System Dynamics

www.cise.ufl.edu
What is System Dynamics
• Computer simulation modeling for studying and managing
complex feedback systems, such as business and other social
systems
• System:
• In general, a collection of interacting elements that function together for
some purpose
• Here, feedback is the differentiating descriptor
• Properties of dynamic problems
• Contain quantities that vary over time
• Variability can be described causally
• Important causal influences can be contained within a closed system of
feedback loops
History
• Cybernetics (Wiener, 1948): the study of how biological,
engineering, social, and economic systems are controlled and
regulated
• Industrial Dynamics (Forrester, 1961): applied principles of
cybernetics to industrial systems
• System Dynamics: Forrester’s work has been broadened to
include other social and economic systems
• Relying on computer, System Dynamics provides a framework in
which to apply the idea of systems theory to social and economic
problems
Ide-ide (konsep-konsep) yang menjadi dasar dalam metodologi
system dynamics diperoleh dari:

 teknik pengendalian (control engineering, the

concepts of feedback and system self-
regulation);
 sibernetika (cybernetics, the nature of
information and
its role in control systems); dan
 teori organisasi (organizational theory, the
structure
of human organizations and the forms of human
decision
making).
Peran beberapa bidang (field) dalam metodologi
system dynamics

 Feedback
theory and  Computer
simulatio
 Cybernetics
n

 Principles  Principles  Low-cost

of selecting of computation
information structure

 Dynamic
 MODEL behavior and
improvement of
policies
 Traditional  Information,
management
 experience,
and political
leadership  judgment
 Mental
Modeling

Conceptual
Modeling &
Computer Modeling
System Dynamics Modeling Tolls
• Mental Modeling • Computer Modeling
• Interview • Powersim
• Focus Group • Stella
Discussion • Isee
• Literature Study • Vensim
• Observation • Dynamo
• Conceptual Modeling • Simtegra
• Sub System Diagram • etc
• Boundary Chart
Diagram
• Causal Loop Diagram
• Flow Diagram
System Dynamics Modeling
• Identify a problem
• Develop a dynamic hypothesis explaining the cause of the
problem
• Create a basic structure of a causal graph
• Augment the causal graph with more information
• Convert the augmented causal graph to a System Dynamics flow
graph
• Translate a System Dynamics flow graph into DYNAMO programs
or equations
Critical Aspects
• Thinking in terms of cause-and-effect relationships
• Focusing on the feedback linkages among components of a
system
• Determining the appropriate boundaries for defining what is
to be included within a system
Understand Cause & Effect
• Causal thinking is the key to organizing ideas in a system
dynamics study
• Instead of ‘cause’, ‘affect’ or ‘influence’ can be used to describe
the related components in the system
• Some are logical (e.g. physics)
• Food intake weight
• Money  happiness
• Fire  smoke
• Some are not (e.g. sociology, economics)
• Use of seatbelts  reduced highway fatalities
• Shortened daylight hours  increased suicide rates
Feedback
• Thinking in terms of “cause and effect” is not enough
• ocean  evaporation  cloud  rain  ocean  …
• Feedback: an initial cause ripples through a chain of
causation ultimately to re-affect itself
• Search to identify closed, causal feedback loops is one
key element of System Dynamics
• The most important causal influences will be exactly
those that are enclosed within feedback loop
Causal Loop Diagram (CLD)
• Represent the feedback structure of systems
• Capture
• The hypotheses about the causes of dynamics
• The important feedbacks
CLD Examples

• Salary VS Performance  Tired VS Sleep

• Salary  Performance  Tired  sleep
• Performance  Salary  Sleep  tired

Salary Performance Tired Sleep

Augmenting CLD 1
(Labeling Link Polarity)
• Signing: Add a ‘+’ or a ‘–’ sign at each arrowhead
to convey more information
• A ‘+’ is used if the cause increase, the effect
increases and if the cause decrease, the effect
decreases
• A ‘-’ is used if the cause increases, the effect
decreases and if the cause decreases, the effect
increases
Signing Arcs

+ +

Salary Performance Tired Sleep

+ -
Augmenting CLD 2
(Determining Loop Polarity)
• Positive feedback loops +
• Have an even number of ‘–’ signs
• Some quantity increase, a “snowball” effect takes over and that
quantity continues to increase
• The “snowball” effect can also work in reverse
• Generate behaviors of growth, amplify, deviation, and reinforce
• Notation: place symbol in the center of the loop
• Negative feedback loops -

• Have an odd number of “–” signs

• Tend to produce “stable”, “balance”, “equilibrium” and “goal-
seeking” behavior over time
• Notation: place symbol in the center of the loop
CLD with Positive Feedback Loop

• Salary  Performance, Performance  Salary

The more salary I get

The better I perform

The better I perform

+
The more salary I get +
Salary Performance
The more salary I get +
The better I perform
CLD with Negative Feedback Loop

