Date Added: Aug 2012
This paper proposes a new methodology for smart classification and retrieval of software mechanism based on user - defined necessities. The classification proposal utilizes a dedicated genetic algorithm which evolves a tiny number of classifiers by dividing the position of available components stored in a database into positive subsets (clusters). Each classifier consequently becomes the leader-representative of its cluster. When a customer wishes to trace a component he/she identifies the preferred characteristics (component profile) which are then compared with the distinctiveness of the available classifiers.