Extraction and reuse of design patterns from genetic algorithms using case-based reasoning

In this paper, we propose a case-based reasoning scheme in which we extract domain knowledge (in the form of design patterns) from a genetic algorithm used to optimize combinational logic circuits at the gate level. Such information is used in two ways: first, we show how the selection pressure of the genetic algorithm is biased by Boolean simplification rules that are normally adopted by human designers, including some which are not completely straightforward. Secondly, we reuse some of these design patterns extracted from the evolutionary process to reduce convergence times of a genetic algorithm using previously found solutions as cases to solve similar problems.