An exact optimization method using ZDDs for linear decomposition of symmetric index generation functions

This paper proposes a method using zero-suppressed binary decision diagrams (ZDDs) to find an exact optimum linear decomposition of symmetric index generation functions. The proposed optimization method recursively divides an index set of a symmetric index generation function, based on a branch and bound approach. The method uses ZDDs to represent partitions of an index set compactly and uniquely, and thus, it reuses partial solutions (partitions of an index set) efficiently to prune redundant solution search. In addition, by taking advantages of the symmetry property, the method reduces search space significantly, and can find an optimum solution quickly. Experimental results using benchmark symmetric index generation functions show effectiveness of the proposed method. © 2018, College Publications. All rights reserved.