Hi Robert,
I started a jobshop project in Oz 1 year and a half ago, and now I have
some customers in beta-version.
In my experience most of the time of the project was spent in the user
interface and database i/o trying to simplify the life of the user (70% vs.
30%).
So my suggest is to concentrate first on the user interface and basic
requirements simulating by hand the work of the scheduler and, after all is
well understood you can play with the scheduling problems.
Because I am forced to work in WINXX I used VB as front-end language and OZ
as numerical server with a socket communication with a simple ad-hoc
protocol.
Among many problems I faced in the building of the scheduler I mention only
the most important:
Numbers - towsends of tasks had to be scheduled in an optimal or suboptimal
mode
Machines - as the problem is a "flexshop" problem many alternative machines
per task exist so a preliminar distribution on those was necessary
Calendars - machines have their own calendar (vacancies,shifts,breaks) so
the available time has to be harmonized among all
Incremental - the schedule had to be incremental that is a previous
situation, possibly in conflict with global constraints (e.g. manually
modified), had to be considered
but, until now, no lot-size constraints neither order management I tackled
because an already working program is feeding the scheduler with the
inventory, leadtimes, machines, tools, setups, etc.. datas from the real
world so what we can do is eventually collaborate on a discussion about the
whole Material Requirements Planning.
Let me know your opinion,
Adriano
-
Please send submissions to users@mozart-oz.org
and administriva mail to users-request@mozart-oz.org.
The Mozart Oz web site is at http://www.mozart-oz.org/.