The polyphase representation with respect to sampling lattices in mult
idimensional (M-D) multirate signal processing allows us to identify p
erfect reconstruction (PR) filter banks with unimodular Laurent polyno
mial matrices, and various problems in the design and analysis of inve
rtible MD multirate systems can be algebraically formulated with the a
id of this representation. While the resulting algebraic problems can
be solved in one dimension (1-D) by the Euclidean Division Algorithm,
we show that Grobner bases offers an effective solution to them in the
M-D case.