Optimizing Variable Ordering of BDDs with Double Hybridized Embryonic Genetic Algorithm

Abstract

This paper presents a new double hybridized genetic algorithm for optimizing the variable order in Reduced Ordered Binary Decision Diagrams. The first hybridization adopts embryonic chromosomes as prefixes of variable orders instead of complete variable orders and combines a branch & bound technique with the basic genetic algorithm. The second hybridization is done with the existing sifting algorithm, known as one of the most effective heuristic for this problem, which is incorporated as a hyper mutation operator.