|
|
 | | From: | John Sturgul | | Subject: | Train simulations | | Date: | Sat, 06 Nov 2004 12:28:54 -0600 |
|
|
 | This is for Kamlish who is interested in modeling train.
I have used GPSS/H and PROOF to model some very complex train situations.
One example had multiple train (up to 20) with up to 1,500 cars. The
trains were continually adding cars, leaving them off, backing up, moving
forward, etc. GPSS/H has the neat ATTACH and DETACH Blocks that allow
this to be easily done. The model was for a system of 14 underground coal
mines working on two different levels in the NW of Spain.
I have also used GPSS/H and PROOF for a train system for an underground
platinum mine in South Africa. There was no problem in doing this.
John Sturgul
Univ. of Idaho
|
|
| | From: | Ron Reagan | | Subject: | Re: Train simulations | | Date: | Sun, 05 Dec 2004 17:39:24 -0600 |
|
|
| Dear comp.simulation folks,
I am been interested in simulation since 1985 and I taught a course on
the subject for 8 years. I am also an avid model and prototype rail fan.
During those years I wrote RVL, a program for fitting distributions to
data. It was presented at the Winter Simulation Conference 1992. The
presented paper is available from ACM if you search for "Ronald S. Reagan"
on Google.
In 1991 I obtained 8 days of train arrival data at Rosenberg, TX. I was
curious about the best time(s) of day to watch trains on the Southern
Pacific. The resulting histogram and 4 best distribution fits are available
at the
URL
http://photos.groups.yahoo.com/group/ggcrf/vwp?.dir=/RR+Ron%27s+Pics&.dnm=Weekdays.jpg&.src=gr&.view=t&.hires=tattached
The Linear6 distribution is my invention. Note there are 3 best
times spaced about 8 hrs apart. I have no explanation except that input
processes at other points are kind of in phase.
I hope to read and regularly contribute to this group.
Cheers,
Ron
Ronald S. Reagan
"John Sturgul" wrote in message
news:Pine.GSO.4.56.0411061027210.11803@tornado.csrv.uidaho.edu...
> This is for Kamlish who is interested in modeling train.
>
> I have used GPSS/H and PROOF to model some very complex train situations.
> One example had multiple train (up to 20) with up to 1,500 cars. The
> trains were continually adding cars, leaving them off, backing up, moving
> forward, etc. GPSS/H has the neat ATTACH and DETACH Blocks that allow
> this to be easily done. The model was for a system of 14 underground coal
> mines working on two different levels in the NW of Spain.
>
> I have also used GPSS/H and PROOF for a train system for an underground
> platinum mine in South Africa. There was no problem in doing this.
>
> John Sturgul
> Univ. of Idaho
>
|
|
| | From: | Ron Reagan | | Subject: | Re: Train simulations | | Date: | Mon, 06 Dec 2004 08:50:19 -0600 |
|
|
| Hi,
I live 0.4 miles from the CSX RR near Gulfport, MS. The line from New
Orleans to Mobile is mostly tangent with little or no gradient. My modeling
(N and HO scales) covers the time around 1980 when CSX was in a major merger
mode. This allows me to run locomotive lashups with units from many lines.
After posting my re:Train Simulation note I had a revelation. That is
simulation of a linear portion of a more complicated railroad system should
be quite feasible if the 2 (or more) random variables describing the trains
arriving at each end are available by hypothesis or measurement.
Models of sidings, spurs, draw bridges, train speeds and progression by
milepost, etc. would be needed for the middle portion.
My next step is to separate the E and W bound trains in the Rosenberg, TX
data and fit distributions for each direction. This exercise may help in
explaining the 3 peaks in the density for the combined directions.
I have no budget so I am fiddling with the EZSIM program from Behrokh
Khoshnevis. Does anyone know of any inexpensive or free simulation codes. I
guess there is always the one in Law and Kelton.
Thanks for the headsup on the Simulation MultiConference in New Orleans. NO
is only 90 miles from here so I will try to make it. I alo contacted one of
my former students who is in Medical Research.
Cheers on a foggy morning,
Ron
"Ron Reagan" wrote in message
news:cp00dn$4ck$1@news.datasync.com...
