Two algorithms are presented, the first is a technique to determine go
od, though not necessarily optimum. fixed polarity Reed-Muller expansi
ons of completely specified boolean functions. The second algorithm de
termines the allocations of the 'don't care' terms of incompletely spe
cified boolean functions resulting in optimum positive polarity Reed-M
uller expansions. Additionally, investigations are made into combining
these two techniques to determine fixed polarity Reed-Muller expansio
ns of incompletely specified boolean functions. Results are presented
which show the effectiveness of the techniques and comparisons are mad
e with existing methods.