On the Weibull cost estimation of building frames designed by simulated annealing
- First Online:
- Cite this article as:
- Paya-Zaforteza, I., Yepes, V., González-Vidosa, F. et al. Meccanica (2010) 45: 693. doi:10.1007/s11012-010-9285-0
- 169 Downloads
This paper proposes a general methodology to determine the number of numerical tests required to provide a solution for a heuristic optimization problem with a user-defined accuracy as compared to a global optimal solution. The methodology is based on the extreme value theory and is explained through a problem of cost minimization for reinforced concrete building frames. Specifically, 1000 numerical experiments were performed for the cost minimization of a two-bay and four-floor frame using the Simulated Annealing (SA) algorithm. Analysis of the results indicates that (a) a three-parameter Weibull distribution function fits the results well, (b) an objective and general procedure can be established to determine the number of experiments necessary to solve an optimization problem with a heuristic which generates independent random solutions, and (c) a small number of experiments is enough to obtain good results for the structural engineer.
KeywordsOptimization Reinforced concrete Weibull distribution Extreme value theory
Unable to display preview. Download preview PDF.