a) $ \lim_{n\rightarrow\infty}n\int_{-1}^{1}e^{-(\frac{nx+1}{n})^2}-e^{-x^2}dx $

$ = \lim_{n\rightarrow\infty}\int_{-1}^{1}\frac{e^{-(x+\frac{1}{n})^2}-e^{-x^2}}{\frac{1}{n}}dx $

The Mean Value Theorem implies $ \exist h \in (0,\frac{1}{n}) $ s.t. $ \frac{e^{-(x+\frac{1}{n})^2}-e^{-x^2}}{\frac{1}{n}} = \frac{d}{dx}(e^{-x^2})|_{h} $

$ = \lim_{n\rightarrow\infty}\int_{-1}^{1}-2he^{-h^2}dx $

$ = \lim_{n\rightarrow\infty}-4he^{-h^2} $

as $ n\rightarrow\infty $, $ h\rightarrow 0 $ so

$ = 0 $

b) $ \lim_{n\rightarrow\infty}n\int_{-\infty}^{\infty}e^{-(\frac{nx+1}{n})^2}-e^{-x^2}dx $

$ = \lim_{n\rightarrow\infty}n\left[ \int_{-\infty}^{\infty}e^{-(\frac{nx+1}{n})^2}dx - \int_{-\infty}^{\infty}e^{-x^2}dx \right] $ To do this we need both integrals to exist, which they do since they are just a scale/shift of the normal distribution.

Let $ y = (\frac{nx+1}{n}) $ using the Riemann u-substitution since this is Riemann integrable.

$ = \lim_{n\rightarrow\infty}n\left[ \int_{-\infty}^{\infty}e^{-y^2}dy - \int_{-\infty}^{\infty}e^{-x^2}dx \right] $

$ = \lim_{n\rightarrow\infty} 0 $

$ = 0 $

Alumni Liaison

Ph.D. on Applied Mathematics in Aug 2007. Involved on applications of image super-resolution to electron microscopy

Francisco Blanco-Silva