Flexible optimization of fixed polarity Reed-Muller expansions for multiple output completely and incompletely specified Boolean functions

A new algorithm that generates optimal fixed polarity Reed-Muller expansions based on user specified optimization criteria is shown. The algorithm accepts reduced representation of Boolean functions in form of an array of cubes and operates on an Algebraic Ternary Decision Tree together with lookup tables of flexible sizes. Allocation of don't care minterms is performed in an non exhaustive way by a heuristic approach based on the properties of Reed-Muller expansions.