PLA decomposition with generalized decoders
S. Yang, Maciej J. Ciesielski · 2003
A method for PLA (programmable logic array) decomposition is presented whereby a single two-level Boolean function (PLA) is decomposed into two levels of cascaded PLAs. Primary inputs are partitioned into disjoint sets and represented as multiple-valued variables so as to minimize the total PLA area. Two new algorithms are presented for input variable assignment, based on (a) graph partitioning and (b) a mathematical optimization method. The constrained output encoding of first-level PLAs is performed by a fast and efficient procedure based on a graph-coloring technique.>