Multifidelity Global Optimization Using DIRECT
Multifidelity global optimization methods have the potential to allow the optimization of problems which are computationally expensive and have high dimensionality. The DIRECT algorithm has been modified to incorporate multifidelity information. Two correction strategies have been employed with problems with up to ten dimensions and thousands of local optima. DIRECT has demonstrated the ability to reach the same global optimum using high fidelity analyses or corrected low fidelity function values. Initial work with two correction strategies have shown a minor reduction in the number of high-fidelity analyses required but refinements to the stopping criteria have the potential for greatly reducing the cost of the optimization.
Unable to display preview. Download preview PDF.
- 6.Lamberti, L., Venkataraman, S., Haftka, R T. and Johnson, T. F., “Preliminary Design Optimization of Stiffened Panels Using Approximate Analysis Models”, Submitted For Review to the International Journal for Numerical Methods in Engineering, 2001Google Scholar
- 7.Swift, R A. and Batill, S. M., “Simulated Annealing Utilizing Neural Networks for Discrete Variable Optimization Problems in Structural Design”,33rd AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference, Dallas, TX, April 1992, pp. 2536–2544Google Scholar
- 9.Baker, C. A., Grossman, B., Haftka, R T., Mason, W. H. and Watson, L. T., “HSCT Configuration Design Space Exploration using Aerodynamic Response Surface Approximations”, 7th AIAA/USAF/NASA/ISSMO Symposium on Multidisciplinary Analysis and Optimization, St Louis, Missouri, Sept. 2–4, 1998, pp. 769–777, AIAA Paper-98-4803Google Scholar
- 13.Cox, S., Haftka, R.T., Baker, C., Grossman, B., Mason, W. and Watson, L. T., “Global Multidisciplinary Optimization of a High Speed Civil Transport”.Proc. Aerospace Numerical Simulation Symposium’ 99, Tokyo, Japan, June, 1999, pp. 23–28Google Scholar