Page 1 of 1

SD Modelling Time frame

Posted: Thu Apr 13, 2000 12:55 pm
by "Achara Changsiri"
I am looking at applying SD to materials recycling and management and found
some interesting behaviour to the system in the short run.

I am trying to determine the validity of using SD modelling, when examining
systems behaviour in the short term (before the system reaches a steady
state). I understand and agree with the use of SD to determine the systems
trends/behaviour in the long-run, under different scenario.

What about the use of SD to examine results of the first 10% of the
simulation run ? (Assuming that at 100% simulation run the system reaches
some sort of a steady state/pattern)

What major problems do you foresee with the interpretation of the short-term
results when compared to examining the entire run?

Achara Changsiri
School of Civil and Environmental Engineering
The University of New South Wales
Sydney NSW 2052

Ph: (02) 9385 5064
Fax: (02) 9385 6139
Email: achara@civeng.unsw.edu.au

SD Modelling Time frame

Posted: Fri Oct 13, 2000 9:23 am
by John Sterman
Achara Changsiri wonders whether SD mdoels can address the transient
behavior of a system and asks about SD models of recycling.

System dynamics models are designed as behavioral, dynamic, disequilibrium
models - the main value is their ability to capture the nonequilibrium
(transient) behavior of a system. Rather than envisioning the evolution of
a system as the smooth transition from one near-equilibrium state to
another, or the optimal trajectory of the system (as in many economic
models), we seek to capture the behavioral decision processes used by the
actors in the system through which they respond to the pressures that arise
from imbalances. These decisions then feed back to alter the state of the
system. A good SD model endogenously generates the disequilibrium dynamics
of the system. Whether it reaches a stable long run equilibrium or not is
then an emergent property of the model and not a built-in assumption or
constraint imposed by the modeling method.

On recycling, see the Ph.D. thesis of Hank Taylor, from the MIT SD group:

Taylor, Henry 1999, Modeling Paper Flows and Recycling in the US
Macroeconomy, Ph.D., Dept. of Civil and Environmental Engineering, MIT,
Cambridge MA 02139.

This work develops a model of the US pulp and paper industry with a focus
on the development of recycling. It endogenously represents paper
production, investment, capacity, inventories, and prices, and includes the
collection and recycling of paper endogenously as well. The model is
particularly strong on the short-term effects of policies designed to
promote higher fiber recycling rates.

Hank can be reached at <hank.taylor@strategicsimulation.com>

You may also be interested in the work of Pavel Zamudio Ramirez, whose MIT
masters thesis examined automobile recycling. A summary of this work is
availabe in chapter 6 of my book, Business Dynamics (see
http://www.mhhe.com/sterman). The thesis is

Zamudio-Ramirez, Pavel (1996) The Economics of Automobile Recycling. MS
Thesis, MIT Leaders for Manufacturing Program, Cambridge MA 02139.

This model also illustrates some interesting short and medium term impacts
of policies designed to promote part and materials recycling from
automobiles.

Pavel can be reached at Pzamudio@monitor.com.

John Sterman
jsterman@mit.edu

SD Modelling Time frame

Posted: Fri Oct 13, 2000 9:23 am
by John Sterman
Achara Changsiri wonders whether SD mdoels can address the transient
behavior of a system and asks about SD models of recycling.

System dynamics models are designed as behavioral, dynamic, disequilibrium
models - the main value is their ability to capture the nonequilibrium
(transient) behavior of a system. Rather than envisioning the evolution of
a system as the smooth transition from one near-equilibrium state to
another, or the optimal trajectory of the system (as in many economic
models), we seek to capture the behavioral decision processes used by the
actors in the system through which they respond to the pressures that arise
from imbalances. These decisions then feed back to alter the state of the
system. A good SD model endogenously generates the disequilibrium dynamics
of the system. Whether it reaches a stable long run equilibrium or not is
then an emergent property of the model and not a built-in assumption or
constraint imposed by the modeling method.

On recycling, see the Ph.D. thesis of Hank Taylor, from the MIT SD group:

Taylor, Henry 1999, Modeling Paper Flows and Recycling in the US
Macroeconomy, Ph.D., Dept. of Civil and Environmental Engineering, MIT,
Cambridge MA 02139.

This work develops a model of the US pulp and paper industry with a focus
on the development of recycling. It endogenously represents paper
production, investment, capacity, inventories, and prices, and includes the
collection and recycling of paper endogenously as well. The model is
particularly strong on the short-term effects of policies designed to
promote higher fiber recycling rates.

Hank can be reached at <hank.taylor@strategicsimulation.com>

You may also be interested in the work of Pavel Zamudio Ramirez, whose MIT
masters thesis examined automobile recycling. A summary of this work is
availabe in chapter 6 of my book, Business Dynamics (see
http://www.mhhe.com/sterman). The thesis is

Zamudio-Ramirez, Pavel (1996) The Economics of Automobile Recycling. MS
Thesis, MIT Leaders for Manufacturing Program, Cambridge MA 02139.

This model also illustrates some interesting short and medium term impacts
of policies designed to promote part and materials recycling from
automobiles.

Pavel can be reached at Pzamudio@monitor.com.

John Sterman
jsterman@mit.edu