A Nonhomogenous Poisson Process Question - Harry's Stressful Life

215 Views Asked by At

Due to stress of coping with business, Harry begins to experience migraine headaches of random severities. Headaches are instantaneous and has zero duration. The times when headaches occur follow a Poisson process of rate λ. Headache severities are independent of times of occurrences and are i.i.d. random variables with common exponential distribution given by: $$ \mathbb{P}[H \le x] = 1 − e^{-x}, \quad x > 0 $$ Harry decides to commit himself to the hospital if a headache of severity greater than $c > 0$ occurs in the time period $[0, t]$.

Compute the probability that Harry does not commit himself in $[0, t]$.

I think that probability of not committing is $\mathbb{P}[H(t)=0]$

$\mathbb{P}[H(t)=0]$ should be equal to $e^{-λt \mathbb{P}[H_s]}$ where $H_s$ is the occurrence probability of a severe headache.

But how can I calculate $\mathbb{P}[H_s]$? Should I find $\mathbb{P}[H_s>t] = e^{-tm}$ ??

1

There are 1 best solutions below

0
On

Let $N(t)$ be the number of headaches experienced in $(0,t]$. Then $N(t)\sim\mathsf{Pois}(\lambda t)$, so $$\mathbb P(N(t)=n) = \frac{(\lambda t)^n}{n!}e^{-\lambda}, n=0,1,2,\ldots. $$ The probability that a given headache exceeds severity $c$ is $$\mathbb P(H_1>c) = e^{-c}.$$ Let $N_1(t)$ be the number of non-critical headaches in $(0,t]$ and $N_2(t)$ be the number of critical headaches in $(0,t]$. Then $N(t)=N_1(t)+N_2(t)$ and $N_1(t)$, $N_2(t)$ are independent Poisson processes with rates $\lambda(1-e^{-c})$ and $\lambda e^{-c}$. Therefore the probability that Harry does not commit himself in $(0,t]$ is $$\mathbb P(N_2(t)=0) = e^{-\lambda t e^{-c}} $$ and the probability that he does is $1- e^{-\lambda te^{-c}}$.