 |
    |
Tracking Differential Evolution Algorithms: An Adaptive Approach through Multinomial Distribution Tracking with Exponential Forgetting
Michael G. Epitropakis1, Dimirtis K. Tasoulis2, Nicos G. Pavlidis3, Vassilis P. Plagianakos4, and Michael N. Vrahatis1
1Computational Intelligence Laboratory, Department of Mathematics, University of Patras, GR-26110, Patras, Greece
[email protected]
2Winton Capital Management, 1-5 St Mary Abbots Place, London SW8 6LS, U.K.
[email protected]
3Department of Management Science, Lancaster University, LA1 4YX, U.K.
[email protected]
4Department of Computer Science and Biomedical Informatics, University of Central Greece, GR-35100, Lamia, Greece
[email protected]
Abstract. Several Differential Evolution variants with modified search dynamics have been recently proposed, to improve the performance of the method. This work borrows ideas from adaptive filter theory to develop an “online” algorithmic adaptation framework. The proposed framework is based on tracking the parameters of a multinomial distribution to reflect changes in the evolutionary process. As such, we design a multinomial distribution tracker to capture the successful evolution movements of three Differential Evolution algorithms, in an attempt to aggregate their characteristics and their search dynamics. Experimental results on ten benchmark functions and comparisons with five state-of-the-art algorithms indicate that the proposed framework is competitive and very promising.
Keywords: Differential Evolution, Adaptation, Multinomial Distribution, Exponential forgetting
LNAI 7297, p. 214 ff.
[email protected]
© Springer-Verlag Berlin Heidelberg 2012