[comp.sources.wanted] PD Discrete-event simulation software

leeke@cascade.UUCP (02/27/87)

Does anyone know of a PD discrete-event simulation software package based on
a standard programming language that is available? Preferebly event-oriented
and not process or activity oriented.  Pascal or C prefered, but I'll be happy
with most anything.

Thanks for your help,

Steve Leeke
-- 
Steven D. Leeke, Center for Integrated Systems, Stanford University
    {ucbvax,decvax}!decwrl!glacier!leeke, leeke@cascade.stanford.edu

"I suppose they don't use money in the 23rd century?"

brachman@ubc-cs.UUCP (Barry Brachman) (03/08/87)

In article <284@cascade.STANFORD.EDU> leeke@cascade.UUCP (Steven D. Leeke)
writes:
>Does anyone know of a PD discrete-event simulation software package based on
>a standard programming language that is available? Preferebly event-oriented
>and not process or activity oriented.  Pascal or C prefered, but I'll be happy
>with most anything.
>
>Thanks for your help,
>
>Steven D. Leeke, Center for Integrated Systems, Stanford University
>    {ucbvax,decvax}!decwrl!glacier!leeke, leeke@cascade.stanford.edu

Over the past year or two I've been working intermittently on a discrete event
simulator that is based on the Xinu operating system ("Operating System Design,
the Xinu Approach", Douglas Comer, Prentice-Hall, 1984).  It runs on top of Unix
4.2/4.3BSD on a Vax or Sun and consists of a library of functions that manage
processes and queues, record statistics, and generate variates from several
different distributions.  The simulator is written in C as is the user's code.
All I/O is done using Unix routines (no Xinu I/O stuff is used).
(BTW, since almost everything is in C it should be straightforward to port to
non-Unix machines.)

A simulation program consists of one or more Xinu processes. These processes
normally represent servers but they need not.  A server usually has one or more
queues associated with it.  A single server simulation might have a
'producer' process simply put items on a queue while a 'consumer' process
removes items from the queue.  When the consumer tries to remove an item from
an empty queue the process blocks.  Eventually the producer should
enqueue an item after which the consumer will unblock and continue.
Arbitrary data items are put on the queue; the queue management routines simply
copy the data.  In the producer/consumer example, items placed on the queue
might include the value of the simulation clock at the time the item was
enqueued and the amount of service time the item requires (i.e., how long the
consumer process should wait once the item has been dequeued).

Anyhow, there are two problems:
	1) I'd have to get Doug Comer's permission.  The simulator does not
	   use any code derived from Unix sources so I don't think there would
	   be a problem.
	2) The simulator has not been tested extensively.  You'd pretty much be
	   on your own, although I'd be very interested in suggestions, bug
	   reports, etc.

Also: I'd like to know if there is sufficient interest out there before I go
to the trouble of making up a distribution.

Any interest?
-----
Barry Brachman		 | {ihnp4!alberta,uw-beaver}!ubc-vision!ubc-cs!brachman
Dept. of Computer Science| brachman@cs.ubc.cdn
Univ. of British Columbia| brachman%ubc.csnet@csnet-relay.arpa
Vancouver, B.C. V6T 1W5  | brachman@ubc.csnet
(604) 228-4327

mgb@macs.UUCP (03/11/87)

In article <893@ubc-cs.UUCP>, brachman@ubc-cs.UUCP (Barry Brachman) writes:


> Over the past year or two I've been working intermittently on a discrete event
> simulator that is based on the Xinu operating system ("Operating System Design,
> the Xinu Approach", Douglas Comer, Prentice-Hall, 1984).  It runs on top of Unix
> 4.2/4.3BSD on a Vax or Sun and consists of a library of functions that manage
> processes and queues, record statistics, and generate variates from several
> different distributions.  The simulator is written in C as is the user's code.
> 
> 
> Also: I'd like to know if there is sufficient interest out there before I go
> to the trouble of making up a distribution.
> 
> Any interest?


   I'd be interested in seeing this posted....


   Michel Branton
   ...ihnp4!codas!ki4pv!macs!mgb