Background Notes For Solving Case

Background Notes for Solving the Case Queuing Quandary Queuing theory is one of the quantitative techniques that
is in most demand in variety of business sectors like computer networking, telecom industry, share trading, aviation industry, banking, call centers and so on. What you have studied in lectures is simple Poisson Arrival- Exponential Service queues with single server or multi servers, finite or infinite population queues. These are very elementary queues where you have seen few algebraically simplified formulae. These are not the cases in real life. The probability distributions could be different for arrival as well as service (you need to use probability formulae learnt in SMRM); queues could be combination of parallel and series queues; there could be priority customers; holding capacity in queue could be limited; balking (you may not join long queues), jockeying (jumping between queues) or reneging (leave queue if you anticipate its taking long time) of the queue may be common. In many cases you need to develop the model and solve using fundamental theory of BirthDeath process of queue. However, many common situations could be approximated to Poisson Arrival- Exponential Service queues. As discussed above, many times probability of waiting beyond certain time could be a decision factor. This is the case in Queuing Quandary. We need to use probability formulae. These are provided to you by faculty. Purpose of this note is to explain theoretical basis for these formulae. Now there are two events that determine your waiting and associated probability in a queue. One is arrival of the customer and other is completing service to a customer. If there are two independent events; one at an exponential rate and another at an exponential rate . In such a case by theory of probability of joint distributions we can show that these two events occur together at an exponential rate + . (I wont trouble you with the proof. Any discerning student could come to me for that.) Since the arrival rate is of entering the system and leaving the system, we consider sign of as minus. Thus in our simple Poisson Arrival- Exponential Service queue, combined events occur at an exponential rate ( - ). Now do you remember that we have learnt in probability distribution that for exponential distribution with parameter (rate), the c.d.f. (cumulative distribution function) is, F(t) = P (X t) = 1 Hence P ( X > t) = Now in Poisson Arrival- Exponential Service queue parameter is ( - ), and we use symbol WS for time in system, Hence, P ( WS > t) = This is the result you will have to use for the case. When , and t is given, we can find the probability of waiting in the system for more than time t, using algebra. When , and probability is given we can inverse the function (taking natural logarithms ln and find t or simply use MS Excel function) Just to show how the formula for Average time in system WS has come, WS is nothing but expectation of W or E[W]. We have seen in exponential probability distribution leant in SMRM (I hope some of you
could at least find it now in you book Statistical Tools) that, for exponential distribution with parameter (rate), the Expectation is, E[X] = So obviously for our Poisson Arrival- Exponential Service queue, at an exponential rate ( - ) E[X] = Now for probability of waiting in queue, consider customer arrives and finds that there are n customers in queue ahead of him (including one being served). The customer will have to wait through n exponential services times until his turn comes. This could be found out from summation. Let T1, T2, .etc are exponential random service times for first, second etc customers in the queue, the total time the new customer will spend in queue (including service) is Sn+1 = T1 + T2 +..+ Tn+1 However, we may find 0, 1, 2, .. customers actually in queue with probabilities P0, P1, P2, . Hence using expectation formula, P ( WS > t) = In case of probability for waiting in queue (other than service time), we similarly write P ( Wq > t) = Note summation is taken from 1 to infinity because for standing in queue there must be at least one customer in the system being served. Using formula Pn= (1- ) P ( Wq > t) = Replacing n by n+1 P ( Wq > t) = P ( Wq > t) = This is another formula that you will have to use for the case. This gives probability of waiting in queue for time more than t. =
n
for Poisson Arrival- Exponential Service queue, =

Background Notes For Solving Case

Uploaded by

Copyright:

Available Formats

Background Notes For Solving Case

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Background Notes For Solving Case

Uploaded by

Copyright:

Available Formats

Background Notes for Solving the Case Queuing Quandary Queuing theory is one of the quantitative techniques that

for Poisson Arrival- Exponential Service queue, =

You might also like