High level design reuse through fuzzy learning

Object-oriented design has made possible the reuse of individual components during computer system synthesis. We demonstrate how the degree of reuse can be increased to a larger extent through a high level reuse of complete subsystems by the application of machine learning techniques and fuzzy logic in an object-oriented system-level synthesis environment. The fuzzy learning techniques were implemented in the ICOS methodology. It is shown by experiments that efficiency of synthesis improved by a factor of two to three after fuzzy learning was applied.