Talk:Queueing model

Looks like this should be merged into queuing theory. Charles Matthews 14:26, 11 December 2005 (UTC)
 * It looks like it has been added. (Queueing theory) John Reed Riley 02:52, 17 March 2006 (UTC)
 * The existing content only talks about the general notation. However, explanation of the classic types of queuing models, such as M/M/1 and M/M/n and how they are analysed is probably needed. This, together with a generalised explanation of how models are formed to reflect reality would distinguish it from queueing theory. -- Cameron Dewe 23:56, 29 July 2006 (UTC)
 * I have now expanded the existing content to a better article. I think that this article could be expanded into a substantial article on its own, so merger into queuing theory may not be appropriate any longer. If anything, the explanation of Kendal notation fits better in this article, now, with only a short mention needed in the queuing theory article. -- Cameron Dewe 23:18, 30 September 2006 (UTC)

Queuing theory vs queuing model
I like the additions to this article and it could become a good article if this keeps up. However, it is mirroring some of queuing theory. How can we best differentiate the two articles. --Richard Clegg 09:42, 3 October 2006 (UTC)


 * I think, some of the content in queueing theory, especially that about notation, would actually be better in queueing model. I have not touched it yet because it is probably worth its own article and the details in queueing theory are very good. However, I see queueing theory as being the over-arching overview article, that just highlights the key information, with queueing model being a specific sub-article that explains theoretical models of queues in far more detail. This is to avoid too much detail being in the top level article. Next, I see specific sub-articles about each classic queueing model class, such as M/M/1, M/G/1, M/M/1/K, M/M/c, M/M/c/K, M/M/c/K/K, etc. It would then be in each of these articles that the mathematics for each queue model could be explained and derived. It is either that or build a huge single monolithic queueing theory article that doesn't flow, is hard to read and very obscure. -- Cameron Dewe 08:42, 5 October 2006 (UTC)


 * Good thinking. OK -- I will make a change to the other article.  Good work on this by the way.  --Richard Clegg 10:24, 5 October 2006 (UTC)


 * I've chosen to forward this page to queueing theory as the content on this page was duplicated elsewhere on articles now linked to from Queueing theory and Template:Queueing theory. I feel queueing theory is the use of queueing models, so the arguments above seem weaker to me now that articles like M/M/1 queue, M/M/c queue, M/G/1 queue etc exist. We need a single article as a hub linking to more specific models. Gareth Jones (talk) 18:09, 17 January 2013 (UTC)

venkatraman is doing a project using witness a software used for Discrete Event Simulation. —Preceding unsigned comment added by 203.129.195.140 (talk) 09:48, 17 March 2008 (UTC)

Duplication of Kendall's notation
This article seems to duplicate much of the content of the Kendall's notation article. It seems like this article should focus more on models than notation, so can we remove most of it from this article, with appropriate linking? 70.250.190.30 (talk) 02:54, 3 October 2010 (UTC)

Analytical and simualation approaches to queueing systems modeling. Limits of the analytic approach
I think it would be worth adding a section on how the alaytic approach compares to the simulation modeling approach, like the one in Queueing theory but more specific. It is important to understand that the formulas giving accurate reults for waiting time, queue length, etc. exist only for a limited munber of cases. For example, there are formulas for M/M/1, M/M/c and M/G/1, but there are none for M/G/c. But, what is even more important from practical viewpoint, real service systems are virtually never approximated well with those classical queueing models. Consider a bank where customers are served by tellers. The service time distribution is nowhere near exponential distribution, the process may include redirection of clients from one teller to another, the tellers may have different skills, may share resources like printer, etc. Adding any of those complications disposes the classical analytical model, and in most cases it is not possible to derive a new set of formulas. (Transaction processing in computer systems, call center operations - they all are full of such critical details.) Simulation, on the contrary, would always give you the answer with predictable efforts on model building. So, is the queueing theory useful at all? Definitely yes. It provides a good foundation for undertanding the general behavior of service systems. For example, the fact that when mean sevice time equals mean arrival time the queue length may grow infintely is an analytica result of great importance. In a simulation model you will observe that queue length goes up and down but you cannot be sure that there is no finite mean value. I have created anApplet that compares the two approaches: and would like to add the link to the wikipedia article, should the commmunity consider the comparison interesting and relevant to the article. Andrei Andreiborshchev (talk) 11:41, 19 July 2011 (UTC)