Biconditional-BDD Ordering for Autosymmetric Functions

Autosymmetric functions are particular ``regular'' Boolean functions that are exploited for logic optimization, since it is possible to reduce the number of variables and the number of points of the original autosymmetric function before its synthesis. In this paper we study this regularity in oder to derive a suitable variable ordering for Biconditional Binary Decision Diagrams (BBDDs). BBDDs are a new version of BDD that have EXOR of two variables (instead of a variable) in the nodes. These diagrams are employed for logic synthesis in new technologies such as silicon nanowires and DG-SiNWFETs. We show that it is possible to find a useful variable ordering for these functions and the experimental results validate our approach showing that in the 97% of the cases we get an ordering that gives a number of nodes that is lower or equal to the one obtained with the standard ordering.