Abstract

VELASQUEZ-HENAO, Juan; PULGARIN-AGUDELO, Yeiny  and  CASTANO-ARIAS, Eliana. Monte Carlo Optimization Using Beta Distribution. Ing. Univ. [online]. 2011, vol.15, n.1, pp.61-76. ISSN 0123-2126.

This paper presents an innovating Monte Carlo method for exploring n-dimensional non-linear functions defined in a compact domain which is transformed to the hypercube [0;1 ]ⁿ. This approach uses the beta distribution for generating random samples. Te distribution parameters, named alpha and beta, are dynamically adjusted so that, in the first iterations, the beta distribution looks like the uniform distribution. ; in the last iterations, the beta distribution is centered in the known minimum and the variance is near zero, so that only the neighborhood around the optimum is sampled. The method proposed is tested through four well known benchmark functions.

Keywords : Monte Carlo method; heurísticas.

        · abstract in Spanish | Portuguese     · text in Spanish     · Spanish ( pdf )