comp.lang.ada
 help / color / mirror / Atom feed
From: "Arthur Schwarz" <aschwarz@acm.org>
Subject: Re: Bus Schedules
Date: 2000/01/17
Date: 2000-01-17T00:00:00+00:00	[thread overview]
Message-ID: <s87jcq57eh638@corp.supernews.com> (raw)
In-Reply-To: 85vaqu$b2a$1@nnrp1.deja.com


"Ted Dennison" <dennison@telepath.com> wrote in message
news:85vaqu$b2a$1@nnrp1.deja.com...
> In article <s83og4rbeh665@corp.supernews.com>,
>   "Arthur Schwarz" <aschwarz@acm.org> wrote:
> > Anyone know any Ada software which can calculate a
> > bus schedule? This includes major/minor frame calculations
> > and so on.
>
> That sounds like it would make a really good programming assignment for
> students.
>
> --
> T.E.D.
>
> http://www.telepath.com/~dennison/Ted/TED.html
>

Your probably right. I just happen to be the student :-(

I've seen 1553-bus I/F source code on one of the public sites (PAL?)
and thought to save some work by getting software for fixed
scheduling. Maybe not.

Well, maybe you can help me with the algorithm. Some of the
requirements (or objectives) are listed below. I've thought
about using dynamic programming (for some), maybe Integer
programming, and certainly a branch and bound algorithm should
work since the simplist form of the problem is the same as
the 0-1 knapsack. Any ideas?

The objectives are (in what I think is increasing difficulty):
[1] Variable size Minor/Major Frames. All messages are untagged,
    message slots within a Minor Frame are fixed, and Minor Frames
    dense (no slack bits). n-Minor Frames / Major Frame.
[2] Fixed sized Minor Frames. Packing to achieve message
    repetition rates, Minor Frame size optimized for throughput.
    All Minor Frames fixed size and slack bits minimized.
[3] Tagged messages. Minor Frames contain a fixed number of
    messages slots of fixed size. Each message slot in a Minor
    Frame has a set of possible messages which can 'fit'.
    Each message distinguished by their tag. The constituent
    message sets and message set distributions per Major Frame
    needs to be defined.
[4] System performance optimization. Given a command (out message)
    and a response (in message) place responses in in-messages in
    such a way that total system latency is reduced.

And of course there are binary channels, and turn-around latency
issues, inter-frame latencies, error frames, and a host of other
issues TBD after the above. Any ideas?

Thanks Ted.

art
schwarza@gdls.com






  reply	other threads:[~2000-01-17  0:00 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2000-01-16  0:00 Bus Schedules Arthur Schwarz
2000-01-17  0:00 ` Ted Dennison
2000-01-17  0:00   ` Arthur Schwarz [this message]
2000-01-18  0:00     ` Ted Dennison
2000-01-18  0:00   ` Charles Hixson
replies disabled

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox