×

Embryonic genetic algorithm with random generational growing strategy for optimizing variable ordering of BDDS. (English) Zbl 1265.90345

Summary: This paper addresses the problem of optimizing the variable ordering in binary decision diagrams (BDDs). A new hybrid embryonic genetic algorithm is proposed for optimizing the variable ordering that combines a branch-and-bound technique with the basic genetic algorithm. It uses fitness based on a lower bound and embryos instead of full chromosomes. A novel growing technique introduces two new growing operators. The results of an experimental evaluation illustrate the efficiency of the approach.

MSC:

90C59 Approximation methods and heuristics in mathematical programming
68T20 Problem solving in the context of artificial intelligence (heuristics, search strategies, etc.)
68W20 Randomized algorithms
PDFBibTeX XMLCite