Instability in Pull Systems (flexible kanbans)

This forum contains all archives from the SD Mailing list (go to http://www.systemdynamics.org/forum/ for more information). This is here as a read-only resource, please post any SD related questions to the SD Discussion forum.
Locked
"George W. L. Sousa"
Junior Member
Posts: 2
Joined: Fri Mar 29, 2002 3:39 am

Instability in Pull Systems (flexible kanbans)

Post by "George W. L. Sousa" »

Dear SDS friends,

It has been argued in the Production Planning and Control literature that
"pull" (or Just-in-Time type systems) are not suitable for operation in
environments where demand is unstable or volatile.
Has anybody been involved with supply chain instability studies in this
area? Or can anyone point me to previous related studies?

I have conducted a comprehensive search in past e-mails exchanged in this
listserv and it indicates that discussions around this topic did not take
place.

I am a Ph.D. student working on investigating the impacts of "flexible
kanban systems" on the on-time-delivery performance of pull systems from a
system dynamics point of view. In this case, a flexible kanban system
differs from a "traditional kanban system" (as originally proposed at
Toyota) in the sense that the number of kanban cards regulating the levels
of work in process inventory is allowed to change over time. Traditional
pull systems using kanbans tend to keep the number of cards constant during
production cycles and only change them in the event of major changes in
production capacity.

Thanks you for your attention!
Sincerely,
George
___________________________________________________________
George W. L. Sousa, Ph.D. Candidate & Graduate Research Assistant
Grado Department of Industrial and Systems Engineering
Enterprise Engineering Research Lab
302B Whittemore Hall, Virginia Tech
Blacksburg, VA 24061-0118 USA
phone: +1(540)231-6920
email:
gsousa@vt.edu
http://www.ise.vt.edu/eerl_lab.htm
John Sterman
Senior Member
Posts: 117
Joined: Fri Mar 29, 2002 3:39 am

Instability in Pull Systems (flexible kanbans)

Post by John Sterman »

Paulo Goncalves has been working on exactly the issue of instability
in pull systems, with Intel as a case study. Paulo did his doctoral
work here at the MIT System Dynamics Group and is now on the faculty
of the University of Miami. He can be reached at Paulo Goncalves
<paulog@miami.edu>.

John Sterman
From: John Sterman <jsterman@MIT.EDU>
Paulo Goncalves
Junior Member
Posts: 2
Joined: Fri Mar 29, 2002 3:39 am

Instability in Pull Systems (flexible kanbans)

Post by Paulo Goncalves »

Dear George,

Jim Hines, Charlie Lertpattarapong, and I have been developing models
dealing with supply chain instability applied to Intel Corporation. Since
Intel is aware of the costs associated with operating a pull systems when
demand is unstable, they run a hybrid "push-pull" system instead. That is,
the first stages of production are characterized by a "push"
(make-to-stock) system and the final stages by a "pull" system.

We adopted a (relatively) high level of aggregation in our modeling, so we
did not address directly the issues associated with traditional versus
flexible kanban systems. We did assume constant inventory coverage,
however, which I think would map more closely a traditional kanban system.

Our research suggests that supply chain instability can shift the mode of
operation from the designed push-pull system to a pure push, causing a
reduction in performance. A working paper summarizing this work (The Impact
of Shortages on Push-Pull Production Systems) is part of my Ph.D.
dissertation and it is available upon request.

Gonçalves, Paulo. 2003. "Demand Bubbles and Phantom Orders in Supply
Chains." Unpublished Ph.D. Disssertation, Sloan School of Management, MIT,
Cambridge, MA.

Paulo
paulog@miami.edu
Khalid Saeed
Senior Member
Posts: 79
Joined: Fri Mar 29, 2002 3:39 am

Instability in Pull Systems (flexible kanbans)

Post by Khalid Saeed »

One of my students did a masters thesis on this subject a long time ago
(1986) at the Asian Institute of Technology. His thesis can be obtained
from AIT libraries. The author is Tanvir Ahmad and the thesis number is:
IE-86-3. His thesis title is:

An assessment of the performance of just-in time production system: A
system dynamics approach

Khalid Saeed
From: Khalid Saeed <saeed@WPI.EDU>


Khalid Saeed
Professor and Department Head
Social Science and Policy Studies
Worcester Polytechnic Institute
100 Institute Road
Worcester, MA 01609, USA

Ph: 508-831-5563; fax: 508-831-5896
email: saeed@wpi.edu

SSPS Dept: http://www.wpi.edu/+SSPS
"Jim Hines"
Senior Member
Posts: 88
Joined: Fri Mar 29, 2002 3:39 am

Instability in Pull Systems (flexible kanbans)

Post by "Jim Hines" »

Our (Paulo Goncalves, Charlie Lerpattarapongs, and my) work suggests
that pull components in a supply chain add significantly to stability.
The reason is that a pull component creates a very fast negative
feedback loop. (Paulos thesis is the best place to see the work ---
hes provided a citation in his response).

A number of years ago, working with John Morecroft I did some modeling
of Kanban systems. Kanban also increased stability significantly
because the number of "slots" for inventory is fixed. Your "flexible
Kanban" removes this feature, so I would guess that it would also remove
at least some of the stabilizing effect of Kanban.

Jim Hines
MIT and WPI
jhines@mit.edu
jhines@wpi.edu
"Jack Ring"
Junior Member
Posts: 12
Joined: Fri Mar 29, 2002 3:39 am

Instability in Pull Systems (flexible kanbans)

Post by "Jack Ring" »

One can certainly model, simulate and discern the system stability limits as
a function of a) buffer size (push systems which operate in a "just in case"
mode) and b) latency (time delays in either a push or pull mode of prompting
system activity). Perhaps this should be seen as suppression of the
whiplash effect likelihood vs. cost of lost or useless production.

Must not an investigation of system stability vs. Pull include notions of
waste (net cost of goods sold) and decision error (or fidelity of the mental
model of the system)vs. the real state of the system)?

In the limit the system can remain stable in all sorts of scenarios simply
by doing nothing at all times. The whole issue is System effectiveness
(getting the most useful done) vs. Pull or Push vs. dynamics of situation.

From: "Jack Ring" <jring@amug.org>
Locked