 |
    |
Learning the Ideal Evaluation Function
Edwin D. de Jong* and Jordan B. Pollack
DEMO Lab,
Volen National Center for Complex Systems,
Brandeis University MS018,
415 South street,
Waltham MA 02454-9110, USA
{edwin,pollack}@cs.brandeis.edu,
http://demo.cs.brandeis.edu
Abstract.
Designing an adequate fitness function requires substantial
knowledge of a problem and of features that indicate progress
towards a solution. Coevolution takes the human out of the loop by
dynamically constructing the evaluation function based on
interactions between evolving individuals. A question is to what
extent such automatic evaluation can be adequate. We define the
notion of an ideal evaluation function. It is shown that
coevolution can in principle achieve ideal evaluation. Moreover,
progress towards ideal evaluation can be measured. This
observation leads to an algorithm for coevolution. The algorithm
makes stable progress on several challenging abstract test
problems.
Keywords: Coevolution, Pareto-Coevolution, Complete Evaluation Set, ideal evaluation, underlying objectives, Pareto-hillclimber, over-specialization
*Current address: DSS Group, Utrecht University. dejong@cs.uu.nl
LNCS 2723, p. 274 ff.
lncs@springer.de
© Springer-Verlag Berlin Heidelberg 2003