Application of Genetic Algorithms to Function Decomposition in Pattern Theory.
Abstract
This report documents use of genetic algorithms for finding partitions which lead to optimal decomposition of boolean functions in the Ashenhurst-Curtis method of functional decomposition. This problem apparently grows exponentially as the number of input variables increase, but is useful to study since it has a myriad of potential applications in algorithm design, circuit design, image processing, data compression, logic minimization, and machine learning. The report presents some background on function decomposition, genetic algorithms and results of some experiments. Although use of genetic algorithms still result in exponential growth; they provide a much lower rate of growth.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 26, 1994
- Accession Number
- ADA327931
Entities
People
- David A. Gadd
- Mark Axtell
- Michael J. Noviskey
- Timothy D. Ross