Stochastic Tunneling for Improving the Efficiency of Stochastic Efficient Global Optimization
This paper proposes the use of a normalization scheme for increasing the performance of the recently developed Adaptive Target Variance Stochastic Efficient Global Optimization (sEGO) method. Such a method is designed for the minimization of functions that depend on expensive to evaluate and high dimensional integrals. The results showed that the use of the normalization in the sEGO method yielded very promising results for the minimization of integrals. Indeed, it was able to obtain more precise results, while requiring only a fraction of the computational budget of the original version of the algorithm.
KeywordsStochastic efficient global optimization Stochastic tunneling Global optimization Robust design
The authors acknowledge the financial support and thank the Brazilian research funding agencies CNPq and CAPES.
- 4.Carraro, F., Lopez, R.H., Miguel, L.F.F., Andre J.T.: Optimum design of planar steel frames using the search group algorithm. Struct. Multidiscip. Optim. (2019) (p. to appear)Google Scholar
- 12.Wenzel, W., Hamacher, K.: Stochastic tunneling approach for global minimization of complex potential energy landscapes. Phys. Rev. Lett. 82, 3003–3007 (1999). https://doi.org/10.1103/PhysRevLett.82.3003, https://link.aps.org/doi/10.1103/PhysRevLett.82.3003MathSciNetCrossRefGoogle Scholar