features should be noted.
* The codes have little error checking of input arguments.
* The codes do not print intermediate results or the progress of an iteration.
* For the iterative algorithms a convergence tolerance is hard-coded (in function mft_tolerance). For greater flexibility this tolerance could be made an input argument.
* The codes are designed for simplicity and readability rather than maximum efficiency.
* Algorithmic options such as preprocessing are omitted.
* The codes are intended for double precision matrices. Those algorithms in which the parameters can be adapted to the precision have not been written to take advantage of single precision inputs. |