Abstract:
Decomposition is an important operation for data structures representing switching functions. Decomposition algorithms have been developed for many data structures representing either a single function or incompletely specified functions (ISF). EXOR-decomposition of ISF leads to sets of subfunctions which cannot be described by ISF. Conventional methods of decomposition consider only a small fraction of these sets, usually described by an ISF. This paper introduces LC-ISF (linearly combinated ISF) which can describe the set of functions produced by the decomposition method EXOR-grouping for ISF. There are new possibilities for optimization because LC-ISF provide a larger set of functions than ISF. We also show an EXOR-grouping algorithm for LC-ISF, and show that the resulting set of subfunctions can be described by LC-ISF.