Give an upper bound on the probability that you wait more than 30min for the bus.

75 Views Asked by At

Consider the time to wait for your bus is modeled by $X\text{~}Exp(\lambda)$.

What is the $90/100th$ quantile for $\lambda=1$?

Given $90/100th$ quantile is 20 min ($\lambda$ is unknown). Give an upper bound on the probability that you wait more than 30 min. (The upper bound must be less than $0.5$).

My attempt:

For the first part, I can just use the formula that $\mathbb{P}(X\geq x_p)=1-p$ where $p=90/100$, and the formula for Exponential distribution, which is $f(x)=\lambda e^{-\lambda x}$. $$1e^{-1x}=1-90/100$$ $$x=-\ln(10/100)=2.3025$$

However, I don't understand what is meant by find the upper bound for the probability. Which formulas are involved and how should I approach this problem? Thank you.

1

There are 1 best solutions below

0
On

$X \sim Exp(\lambda)$

The PDF $f(x) = \lambda.e^{-\lambda x}$

The CDF $F(x) = P(X \leq x) = 1 - e^{-\lambda x}$

The quantile function $F^{-1}(p) = \frac{1}{\lambda}\ln \frac{1}{1-p}$

For $\lambda=1$ the $\frac{90}{100}$ quantile is $\ln \frac{1}{1-\frac{9}{10}} = \ln 10 = 2.302585$ mins

Given, $F^{-1}(90/100) = \frac{1}{\lambda}\ln \frac{1}{1-90/100} = \frac{\ln 10}{\lambda} = 20 \implies \lambda = \frac{\ln 10}{20}$

Note that it implies $P(X \leq 20) = F(20) = \frac{90}{100}$ and $P(X > 20) = \frac{10}{100}$

$\therefore P(X > 30)$

$= 1 - P(X \leq 30) = 1 - F(30) = e^{-\lambda. 30} = e^{-\frac{\ln 10}{20}. 30}$

$= (e^{-\ln 10})^\frac{3}{2} = \left(\frac{1}{10}\right)^\frac{3}{2} = 0.03162278$