On a Memory-Based Realization of Sparse Multiple-Valued Functions

Abstract

This paper presents multi-valued (MV) functions, which are generalizations of index generation functions and switching functions. First, an efficient memory-based realization of sparse MV functions, where the number of specified combinations is much smaller than the number of possible input combinations, is presented. Then, a formula for the expected number of variables to represent random sparse MV functions is derived. Finally, the theoretical analysis is compared with the experimental results.