Sequence Construction For Integral Estimation
暂无分享,去创建一个
We consider a deterministic numerical code which computes the infrared signature of an aircraft in its surroundings. This code takes as input parameters a large number of variables describing the aircraft and its environement characteristics. However, the code gives a result for a known physical configuration. Its use is then limited. Indeed, the code doesn’t allow to know the envelope describing the set of possible values of the signature, this envelope resulting from a partial knowledge of the aircraft and its environement characteristics. More precisely, for a given attack scenario, some of the variables are only known through their probability density function, like meteorological variables, and some other variables through their variation interval. This envelope is essential to estimate the detection performance of infrared sensors. The search of this envelope leads to the estimation of the integral of a function h, with unknown analytic form, defined on a large dimensional space. Typically, the dimension can be of order 30 up to 60.
[1] E. Hlawka. Funktionen von beschränkter Variatiou in der Theorie der Gleichverteilung , 1961 .
[2] Russel E. Caflisch,et al. Quasi-Random Sequences and Their Discrepancies , 1994, SIAM J. Sci. Comput..
[3] B. Efron,et al. The Jackknife Estimate of Variance , 1981 .