Brian's Digest: Simulation
1996
SCI.OP-RESEARCH Digest Mon, 25 Aug 97 Volume 4: Issue 34

Date: 19 Aug 1997 21:33:40 GMT
From: tocon23458@aol.com (TOcon23458)
Subject: Ehrlang C- Stochastic Modeling
Can anyone provide me with the equation which calculates Ehrlang C tables for call-centre modeling purposes.

I'm trying to plug one straight into a service model that I'm developing for my employer in the UK and cannot find it in any text.

Thanks in advance.

Tracy O'Connor

Date: Wed, 20 Aug 1997 09:55:15 +0000
From: Alfred Wassermann <Alfred.Wassermann@uni-bayreuth.de>
Subject: Ehrlang C- Stochastic Modeling
See http://pass.maths.org.uk

Alfred

Date: Mon, 25 Aug 1997 16:29:27 GMT
From: charp@ford.com (Charley Harp)
Subject: Ehrlang C- Stochastic Modeling
In article <33FABF03.49B9B63E@uni-bayreuth.de> Alfred Wassermann The probability that a call will be delayed is given by

a^n n --- * --- n! n-a Pr(t>0) = ----------------------- n-1 (a^i) a^n n Sum ----- + --- * --- i=0 i! n! n-1 and the probability of a delay greater than T, as a multiple of h, is given by

Pr(t>T) = Pr(t>0) * exp(-(n-a)*T). where a = Total traffic, in Erlangs, n = Number of circuits in the group, and h = Average call length, in hours. (from h = a / calls-per-hour) Also, h h D1 = P(t>0) * ---, and D2 = --- n-a n-a give D1 = average call delay for all calls, in multiples of h, and D2 = average call delay for calls delayed, in multiples of h.

You can compare your results with the this table:

P(t>T) for T = a n P(t>0) D1 D2 0.2 0.4 0.8 1.0 2.0 --- --- ------ ----- ---- ----- ----- ----- ----- ----- 0.1 1 0.1000 0.111 1.11 0.084 0.070 0.049 0.041 0.017 0.1 2 0.0048 0.003 0.53 0.003 0.002 0.001 0.001 - 0.1 3 0.0002 - 0.34 - - - - - 0.2 1 0.2000 0.250 1.25 0.170 0.145 0.105 0.090 0.040 0.2 2 0.0182 0.010 0.56 0.013 0.009 0.004 0.003 - 0.2 3 0.0012 - 0.36 0.001 - - - - 0.2 4 0.0001 - 0.26 - - - - - 0.5 1 0.5000 1.000 2.00 0.452 0.409 0.335 0.303 0.184 0.5 2 0.1000 0.067 0.67 0.074 0.055 0.030 0.022 0.005 0.5 3 0.0152 0.006 0.40 0.009 0.006 0.002 0.001 - 0.5 4 0.0018 0.001 0.29 0.001 - - - - 0.5 5 0.0002 - 0.22 - - - - - 1.0 2 0.3333 0.333 1.00 0.273 0.223 0.150 0.123 0.045 1.0 3 0.0909 0.045 0.50 0.061 0.041 0.018 0.012 0.002 1.0 4 0.0204 0.007 0.33 0.011 0.006 0.002 0.001 - 1.0 5 0.0038 0.001 0.25 0.002 0.001 - - - 1.0 6 0.0006 - 0.20 - - - - - 1.0 7 0.0001 - 0.17 - - - - - Good luck! ---------------------------------------------------------------------- Charley Harp, N8MQL FBP-425 charp@ford.com Operations Research Dept. 555 Republic Dr. V: (313) 845-5873 Ford Motor Company Allen Park, MI 48101 F: (313) 621-8381 ----------------------------------------------------------------------
SCI.OP-RESEARCH Digest Mon, 28 April 97 Volume 4: Issue 17

Date: Wed, 23 Apr 1997 15:17:31 -0700
From: Jun Yoshioka <junyoshi@fog.net>
Subject: Computer Simulation of Urban Transit System
I am a graduate business student at San Francisco State University, and I have just completed my thesis on a computer simulation of a public transit system.

http://www.rescuemuni.org/yoshioka/38geary.htm

Does anyone know any similar study has been done?

Any comment on my thesis is also welcome!

Thank you

Date: Sat, 18 Jan 1997 22:12:30 +0800
From: Richard LAplace <R.Laplace@202.117.7.150>
Subject: How to evaluate air warfare effectivness?
Hello!

I'm interest in M&S ,you know modeling is very important in computer simulation.Many models have been used in US army .If we only talk about weapon ,it's some easy to study,but there is human involving .How to describe the preference of human in warfare?Espically when focus on air combat,how to evaluate the effectiveness of the human-fighter system?Is there some useful models?

SCI.OP-RESEARCH Digest Mon, 20 Jan 97 Volume 4: Issue 3

Date: 16 Jan 1997 16:14:15 -0500
From: seonho@cs.umd.edu (Seonho Choi)
Subject: Incremental Shortest Path Algorithm
Hi!

I need a help on the following problem.

I want to repeatedly apply all-pair shortest path algorithm on directed graphs. However, each time I apply the algorithm, the graph is different very little from the previous graph. The structure of the graph remains the same and just the weights of few arcs are changed. Definitely, I can apply the all-pair shortest algorithm each time, but, it seems that this approach is very inefficient.

Do you know any incremental algorithm developed for this kind of situation? Any help would be appreciated.

Thanks.

Seonho Choi

e-mail: seonho@cs.umd.edu

1996

SCI.OP-RESEARCH Digest Mon, 18 Dec 95 Volume 2: Issue 51

Date: 11 Dec 1995 19:02:51 GMT
From: tsakiris@pt9225.ped.pto.ford.com (Tony Tsakiris)
Subject: Info on Article: Denver Airport Simulation

: >Sometime between January and November 1994, I recall reading an article : >that described some of the issues with the new Denver Airport project. : >Does anyone know where this article appeared? : >Do you have a copy of the article? : >Do you know of any similar case studies? -- See also an article entitled "Software's Chronic Crisis" by W. Wyat Gibbs in the September 1994 issue of Scientific American.

Anthony Tsakiris

Date: 13 Dec 1995 13:49:18 GMT
From: Joe Deasy <jodeasy@roentgen.bcc.louisville.edu>
Subject: Info on Article: Denver Airport Simulation
You might also be interested to know that Digital was able to get its current lead in microprocessor technology by doing full scale state simulations, following charge, etc., of their alpha processors. The first alpha processor had few enough transistors to be simulated on a 32-bit hardware platform. But once the first 64-bit alpha was built it was used immediately to simulate processors which could not have been simulated in the pre-alpha days. And onwards, etc.

I would think that software simulation of complex interacting mechanisms should be accepted as standard where possible.

--Joe

SCI.OP-RESEARCH Digest Mon, 11 Mar 96 Volume 3: Issue 11

Date: Wed, 6 Mar 1996 12:53:47 GMT
From: rvdham@ect.nl (Ruud van der Ham)
Subject: Algorithm for event chain
For an implemetation of an event driven simulation system, I need to maintain a list of eventtime (along with information on those events). Of I can use a simple linked list for this purpose, but I think there must be more efficient ways to do this (I remember an article somewhere, but cannot locate it anymore),

The algorithm should provide a way of inserting events (according to a 'random' time) and must be able to have fast access to the first event of the list, which will normally be deleted in sync with that. There must be ways to traverse the list and to delete an item somewhere on the list, but this will be rare, so efficiency is not important there. Concluding I need some routines for: