Global optimization by multilevel search
- 76 Downloads
A new approach to the global optimization of functions with extremely rugged graphs is introduced. This multilevel search method is both an algorithm and a meta-algorithm, a logic for regulating optimization done by other algorithms. First, it is examined in the one-dimensional case theoretically and through simple examples. Then, to deal with higher dimensions, multilevel search is combined with the Monte Carlo method; this hybrid algorithm is tested on standard problems and is found to perform extremely well for a derivative-free method.
Key WordsMultilevel methods global optimization Monte Carlo methods
Unable to display preview. Download preview PDF.
- 1.Dixon, L. C. W., andSzego, G. P.,Toward Global Optimization 2, North-Holland, New York, New York, 1978.Google Scholar
- 2.Parker, G., andHardin, R.,Discrete Optimization, Academic Press, New York, New York, 1988.Google Scholar
- 3.Conley, M.,Computer Optimization Techniques, Petrocelli Books, New York, New York, 1984.Google Scholar
- 4.Rinnooy-Kan, A. H. G., andTimmer, G. T.,Stochastic Methods for Global Optimization, American Journal of Mathematics and Management Sciences, Vol. 4, pp. 7–40, 1984.Google Scholar
- 5.Wong, D. F., Leong, H. W., andLiu, C. L.,Simulated Annealing for VLSI Design, Klüwer Academic Press, Boston, Massachusetts, 1988.Google Scholar
- 6.Aarts, K.,Simulated Annealing, Academic Press, New York, New York, 1987.Google Scholar
- 7.Goertzel, B. N.,A Multilevel Approach to Global Optimization, PhD Thesis, Temple University, 1990.Google Scholar
- 8.Dixon, L. C. W., andSzego, G. P.,The Global Optimization Problem: An Introduction, Toward Global Optimization 2, Edited by L. C. W. Dixon and G. P. Szego, North-Holland, New York, New York, pp. 1–15, 1978.Google Scholar
- 9.Stuckman, B.,Global Search Methods for Optimizing Nonlinear Systems, IEEE Transactions on Systems, Man, and Cybernetics, Vol. 5, pp. 965–971, 1988.Google Scholar