Optimal design method to automobile problems using holographic neural network’s approximation

  • Qinzhong Shi
  • Ichiro Hagiwara


A great number of functional evaluations may be required until reaching the convergence in the process of optimization. Although the approximation models constructed by the response surface methodology are usually used to get the optimal designs, it is thought that the design accuracy is dependent on the type of activate functions and the design region of interest. In this paper, techniques to search all the local optimal designs within the feasible design region, and techniques for finding more accurate approximation using Holographic Neural Network (HNN) are investigated. Furthermore, the proposed method is applied to the problems frequently encountered in design of automobiles such as increasing of energy dissipation in crashworthiness and reducing of interior noise to illustrate the effectiveness of the proposed method.

Key words

nonlinear problem holographic neural network function approximation response surface methodology structural analysis vehicle crashworthiness vehicle interior noise 


  1. [1]
    G. Box, W. Hunter and J. Hunter, Empirical Model Building and Response Surface. John Wiley & Sons, New York, 1987.Google Scholar
  2. [2]
    W. Carpenter and J. Barthelemy, A comparison of polynomial approximation and artificial neural nets as response surfaces. Proc. of the 33rd Structure, Structural Dynamics, and Material Conf. AIAA, Washington DC (AIAA Paper 92-2247), 1994. 2474–2483.Google Scholar
  3. [3]
    I. Hagiwara, Buckling dynamic analysis of thin-walled columns. Japan Journal of Industrial and Applied Mathematics, Proc., Sept. 19–21,1994, 126–127, (in Japanese).Google Scholar
  4. [4]
    I. Hagiwara, et al., Technical development of eigenmode sensitivity analysis for coupled acoustic-structural systems. JSME Transaction,56, No.527 (1990), 1704–1711, (in Japanese).Google Scholar
  5. [5]
    K. Hornik, et al., Multilayer feedforward networks are universal approximators. Neural Net- works,2 (1989), 359–366.CrossRefGoogle Scholar
  6. [6]
    H. Miura, et al., Experience in large scale structural design optimization. Applications of Supercomputers in Engineering: Fluid Flow and Stress Analysis Applications, Elsevier, Am- sterdam, 1989, 251–268.Google Scholar
  7. [7]
    H. Miura, Design optimization with LS-DYNA. 4th International LS-DYNA3D Conference, Mineapolis, Minnesota, U.S.A., Sept., 1996, 59–70.Google Scholar
  8. [8]
    W.J. Roux, N. Stander and R.T. Haftka, Response surface approximation for structural optimization. AIAA-96-4042-CP, 1996, 565–578.Google Scholar
  9. [9]
    J.G. Sutherland, The holographic model of memory, learning and expression. International Journal of Neural System,1, No.3 (1990), 259–267.CrossRefGoogle Scholar
  10. [10]
    K.P. White Jr, et al., Simulation optimization of the crashworthiness of a passenger vehicle in frontal collision using response surface methodology. SAE Transaction, Sec.3, 1985, 798–811.Google Scholar

Copyright information

© JJIAM Publishing Committee 2000

Authors and Affiliations

  • Qinzhong Shi
    • 1
  • Ichiro Hagiwara
    • 1
  1. 1.Department of Mechanical Science and EngineeringTokyo Institute of TechnologyTokyoJapan

Personalised recommendations