Genetische Algorithmen
| project staff | project description | project structure | publications
|
| Chair of Computer Architecture | |
| Frank Schmiedle, Dr.-Ing. | |
| Rolf Drechsler, Prof. Dr. | |
| Nicole Drechsler, Dr. | |
In VLSI CAD often difficult optimization problems have to be solved. One method, which has been analysed and used with growing interest in the past years are genetic algorithms (GA). This method of optimization is based on the theory of evolution: characterisations of solutions for problems are encoded and generate so-called individuals, which are rated with the help of a fitnessfunction. Through reiterated selection, recombination and elimination new solutions are created constantly and the aim is, that these solutions get better from generation to generation concerning their fitness. Genetic algorithms resp. evolutionary strageties often are able to find better solutions than other optimizationmethods. One disadvantage of this method however are the sometimes very long running-times.
Genetic Algorithm Managing Environment
GAME is a C++ softwarepackage, which provides the framework for genetic algorithms including all necessary modules e.g. selection or recombination on the one hand, but also provides a standardised interface and the possibility to integrate user-defined modules on the other hand. With help of this concept GAME can be adapted to any particular application and can be used multi-purposely in the field of genetic algorithms.
Applications in CAD
At the Chair of Computerarchitecture genetic algorithms were and still are used successfully in different subareas of the VLSI CAD, to be exact in logic synthesis, physical synthesis and also in the field of testing. In logic synthesis for example AND/EXOR-Minimizations and multilayered circuit-minimizations were examined this way. In physical synthesis GAs were used on the fields of floorplanning and routing. Finally in testing evolutionary methods were used for test set generation and integrated self-tests.
|
Christian Matuszweski, Frank Schmiedle, Dr.-Ing., Robby Schönfeld Routing with Genetic Algorithms Albert-Ludwigs-University Freiburg, Technical Report 125, 1999 |
|
Nicole Drechsler, Dr., Rolf Drechsler, Prof. Dr., Bernd Becker, Prof. Dr. A New Model for Multi-Objective Optimization in Evolutionary Algorithms International Conference on Computational Intelligence (Fuzzy Days), 1999 |
|
Martin Keim, Dr., Nicole Drechsler, Dr., Bernd Becker, Prof. Dr. Combining GAs and Symbolic Methods for High Quality Tests of Sequential Circuits ASP Design Automation Conference, 1999 |
|
Nicole Drechsler, Dr., Rolf Drechsler, Prof. Dr., Bernd Becker, Prof. Dr. GAME: A Software Environment for Using Genetic Algorithms in Circuit Design International Conference on Applied Computer Systems, 1997 |
|
C. Ökmen, Martin Keim, Dr., R. Krieger, Bernd Becker, Prof. Dr. On Optimizing BIST Architecture by Using OBDD-based Approaches and Genetic Algorithms VLSI Test Symposium, 1997 |
|
Rolf Drechsler, Prof. Dr., Bernd Becker, Prof. Dr. Learning Heuristics by Genetic Algorithms ASP Design Automation Conference, 1995 |
|
Bernd Becker, Prof. Dr., Rolf Drechsler, Prof. Dr. OFDD based Minimization of Fixed Polarity Reed-Muller Expressions Using Hybrid Genetic Algorithms International Conference on Computer Design, 1994 |
|
Rolf Drechsler, Prof. Dr. Evolutionary Algorithms for VLSI CAD Kluwer Academic Publishers, 1998 |
|
Nicole Drechsler, Dr., Frank Schmiedle, Dr.-Ing., Daniel Große, Dipl.-Inf., Rolf Drechsler, Prof. Dr. Heuristic Learning based on Genetic Programming EuroGP, volume 2038 of LNCS, pp.1-10, Springer Verlag, 2001 |
|
Frank Schmiedle, Dr.-Ing., Nicole Drechsler, Dr., Daniel Große, Dipl.-Inf., Rolf Drechsler, Prof. Dr. Priorities in Multi-Objective Optimization for Genetic Programming Genetic and Evolutionary Computation Conference, pp.129-136, 2001 |
|
Frank Schmiedle, Dr.-Ing., Daniel Große, Dipl.-Inf., Rolf Drechsler, Prof. Dr., Bernd Becker, Prof. Dr. Too Much Knowledge Hurts: Acceleration of Genetic Programs for Learning Heuristics Int'l Conference on Computational Intelligence (Fuzzy Days), volume 2206 of LNCS, pp.479-491 |
|
Frank Schmiedle, Dr.-Ing., Nicole Drechsler, Dr., Daniel Große, Dipl.-Inf., Rolf Drechsler, Prof. Dr. Heuristic Learning based on Genetic Programming Genetic Programming and Evolvable Machines, 3(4):363-388 |
|
Frank Schmiedle, Dr.-Ing. Exact Routing with symbolic methods PhD thesis, Logos Verlag, 2003 |