2
$\begingroup$

Here's the problem:

A Flight has 20 seats and enough demand to sell out all flights (not enough to justify buying bigger plane). All seats sell for $200.

Probability a passenger with reservation shows up = p. The probability of “No show”=1-p. The occurrence of “Show/No show” is independent among passengers. Passengers who are turned away due to “overbooking” are given $240 (the purchase price plus a 20% penalty to airline) The number of reservations made for each flight is chosen by airline (not random): n

The number of booked passengers who show up is a random variable: X

  • What is the probability distribution of X?
  • Write out the Revenues to the airline for a flight as a function of X
  • Write out the expected Revenues to the airline
  • What is the effect of increasing n on expected revenues for a given p?
  • What is the effect of increasing p on expected revenues for a given n?
  • When would it make no sense to overbook flights?

My thoughts: I think the answer to the first part is that it is a binomial distribution (since we have two distinct outcomes and a constant and independent probability of success). However, I am facing problem in the remaining parts. Can someone please explain it to me? Thanks so much :)

$\endgroup$
2
  • $\begingroup$ Are all tickets non-refundable? $\endgroup$ – Fabio Somenzi Mar 26 '17 at 3:36
  • $\begingroup$ Yup. Sorry forgot to include that assumption $\endgroup$ – user3834127 Mar 26 '17 at 4:02
1
$\begingroup$

What is the probability distribution of X?

Yes, $X \sim Binom(n, p)$. That is, \begin{equation} P(X=x) = \binom{n}{x}p^x(1-p)^x \end{equation}

Write out the revenues to the airline as a function of X.

If $X \leq 20$, then it is pure profit for the airlines. If $X > 20$, then the airline loses 40 dollars for each passenger over 20. \begin{equation} R(X) = 200X - \max[0, 240(X-20)] \end{equation} Or... \begin{equation} R(X) = \begin{cases} 200X & ;X \leq 20 \\ 200(20) - 40(X-20) & ;X > 20 \end{cases} \end{equation}

Write out the expected value of R(X).

Remember that $E[R(X)] = \sum_{x=0}^n R(X)P(X=x)$. The math gets a bit tricky, but it can be done using R(X) as given above.

What is the effect of increasing n on expected revenues for a given p?

Eventually, it is going to decrease the expected revenue for the airline. Basically if the airline overbooks too often, they will lose a lot of money. However if $p$ is small, the expected value of the revenue may be optimal for some $n > 20$. Revenue as f(n)

What is the effect of increasing p on expected revenues for a given n?

Again, it should decrease the expected revenue. If $n$ is fixed (and I'm assuming greater than 20), then increasing $p$ means the airline is more likely to have to refund somebodies money. You should be able to show this formally if you evaluate $E[R(X)]$.Revenue as f(p)

When would it make no sense to overbook flights.

If $p=1$ (or in practice, if it is close to 1), then it makes no sense. Because they will always have to refund somebodies money.

$\endgroup$
9
  • $\begingroup$ Hi! Thank you so so much for the solution! I finally understand it now. However, I am confused with the notation you used R(X)=200X−max[0,40(X−20)]. What does the max part mean? I'm sorry I'm new to this as I am still a high school student. $\endgroup$ – user3834127 Mar 26 '17 at 4:14
  • $\begingroup$ Also, I was wondering if there is a method to write the revenue as a function of n? $\endgroup$ – user3834127 Mar 26 '17 at 4:15
  • $\begingroup$ You're correct. But even if the tickets are non-refundable, they get monetary compensation greater than the ticket price. It is thus assumed, for the bumped passengers, that the monetary compensation they receive is over and above their ticket price. I'm sorry for the confusion $\endgroup$ – user3834127 Mar 26 '17 at 4:29
  • $\begingroup$ Yup @FabioSomenzi! Thanks so much for helping out! Also, can you help me write the revenue as a function of n? $\endgroup$ – user3834127 Mar 26 '17 at 5:02
  • 1
    $\begingroup$ Agreed it's not a big deal to change. I asked the OP and that's what (s)he said. (See comments to the question.) $\endgroup$ – Fabio Somenzi Mar 26 '17 at 22:27

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.