Facebook Page
Twitter
RSS
+ Reply to Thread
Results 1 to 2 of 2

Thread: Poisson process

  1. MHB Apprentice

    Status
    Offline
    Join Date
    Aug 2019
    Posts
    5
    Thanks
    2 times
    Thanked
    0 times
    #1
    I have the following two problems that I need to solve:

    1. Suppose that the service time for a student enlisting during enrollment is modeled as an
    exponential RV with a mean time of 1 minute. If the school expects 500 students during
    enrollment period, what is the probability that the registration committee will finish enlisting all
    students within a day if a day constitutes a total of 8 working hours? Use the Central Limit
    Theorem.

    2. Suppose that you have a class at 8:45 am and you left your house at 8:00 am. To get to class, you
    ride a train, ride a taxi and then walk. The train waiting time, train trip time, taxi waiting time, taxi trip time,
    and walking time are all modeled as exponential random variables with a mean time of 10 mins,
    and they are assumed to be independent of each other. What is the probability that you will not
    be late for class?


    My attempted solution:
    1. Since mean time is 1 minute, and there are 480 mins in 8 hours, mean amount of students = 480
    Variance = 1/λ^2, so standard deviation = 1/λ = mean = 480
    Central limit theorem: Z = (S - mean)/standard_deviation --> (500-480)/480 = 0.0416666666

    2. Since all the waiting times are identical, I add them up and get a mean time of 50 mins.
    λ = 1/50 = 0.02
    probability for an exponential random variable is λe^-(λt)
    I then solve for the integral from 0 to 45 for 0.02e^-(0.02t) dt = 0.59344

    Are my answers correct?

  2. # ADS
    Circuit advertisement
    Join Date
    Always
    Posts
    Many
     

  3. MHB Apprentice

    Status
    Offline
    Join Date
    Dec 2018
    Posts
    38
    Thanks
    4 times
    Thanked
    31 times
    #2
    let's tackle part 1 first

    Quote Originally Posted by marcadams267 View Post
    I have the following two problems that I need to solve:

    1. Suppose that the service time for a student enlisting during enrollment is modeled as an
    exponential RV with a mean time of 1 minute. If the school expects 500 students during
    enrollment period, what is the probability that the registration committee will finish enlisting all
    students within a day if a day constitutes a total of 8 working hours? Use the Central Limit
    Theorem.
    ...

    My attempted solution:
    1. Since mean time is 1 minute, and there are 480 mins in 8 hours, mean amount of students = 480
    Variance = 1/λ^2, so standard deviation = 1/λ = mean = 480
    Central limit theorem: Z = (S - mean)/standard_deviation --> (500-480)/480 = 0.0416666666

    ...
    Are my answers correct?
    the fact that you didn't use a poisson distribution here is a red flag you need to think about. You should also explicitly write out the random variable and what is stands for... jumping straight into calculations can lead to problems like this.

    As is you seem to think the problem involves adding 480 iid exponential random variables. I don't see how this can possibly be the case -- we are interested in 500 students not 48. Note: even if your setup was correct, there's a linearity problem -- the combined random variable would in fact have variance of 480 but would have standard deviation of $\sqrt{480}$ -- variance adds in the case of iid random variables but standard deviation does not-- you can't interchange sums of positive numbers and square roots due to negative convexity.

    so writing this out, what you actually want to know is whether

    $S_{500} \leq 480$
    this is standard partial sum notation you may have seen in calculus e.g. $s_3 = x_1 + x_2 + x_3$, except we are summing random variables,

    $S_{500} = X_1 + X_2 + ... + X_{500}$

    where each $X_j$ represents a student, having iid arrival / service time that is an exponential random variable with parameter $\lambda = 1$

    so what is
    $Pr\big(S_{500} \leq 480\big) $

    well this is asking whether the first 500 arrivals occur at some time less than or equal to 480 minutes, which is equivalent to asking whether the number arrivals at time 480 are at least 500 in a poisson process. i.e.

    whether
    $Pr\big(N(t)) \geq 500\big)$
    with $t = 480$
    and $N(t)$ is the 'counting' random variable which counts the number of iid exponentially distributed arrivals in $(0, t]$

    and from your text you should know
    $Pr\big(N(t) = k )$

    is precisely given by a poisson distribution.

    From here I'd suggest calculating the answer exactly in say excel, and then doing the normal approximation and comparing answers.

    = = = = =
    note: you didn't give any background on the course or book you are using so I had to guess on what you know. It is possible that some of what I said won't make sense. My first suggestion of completely writing out the problem and what are the random variables you're interested, and why stands in any case. Second, there are many good texts on poisson processes. These days I'd probably recommend the freely available book by Blitzstein and Hwang which has a nice chapter on Poisson processes that you may want to go through. See here:


Similar Threads

  1. A math proof within a question about homogeneous Poisson process
    By ianchenmu in forum Advanced Probability and Statistics
    Replies: 1
    Last Post: January 26th, 2015, 19:54
  2. Proof about an Inhomogeneous Poisson Process
    By ianchenmu in forum Advanced Probability and Statistics
    Replies: 0
    Last Post: January 25th, 2015, 20:31
  3. Poisson Process - Number of cars that a petrol station can service
    By Usagi in forum Advanced Probability and Statistics
    Replies: 4
    Last Post: April 23rd, 2014, 22:31
  4. probability that all N_Q packets arrived in [0,t], in a Poisson process
    By hemanth in forum Advanced Probability and Statistics
    Replies: 1
    Last Post: September 2nd, 2013, 06:50
  5. Poisson process derivation
    By Poirot in forum Advanced Probability and Statistics
    Replies: 4
    Last Post: May 4th, 2012, 08:27

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
Math Help Boards