> Dear comp.simulation folks,
> I am been interested in simulation since 1985 and I taught a course on
> the subject for 8 years. I am also an avid model and prototype rail fan.
> During those years I wrote RVL, a program for fitting distributions to
> data. It was presented at the Winter Simulation Conference 1992. The
> presented paper is available from ACM if you search for "Ronald S. Reagan"
> on Google.
> In 1991 I obtained 8 days of train arrival data at Rosenberg, TX. I was
> curious about the best time(s) of day to watch trains on the Southern
> Pacific. The resulting histogram and 4 best distribution fits are
> available at the
> URL
> http://photos.groups.yahoo.com/group/ggcrf/vwp?.dir=/RR+Ron%27s+Pics&.dnm=Weekdays.jpg&.src=gr&.view=t&.hires=tattached
> The Linear6 distribution is my invention. Note there are 3 best
> times spaced about 8 hrs apart. I have no explanation except that input
> processes at other points are kind of in phase.
> I hope to read and regularly contribute to this group.
> Cheers,
> Ron
> Ronald S. Reagan
>
> "John Sturgul" wrote in message
> news:Pine.GSO.4.56.0411061027210.11803@tornado.csrv.uidaho.edu...
>> This is for Kamlish who is interested in modeling train.
>>
>> I have used GPSS/H and PROOF to model some very complex train situations.
>> One example had multiple train (up to 20) with up to 1,500 cars. The
>> trains were continually adding cars, leaving them off, backing up, moving
>> forward, etc. GPSS/H has the neat ATTACH and DETACH Blocks that allow
>> this to be easily done. The model was for a system of 14 underground
>> coal
>> mines working on two different levels in the NW of Spain.
>>
>> I have also used GPSS/H and PROOF for a train system for an underground
>> platinum mine in South Africa. There was no problem in doing this.
>>
>> John Sturgul
>> Univ. of Idaho
>>
>
|
|
| | From: | Kamlesh Somani | | Subject: | Re: Train simulations | | Date: | Sun, 07 Nov 2004 12:15:55 -0600 |
|
|
| Thanks. But we have to use either Arena or Siman.
Regards,
Kamalesh
John Sturgul wrote in message news:...
> This is for Kamlish who is interested in modeling train.
>
> I have used GPSS/H and PROOF to model some very complex train situations.
> One example had multiple train (up to 20) with up to 1,500 cars. The
> trains were continually adding cars, leaving them off, backing up, moving
> forward, etc. GPSS/H has the neat ATTACH and DETACH Blocks that allow
> this to be easily done. The model was for a system of 14 underground coal
> mines working on two different levels in the NW of Spain.
>
> I have also used GPSS/H and PROOF for a train system for an underground
> platinum mine in South Africa. There was no problem in doing this.
>
> John Sturgul
> Univ. of Idaho
|
|
| | From: | Bill Harris | | Subject: | Re: Train simulations | | Date: | Mon, 06 Dec 2004 09:30:42 -0600 |
|
|
| "Ron Reagan" writes:
> I have no budget so I am fiddling with the EZSIM program from Behrokh
> Khoshnevis. Does anyone know of any inexpensive or free simulation codes. I
> guess there is always the one in Law and Kelton.
>> I am been interested in simulation since 1985 and I taught a course on
>> the subject for 8 years. I am also an avid model and prototype rail fan.
Ron,
Welcome to this (often somewhat quiet) group! I, too, am a fan of
simulation and and also of railroads.
As for free tools, you might check out
http://www.cis.ufl.edu/~fishwick/simpack/simpack.html. It has the bits
and pieces for many approaches to simulation. There used to be an FAQ
for this group that included a list of tools "every modeler should
have," and SimPack was on that list.
You might also check out CPN Tools
(http://wiki.daimi.au.dk/cpntools/cpntools.wiki). There's a paper
(which I have yet to read) listed at
http://www.daimi.au.dk/CPnets/intro/4.html on "Technical Issues in
Modelling the European Train Control System," as well as a couple of
papers dealing with the modeling of model railways.
Or there's always
http://www.camcentral.com/camviewer.asp?script=location&item=Germany&page=8&id=140
for just doing it (English or German, your choice)! :-)
Bill
--
Bill Harris
Facilitated Systems
http://facilitatedsystems.com/
|
|
|
