Expectation with a variable that satisfies uniform distribution

92 Views Asked by At

Do we have step by step proof for this expression which is concerned with the expectation of an exponential function with a variable (R) that satisfies uniform distribution. I have been trying to understand how it was derived but its seems obscure to me up till now:

$\frac{\lambda_1}{2}E_R (e^{-B_{1}R^2})E_R (e^{-D_{1}R^2}) = \\ \frac{\lambda_1 \pi}{2\sqrt{B_{1}D_{1}}(R_{max}-R_{min})^2}[\phi (2\sqrt{B_{1}}R_{max})-\phi (2\sqrt{B_{1}}R_{min})][\phi (2\sqrt{D_{1}}R_{max})-\phi (2\sqrt{D_{1}}R_{min})]$

Where $\phi$ is the cdf of standard normal distribution

R satisfies uniform distribution in $ [R_{min}, R_{max}]$ $(0\leq R_{min} < R_{max})$

1

There are 1 best solutions below

4
On BEST ANSWER

Consider the general form of the following expectation, where $R$ has a uniform probability density funcion $f$ on the interval $(a,b)$: $$\begin{align}E[e^{-c\,R^2}]&=\int_{-\infty}^\infty e^{-c\,r^2}\,f(r)\,dr\\ \\ &=\int_{-\infty}^\infty e^{-c\,r^2}\frac{1}{b-a}[a<r<b]\,dr\\ \\ &=\frac{1}{b-a}\int_{a}^b e^{-c\,r^2}\,dr\\ \\ &=\frac{1}{b-a}\left(\int_{-\infty}^b e^{-c\,r^2}\,dr\ -\ \int_{-\infty}^a e^{-c\,r^2}\,dr\right)\tag{1} \end{align}$$ Now condider the definition of the standard normal CDF: $$\Phi(x) = \frac{1}{\sqrt{2\pi}}\int_{-\infty}^x e^{-\frac{1}{2}t^2}\,dt.$$ Letting $r=\frac{1}{\sqrt{2c}}\,t$, so $c\,r^2=\frac{1}{2}t^2$ and $dr=\frac{1}{\sqrt{2c}}dt$, we have: $$\begin{align}\int_{-\infty}^a e^{-c\,r^2}\,dr= \int_{-\infty}^{\sqrt{2c}\,a}e^{-\frac{1}{2}t^2}\,\frac{1}{\sqrt{2c}}dt=\sqrt{\frac{\pi}{c}}\,\Phi(\sqrt{2c}\,a) \end{align}$$ Substituting this, and the similar result involving $b$, into (1) gives the following: $$\begin{align}E[e^{-c\,R^2}]=\frac{1}{b-a}\sqrt{\frac{\pi}{c}}\bigg( \Phi(\sqrt{2c}\,b)\ -\ \Phi(\sqrt{2c}\,a)\bigg)\end{align}$$ Applying this formula to each of the expectations in your question will give the desired result. (Note that the factor of $\frac{\lambda_1}{2}$ is superfluous, as it appears on both sides of your equation.)