2
$\begingroup$

Given two independent random variables (which describe two workers' durations to complete a task) $t_1, t_2$ which are normally distributed with ($\mu_1=20,\sigma_1=5$) and ($\mu_2=30,\sigma_2=10$). A task is given at random to worker 1 or 2 with equal probability of $p=0.5$.

A task is completed in $t=10$. Now, I would like to find the probability of this task being completed by worker 2.

It would be great if someone could guide me through this task. So far I came up with the following thought-process:

1) I assume workers durations to be independent, thus $p(t)=p(t_1)p(t_2)$ (But not sure what to do with the equal probability of 0.5)

2) My goal is to find $p(\text{worker}_2\vert t=10)$ but I am not sure how?

Ideas I had:

  • Use $p(t)=\mathcal{N}(\mu_1+\mu_2,\sigma_1^2+\sigma_2^2 )$
$\endgroup$
0
$\begingroup$

There are two type of workers, $W = 1$ and $W=2$. Given an observation of completion time $t$, we can use Bayes' rule to compute the probability that this observation was produced by a worker of type $i$: \begin{equation} p(W=i|t) = \frac{p(t|W=i)\,p(W=i)}{p(t)} , \end{equation} where \begin{equation} p(t) = \sum_{i=1}^2 p(t|W=i)\,p(W=i) . \end{equation}

In the setup we are told $p(W=1) = p(W=2) = 1/2$, $p(t|W=1) = \textsf{N}(t|20,5^2)$, and $p(t|W=2) = \textsf{N}(t|30,10^2)$. Therefore, \begin{equation} p(W=2|t=10) = \frac{\textsf{N}(10|30,10^2)\times \frac{1}{2}}{\textsf{N}(10|20,5^2)\times \frac{1}{2} + \textsf{N}(10|30,10^2)\times \frac{1}{2}} = \frac{1}{3} . \end{equation}

$\endgroup$
  • $\begingroup$ My teacher always tells us that the normal distribution is zero for a single value (e.g. like here t=10) - why does this not apply here? $\endgroup$ – Oliver Apr 26 '18 at 9:07
  • $\begingroup$ The probability of a single value is indeed zero. But the density at a single value is not zero. The answer involves the densities of a single value for two different distributions. $\endgroup$ – mef Apr 26 '18 at 12:24
  • $\begingroup$ It is confusing, then, to call these quantities "probabilities" in your answer and then claim in a comment that they are "densities." In fact, some of them are probabilities and some of them are densities! $\endgroup$ – whuber Feb 3 at 21:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.