Canadian Operational
Research Society
Calgary Section
http://www.corscalgary.org/
PROFESSIONAL DEVELOPMENT
SEMINAR
When: Noon to 1:30 PM,
Friday, April 16, 2004
Room 217
TransCanada Tower
450 - 1 Street SW
(See attached map)
Paul Rogers, P.Eng., Ph.D.
TOPIC
Abstract:
The topic of "workload control" (or input
control) in manufacturing systems has been growing in importance with
manufacturers trying to achieve both high levels of productivity from their
resources and high levels of responsiveness to customer demands. In make-to-order
production environments facing uncertain demand, it is especially challenging
to meet both responsiveness requirements (low flow times and good on-time
delivery performance) and efficiency requirements (high utilization of
production resources). Two critical aspects of workload control in this
make-to-order context include the making of decisions concerning which customer
orders should be accepted and when should accepted orders be released to the
shop floor. In cases where demand is highly variable, accepting all orders can
cause due date performance to deteriorate to unacceptable levels, so the
ability to judiciously reject customer orders represents one way of keeping due
date performance acceptable (hopefully without having to reject too many
orders). Although there has been substantial discrete-event simulation work
reporting on the performance of various alternative
"acceptance/rejection" policies and "order release"
policies, few general conclusions can be drawn from this work, and little real
insight appears to have been gained.
This presentation will focus on work that has been
completed at the University of Calgary on the performance of optimal order
rejection and delayed order release policies for the well-known M/M/1 queuing
system (exponentially distributed job interarrival times and exponentially
distributed service times for a single server), when there are linear earliness
and tardiness penalties for jobs not completed exactly on their due time. Three
different policies will be described during the presentation, the first which
involves just a delayed order release mechanism, the second which involves just
an order rejection mechanism, and the third which involves both mechanisms in
concert. Results presented will include the determination of optimal control
parameters for the three policies and how the optimal policies perform under a
broad range of conditions. The models and results described in the presentation
were developed by the presenter and Mr. Yannai Zev Romer Segal during the
latter's research towards the M.Sc. degree at the University of Calgary under
the supervision of the presenter. For those seeking more detail than can be
covered in a short presentation, full details on the three models and their
results can be found in Mr. Segal's M.Sc. thesis, available in the University
of Calgary library.
About the Speaker
PAUL ROGERS is an Associate Professor in the Department of Mechanical and
Manufacturing Engineering at the University of Calgary. He holds Ph.D. and
M.Eng. degrees from Cambridge University in England where he was involved in
industry-funded research within the Manufacturing Engineering Group on the
design and operation of flexible machining systems. Before joining the
University of Calgary in 1991, he spent two years as a visiting professor of
Industrial Engineering at Purdue University in the USA. His current research
and teaching interests include production planning and control systems,
manufacturing strategy, distributed and dynamic scheduling, and models for the
design, analysis and control of manufacturing systems. He is a Professional
Engineer registered with the Association of Professional Engineers, Geologists,
and Geophysicists of Alberta (APEGGA), a member of IIE, APICS, and INFORMS, and
serves on the Editorial Board of the International Journal of Computer
Integrated Manufacturing. He can be contacted by email at <rogers@ucalgary.ca>.