A comparator on a photodiode channel fires when the noise voltage

PHYSICS · Multi-part · Fermi University Physics · Mathematical Methods for Physics · Probability, statistics and stochastic methods · Probability, statistics and stochastic methods

Error, Chance and the Machine, Ch Monte Carlo integration and variance reduction

Problem

A comparator on a photodiode channel fires when the noise voltage ZZ, a standard Gaussian in units of its rms value, exceeds a threshold of 44. The false-alarm probability P=P(Z>4)=3.167124×10−5P=P(Z>4)=3.167124\times10^{-5} is to be estimated by importance sampling, with f=1f=1 for x>4x>4 and 00 otherwise, p(x)=e−x2/2/2πp(x)=e^{-x^2/2}/\sqrt{2\pi} and the shifted exponential proposal qλ(x)=λ e−λ(x−4),x>4,q_\lambda(x)=\lambda\,e^{-\lambda(x-4)},\qquad x>4, where the rate λ>0\lambda>0 is to be chosen. Every draw from qλq_\lambda has f=1f=1, so the estimate is N−1∑iw(Xi)N^{-1}\sum_iw(X_i) with w=p/qλw=p/q_\lambda. Evaluate Gaussian tails P(Z>t)P(Z>t) with software. ((a)) Show that qλ/p≥cq_\lambda/p\ge c for all x>4x>4 with c=λ2π e4λ−λ2/2c=\lambda\sqrt{2\pi}\,e^{4\lambda-\lambda^2/2}, and conclude that the variance of the estimator is finite for every λ>0\lambda>0. ((b)) Show that the second moment M(λ)=∫4∞p2/qλ dxM(\lambda)=\int_4^\infty p^2/q_\lambda\,\mathrm{d}x equals M(λ)=eλ2/4−4λ2π λ P ⁣(Z>42−λ2),M(\lambda)=\frac{e^{\lambda^2/4-4\lambda}}{2\sqrt\pi\,\lambda}\,P\!\left(Z>4\sqrt2-\frac{\lambda}{\sqrt2}\right), and find the λ∗\lambda^* that minimises it, to three significant figures. ((c)) Find the relative error of the estimate from N=104N=10^4 draws, in per cent, at λ∗\lambda^*, at λ=4\lambda=4, and for the shifted unit Gaussian q(x)=e−(x−4)2/2/2πq(x)=e^{-(x-4)^2/2}/\sqrt{2\pi}, for which M=e16P(Z>8)M=e^{16}P(Z>8) with P(Z>8)=6.22096×10−16P(Z>8)=6.22096\times10^{-16}.

Browse the problem library