User Tools

Site Tools


math121a-f23:september_27_wednesday

September 27, Wednesday

Gaussian integral is of the form (modulo constant) I=ex2dxI = \int_{-\infty}^\infty e^{- x^2} dx How to evaluate this? here residue theorem cannot help you. you need to use a little trick.

Consider the double integral I2=ex2y2dxdyI^2 = \int_{-\infty}^\infty\int_{-\infty}^\infty e^{- x^2 -y^2 } dx dy then switch to radial coordinate I2=r=0θ=02πer2rdrdθ=2πr=0er2rdr I^2 = \int_{r=0}^{\infty}\int_{\theta=0}^{2\pi} e^{-r^2} r drd\theta = 2 \pi \int_{r=0}^{\infty} e^{-r^2} r dr subsitatue r2=ur^2=u, we get 2rdr=du2rdr = du, and I2=π0eudu=π I^2 = \pi \int_0^\infty e^{-u} du = \pi ok, we get I=π. I = \sqrt{\pi}.

OK, not bad. How about more general case? For a>0a>0, consider Ia=eax2dxI_a = \int_{-\infty}^\infty e^{- a x^2} dx we can change variable, let u=axu = \sqrt{a} x, then dx=(1/a)dudx = (1/\sqrt{a}) du, Ia=(1/a)eu2du=(1/a)I=(1/a)π.I_a = (1/\sqrt{a}) \int_{-\infty}^\infty e^{- u^2} du = (1/\sqrt{a}) I = (1/\sqrt{a}) \sqrt{\pi}.

How about a=reiθa = r e^{i \theta}? and 0<θ«10<\theta«1? and then making θ\theta larger, and larger?

math121a-f23/september_27_wednesday.txt · Last modified: 2023/09/26 21:35 by pzhou