Research Professor of Mathematical and Evolutionary Computer Science, with the Interactive Systems Engineering Research Group and the Adaptive Systems Research Group at the University of Hertfordshire.
Previously Professor in the Cybernetics and Software Systems Group at the University of Aizu in northern Japan and a frequent Visiting Professor at Institute for Mathematics and Informatics at the University of Debrecen in Hungary, Chrystopher Nehaniv is also Director of the UK EPSRC Network on Evolvability in Biological and Software Systems, Associate Editor of the Elsevier journal BIOSYSTEMS.
He will address EuroGP2002 on "Open-Ended Evolution, Evolvability and Embodiment in Biology and Other Sciences of the Artificial"
Professor at Computer Science Department at the University of Dortmund. Author of 7 monographs and more than 100 scientific papers, university medal for excellent teaching (1994), elected referee of DFG (19922000), head of this committee (19962000), responsible for the German youth competition in computer science (19952001), elected member of the Academy of Sciences (since 2001); research interests: complexity theory and efficient algorithms.
The subject for his talk at EuroGP2002 is "How to apply methods from complexity theory and classical algorithm analysis to evolutionary algorithms"
Although evolutionary algorithms are randomised algorithms, people from
the EA community usually do not apply methods and results from classical
algorithm analysis or from complexity theory. It is the purpose of this talk
to discuss why the main questions in the analysis of classical randomised al
gorithms, the estimation of the expected optimisation time and the success
probability, are at least as important as the typical EA theories. Moreover,
the usefulness of methods like Chernoff bounds, potential functions, coupon
collector's theorem, delay sequence arguments, and the estimation of the pro
bability that an EA run is not typical is explained by applying the methods
to typical EA problems.