 Tired  Sleep, Sleep  Tired

The more I sleep The less tired I am

The more tired I am The less tired I am

The more I sleep The less I sleep

The less I sleep The more tired I am

Tired - Sleep
-
Loop Dominance
• There are systems which have more than one feedback
loop within them
• A particular loop in a system of more than one loop is
most responsible for the overall behavior of that
system
• The dominating loop might shift over time
• When a feedback loop is within another, one loop
must dominate
• Stable conditions will exist when negative loops
dominate positive loops
CLD with Combined Feedback Loops
(Population Growth)

+ +

Birth rate + Polulation - Death rate

+ -
CLD with Nested Feedback Loops
(Self-Regulating Biosphere)
 Evaporation  clouds  rain  amount of water 
evaporation  …

Sunshine
+
- + +
Earth’s A mount of
-
- temperature Evaporation water on earth
+

+ + + -
+
Clouds Rain
+
Exogenous Items
• Items that affect other items in the system but are not
themselves affected by anything in the system
• Arrows are drawn from these items but there are no
arrows drawn to these items

Sunlight reaching Density of plants

-
each plant
-
Sunlight +
Delays
• Systems often respond sluggishly
• From the example below, once the trees are planted, the
harvest rate can be ‘0’ until the trees grow enough to
harvest

delay
+

# of growing trees - Harvest rate

Planting rate -
+
Casual-loops
• Provide insight into a system's structure
• Often difficult to infer the behavior of a system from
its casual-loop representation
• Need to use computer simulation
• Simulation model: flow diagrams, equations,
simulation language
• DYNAMO (DYNAmic Models):
• Not a general-purpose language but special purpose
language to aid in building computer models
Flow Graph Symbols

Level

Rate Flow
arc

Auxiliar Cause-and-effect
y arc

Source/Si
nk

Consta
nt
Level:
• AKA stock, accumulation, or state variable
• A quantity that accumulates over time
• Change its value by accumulating or integrating rates
• Change continuously over time even when the rates are
changing discontinuously
Rate/Flow:
• AKA flow, activity, movement
• Change the values of levels
• The value of a rate is
• Not dependent on previous values of that rate
• But dependent on the levels in a system along with exogenous
influences
Auxiliary:
• Arise when the formulation of a level’s influence on
a rate involves one or more intermediate
calculations
• Often useful in formulating complex rate equations
• Used for ease of communication and clarity
• Value changes immediately in response to changes
in levels or exogenous influences
Source and Sink:
• Source represents systems of levels and rates outside the
boundary of the model
• Sink is where flows terminate outside the system
Example 1
(Population and birth)

Births Population
+

Birt
hs
Populatio
n
Example 2
(Children and adults)
+ + +
Births Children - Children maturing Adults
-
+ +

Children
Birt maturing
hs
childr Adult
en s
 Tests for Building Confidence in System Dynamics Model
(Forrester and Senge 1980, Richardson and Pugh 1981):

 Test of Model Structure

1. Structure Verification
(Is the model structure consistent with relevant
descriptive
knowledge of the system?)
2. Parameter Verification
(Are the parameters consistent with relevant
descriptive [and
numerical, when available] knowledge of system?)
3. Extreme Conditions
(Does each equation make sense even when its
inputs take
on extreme values?)
  4. Structure Boundary Adequacy
 (Are the important concepts for addressing the
problem
 endogenous of the model?)
 5. Dimensional Consistency
 (Is each equation dimensionally consistent without
Test
the use of Model Behavior
 1. Behavior Reproduction
of parameters having no real-world counterpart?)
(Does the model endogenously generate the
symptoms of
the problem, behavior modes, phasing, frequencies,
and
other characteristics of the behavior of the real
system?)
2. Behavior Anomaly
(Does anomalous behavior arise if an assumption of
the
model is deleted?)

4. Surprise Behavior
(Does the model point to the existence of a previously
unrecognized mode of behavior in the real system?)
5. Extreme Policy
(Does the model behave properly when subjected to
extreme policies or test inputs?)
6. Behavior Boundary Adequacy
(Is the behavior of the model sensitive to the addition
or
alteration of structure to represent plausible
alternative
theories?)
7. Behavior Sensitivity
(Is the behavior of the model sensitive to plausible
variations in parameters?)
8. Statistical Character
(Does the output of the model have the same
 Test of Policy Implications
1. System Improvement
(Is the performance of the real system improved
through
use of the model?)
2. Behavior Prediction
(Does the model correctly describe the results of a
new
policy?)
3. Policy Boundary Adequacy
(Are the policy recommendations sensitive to the
addition or alteration of structure to represent
plausible
alternative theories?)
4. Policy Sensitivity
(Are the policy recommendations sensitive to
plausible
References
• Simulation Model Design and Execution, Fishwick,
Prentice-Hall, 1995 (Textbook)
• Introduction to Computer Simulation: A system
dynamics modeling approach, Nancy Roberts et al,
Addison-wesley, 1983
• Business Dynamics: Systems thinking and modeling
for a complex world, John D. Sterman, McGraw-
Hill,2000