Boosting the Performance of a Multiobjective Algorithm to Design RBFNNs Through Parallelization

  • Alberto Guillén
  • Ignacio Rojas
  • Jesus González
  • Hector Pomares
  • Luis J. Herrera
  • Ben Paechter
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4431)

Abstract

Radial Basis Function Neural Networks (RBFNNs) have been widely used to solve classification and regression tasks providing satisfactory results. The main issue when working with RBFNNs is how to design them because this task requires the optimization of several parameters such as the number of RBFs, the position of their centers, and their radii. The problem of setting all the previous values presents many local minima so Evolutionary Algorithms (EAs) are a common solution because of their capability of finding global minima. Two of the most important elements in an EAs are the crossover and the mutation operators. This paper presents a comparison between a non distributed multiobjective algorithm against several parallel approaches that are obtained by the specialisation of the crossover and mutation operators in different islands. The results show how the creation of specialised islands that use different combinations of crossover and mutation operators could lead to a better performance of the algorithm by obtaining better solutions.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Park, J., Sandberg, J.W.: Universal approximation using radial basis functions network. Neural Computation 3, 246–257 (1991)CrossRefGoogle Scholar
  2. 2.
    Rojas, I., Anguita, M., Prieto, A., Valenzuela, O.: Analysis of the operators involved in the definition of the implication functions and in the fuzzy inference proccess. International Journal of Approximate Reasoning 19, 367–389 (1998)MATHCrossRefMathSciNetGoogle Scholar
  3. 3.
    Deb, K., Agrawal, S., Pratap, A., Meyarivan, T.: A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Trans. Evolutionary Computation 6(2), 182–197 (2002)CrossRefGoogle Scholar
  4. 4.
    Tang, Y., Reed, P., Wagene, T.: How effective and efficient are multiobjective evolutionary algorithms at hydrologic model calibration?. Hydrology and Earth System Sciences 10, 289–307 (2006)CrossRefGoogle Scholar
  5. 5.
    Guillén, A., Rojas, I., González, J., Pomares, H., Herrera, L., Valenzuela, O., Prieto, A.: A Possibilistic Approach to RBFN Centers Initialization. In: Ślęzak, D., Yao, J., Peters, J.F., Ziarko, W., Hu, X. (eds.) RSFDGrC 2005. LNCS (LNAI), vol. 3642, pp. 174–183. Springer, Heidelberg (2005)CrossRefGoogle Scholar
  6. 6.
    Guillén, A., Rojas, I., González, J., Pomares, H., Herrera, L.J., Valenzuela, O., Prieto, A.G.: Improving Clustering Technique for Functional Approximation Problem Using Fuzzy Logic: ICFA Algorithm. In: Cabestany, J., Prieto, A.G., Sandoval, F. (eds.) IWANN 2005. LNCS, vol. 3512, pp. 272–279. Springer, Heidelberg (2005)Google Scholar
  7. 7.
    Marquardt, D.W.: An Algorithm for Least-Squares Estimation of Nonlinear Inequalities. SIAM J. Appl. Math. 11, 431–441 (1963)MATHCrossRefMathSciNetGoogle Scholar
  8. 8.
    Hancock, P.J.B.: Genetic Algorithms and Permutation Problems: a Comparison of Recombination Operators for Neural Net Structure Specification. In: Whitley, D. (ed.) Proceedings of COGANN workshop, IJCNN, Baltimore, IEEE Computer Society Press, Los Alamitos (1992)Google Scholar
  9. 9.
    Herrera, F., Lozano, M., Verdegay, J.L.: Tackling real-coded genetic algorithms: operators and tools for the behavioural analysis. Artificial Intelligence Reviews 12(4), 265–319 (1998)MATHCrossRefGoogle Scholar
  10. 10.
    Knowles, J., Corne, D.: On metrics for comparing non-dominated sets. In: Congress on Evolutionary Computation, CEC 2002 (2002)Google Scholar
  11. 11.
    van Veldhuizen, D.A., Zydallis, J.B., Lamont, G.B.: Considerations in engineering parallel multiobjective evolutionary algorithms. IEEE Trans. Evolutionary Computation 7(2), 144–173 (2003)CrossRefGoogle Scholar

Copyright information

© Springer Berlin Heidelberg 2007

Authors and Affiliations

  • Alberto Guillén
    • 1
  • Ignacio Rojas
    • 2
  • Jesus González
    • 2
  • Hector Pomares
    • 2
  • Luis J. Herrera
    • 2
  • Ben Paechter
    • 3
  1. 1.Department of Computer Science, University of JaénSpain
  2. 2.Department of Computer Architecture and Computer Technology, Universidad de GranadaSpain
  3. 3.School of Computing Napier UniversityScotland

Personalised recommendations