Metaoptimization of Differential Evolution by Using Productions of Low-Number of Cycles: The Fitting of Rotation Curves of Spiral Galaxies as Case Study
In order to increase the efficiency of Evolutionary Algorithms, practitioners include improvements as new operators or modifications of the canonical operators, or the hybridization with other Evolutionary Algorithms. However, an alternative to obtain high-quality solutions is: to tune the parameters which govern the behaviour of the algorithm to the specific problem to optimize. This parameters adjustment can be performed by using other Evolutionary Algorithm (Metaoptimization). Unfortunately, metaoptimization leads to a critical increment in the execution time. In this work, a measure of the quality of the tuned behavioural parameters when executing very low-number of cycles in the optimizer is performed and compared with the case when executing high-number of cycles. The fundamental aspect of this approach is if there is enough information about the quality of the behavioural parameters in the very initial cycles of the optimizer. By ascertaining if productions based on a low-number of cycles harvest high-quality behavioural parameters, one of the main drawbacks of the metaoptimization process —the large execution time— can be overcome. The performed tests —the fitting of experimental data of rotation curves of spiral galaxies— demonstrate that this approach improves the efficiency of the metaoptimizer, while reducing processing time.
KeywordsMetaoptimization Differential Evolution Rotation Curve Spiral Galaxy
Unable to display preview. Download preview PDF.
- 2.Price, K.V., Storn, R., Lampinen, J.: Differential Evolution: A practical Approach to Global Optimization. Springer, Berlin (2005)Google Scholar
- 7.Birattari, M., Stützle, T., Paquete, L., Varrentrapp, K.: A racing algorithm for configuring metaheuristics. In: GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference, New York, USA, July 9-13, pp. 11–18. Morgan Kaufmann (2002)Google Scholar
- 9.Pedersen, M.E.H.: Good Parameters for Differential Evolution. Technical Report Technical report no. HL1002, Hvass Laboratories, University of Zurich, Department of Informatics (2010)Google Scholar
- 10.Smit, S.K., Eiben, A.E.: Comparing Parameter Tuning Methods for Evolutionary Algorithms. In: IEEE Congress on Evolutionary Computation (CEC), pp. 399–406 (May 2009)Google Scholar
- 11.Mezura-Montes, E., Velázquez-Reyes, J., Coello, C.A.C.: A comparative study of differential evolution variants for global optimization. In: GECCO, Genetic and Evolutionary Computation Conference, Seattle, Washington, USA, July 8-12, pp. 485–492. ACM (2006)Google Scholar
- 13.Cárdenas-Montes, M., Vega-Rodríguez, M.A., Gómez-Iglesias, A.: Real-world problem for checking the sensitiveness of evolutionary algorithms to the choice of the random number generator. In: Corchado, E., Snášel, V., Abraham, A., Woźniak, M., Graña, M., Cho, S.-B. (eds.) HAIS 2012, Part III. LNCS, vol. 7208, pp. 385–396. Springer, Heidelberg (2012)CrossRefGoogle Scholar
- 15.Cárdenas-Montes, M., Mollá, M., Vega-Rodríguez, M.A., Rodríguez-Vázquez, J.J., Gómez-Iglesias, A.: Adjustment of observational data to specific functional forms using a particle swarm algorithm and differential evolution: Rotational curves of a spiral galaxy as case study. In: Sarro, L.M., Eyer, L., O’Mullane, W., De Ridder, J. (eds.) Astrostatistics and Data Mining. Springer Series in Astrostatistics, vol. 2, pp. 81–88. Springer, New York (2012)CrossRefGoogle Scholar