Costs and Benefits of Tuning Parameters of Evolutionary Algorithms

We present an empirical study on the impact of different design choices on the performance of an evolutionary algorithm (EA). Four EA components are considered--parent selection, survivor selection, recombination and mutation--and for each component we study the impact of choosing the right operator, and of tuning its free parameter(s). We tune 120 different combinations of EA operators to 4 different classes of fitness landscapes, and measure the cost of tuning. We find that components differ greatly in importance. Typically the choice of operator for parent selection has the greatest impact, and mutation needs the most tuning. Regarding individual EAs however, the impact of design choices for one component depends on the choices for other components, as well as on the available amount of resources for tuning.

[1]  A. E. Eiben,et al.  Introduction to Evolutionary Computing , 2003, Natural Computing Series.

[2]  Marcus Gallagher,et al.  A general-purpose tunable landscape generator , 2006, IEEE Transactions on Evolutionary Computation.

[3]  François E. Cellier,et al.  Artificial Neural Networks and Genetic Algorithms , 1991 .

[4]  K. Dejong,et al.  An analysis of the behavior of a class of genetic adaptive systems , 1975 .

[5]  Rajarshi Das,et al.  A Study of Control Parameters Affecting Online Performance of Genetic Algorithms for Function Optimization , 1989, ICGA.

[6]  Thomas Bartz-Beielstein,et al.  Sequential Parameter Optimization Applied to Self-Adaptation for Binary-Coded Evolutionary Algorithms , 2007, Parameter Setting in Evolutionary Algorithms.

[7]  B. Freisleben,et al.  Optimization of Genetic Algorithms by Genetic Algorithms , 1993 .

[8]  Mauro Birattari,et al.  Tuning Metaheuristics - A Machine Learning Perspective , 2009, Studies in Computational Intelligence.

[9]  Jason M. Daida,et al.  Parameter sweeps for exploring GP parameters , 2005, GECCO '05.

[10]  Zbigniew Michalewicz,et al.  Parameter Control in Evolutionary Algorithms , 2007, Parameter Setting in Evolutionary Algorithms.

[11]  A. E. Eiben,et al.  Efficient relevance estimation and value calibration of evolutionary algorithm parameters , 2007, 2007 IEEE Congress on Evolutionary Computation.

[12]  Dan Boneh,et al.  On genetic algorithms , 1995, COLT '95.

[13]  Zbigniew Michalewicz,et al.  Parameter Setting in Evolutionary Algorithms , 2007, Studies in Computational Intelligence.

[14]  D. J. Smith,et al.  A Study of Permutation Crossover Operators on the Traveling Salesman Problem , 1987, ICGA.

[15]  G. Unter Rudolph.  On Correlated Mutations in Evolution Strategies , 1992 .

[16]  Heinz Mühlenbein,et al.  The Estimation of Distributions and the Minimum Relative Entropy Principle , 2005, Evol. Comput..

[17]  Olivier François,et al.  Design of evolutionary algorithms-A statistical perspective , 2001, IEEE Trans. Evol. Comput..

[18]  Ritu Gupta,et al.  Statistical exploratory analysis of genetic algorithms , 2004, IEEE Transactions on Evolutionary Computation.

[19]  Jason M. Daida,et al.  Parameter Sweeps for Exploring Parameter Spaces of Genetic and Evolutionary Algorithms , 2007, Parameter Setting in Evolutionary Algorithms.

[20]  John J. Grefenstette,et al.  Optimization of Control Parameters for Genetic Algorithms , 1986, IEEE Transactions on Systems, Man, and Cybernetics.

[21]  Kenneth Alan De Jong,et al.  An analysis of the behavior of a class of genetic adaptive systems. , 1975 .

[22]  David E. Goldberg,et al.  Genetic Algorithms in Search Optimization and Machine Learning , 1988 .