Evolutionary Algorithms in Computer-Aided Design of Integrated Circuits
Rolf Drechsler, Nicole Drechsler, Bernd Becker, Henrik Esbensen · World Scientific series in robotics and intelligent systems · 1999
An increasing number of successful applications of Evolutionary Algorithms (EAs) within Computer Aided Design (CAD) of Integrated Circuits (ICs) is reported in the literature. The problems dealt with in this field consist of sequences of sub-problems, which are characterized by being NP-hard, large and mutually dependent. This very high level of complexity should make the EA a well suited approach, at least in principle. However, the EA is of practical interest to the CAD community if and only if it is competitive to the existing approaches with respect to performance. With this fact as the starting point, the purpose of this paper is to discuss how to develop high-performance EAs for CAD of ICs. After reviewing a number of recent, performance competitive EAs for key problems in CAD of ICs, the common characteristics of these algorithms are discussed. They all exploit problem specific knowledge in various ways or are integrated with other heuristics. The paper also discusses performanc...