SD

SD

System Dynamics
or
Systemic of Professional Engineer

Basic concepts
 Generality

3 The course is divided into two main parts

i In the first I present the basic concepts and provides some history on System
Dynamics. I will present several models made in various fields (production,
management, economics, ecology ...) and I will teach you how to use simulation
tools
i Second will cover advanced mathematical functions such as stock, conveyer, oven,
delay… Then students will be divided into small groups to build their own models in
workshops
3 Each session lasts for two days
3 Student evaluation
i Grading will be based on your oral presentations and exam results
i Oral presentation of simulation model: 50%
4 Evaluation of your ability to design operations of a complex system
i Final exam : 50%
4 Evaluation of your understanding of theoretical concepts and examples
3 Daily schedule
i Lunch
4 12:30 – 1:30 pm
i Break
4 10:30 – 10:50
4 2:30 pm – 2:50 pm
i End
4 4:00 pm

SD Page 2
SD Stéphane Copin
« System Dynamics lesson »
 Generality

3 System Dynamics is applied in many fields to

understand and anticipate changes in the mid to long
term development of a structure
i Anticipating the effects of an oil slick on an ecosystem
i Choosing the best scenario for launching a new product
i …
3 To show student the following
i The advantages of using a systemic approach
i How to identify and design operations of a complex system
with feedback loops
i How to use the software Stella ® to simulate and analyze the
running of a complex system

SD Page 3
SD Stéphane Copin
« System Dynamics lesson »
 Course syllabus

3 First Day
i Presentation of the System Dynamics approach
4 Basic concepts
i Examples of application
4 Retarded catastrophe
i Causal analysis
4 Some advices for building your diagram
4 How to start your analysis ?
i Exercise
3 Second day
i Modeling with Stella ®
4 Dissonance in Nikunau
4 Detailed explanation of the model and corresponding simulations
i The war on drug
4 System Dynamics in action
i Exercise
3 Third and fourth day With Vincent Vettier

SD Page 4
SD Stéphane Copin
« System Dynamics lesson »
 Introduction

3 We define a System like a group of interconnected

variables influencing mutually
3 We define the Dynamic Complexity by a system
composing of a set of variables connected between
them by feedback loops
3 Confronted with behavior in the time of a system (its
evolution) we have difficulties understanding
because of
i The temporal behavior of a set of interconnected variables is made
complex by the presence of feedback loops which can occur
simultaneously or successively within the structure
i We don’t have a talent for analyzing what risks to take place within
a system when several variables takes place at the same time the
role of cause and effect

SD Page 5
SD Stéphane Copin
« System Dynamics lesson»
 Introduction

3 System dynamics is a methodology for studying and

managing complex feedback systems, such as one
finds in business and other social systems
3 Feedback loop is the differentiating descriptor here
3 Feedback refers to the situation of X affecting Y and Y
in turn affecting X perhaps through a chain of causes
and effects
3 One cannot study the link between X and Y and,
independently, the link between Y and X and predict
how the system will behave
3 Only the study of the whole system as a feedback
system will lead to correct results

SD Page 6
SD Stéphane Copin
« System Dynamics lesson»
 Introduction

1956 1960 1960 - 1970 1970 - 1975

Beginning of the Forrester The Urban Forrester

field of published the Dynamics model published
system first, and still was the first WORLD2 in a
dynamics classic, book in major non- book titled World
the field titled corporate Dynamics
Industrial application of
Dynamics in 1961 system
dynamics

SD Page 7
SD Stéphane Copin
« System Dynamics lesson»
 Introduction

3 The model "world2" describes in the aggregate principal

forces that influence the state of our global environment
i The main result
4 Addressing one of the forces (population, food, economic
development, etc..) Regardless of others, would always end a
disastrous effect. Only a coordinated and progressive environment
could stabilize our ecological, economic, balance of population,
etc..
3 From 1975, France, led by Jean Lebel, Michel Karsky,
Bernard Paulré ... organizes seminars that led to the
creation of an international scientific society now based in
the United States with some 1,000 members
3 These cover all work areas where time is a critical
variable
i Business Strategy, Economics, Finance, Production, Marketing,
Management, Human Resources, Biology, Medicine ...

SD Page 8
SD Stéphane Copin
« System Dynamics lesson»
 Introduction

3 The Dynamic term means the movement, the change

3 System Dynamics is the science of the change and
the evolution

The SD feature is to analyze the

factors which create the change
or, on the contrary, which
opposes it

SD Page 9
SD Stéphane Copin
« System Dynamics lesson»
 Course syllabus

3 First Day
i Presentation of the System Dynamics approach
4 Basic concepts
i Examples of application
4 Retarded catastrophe
i Causal analysis
4 Some advices for building your diagram
4 How to start your analysis ?
i Exercise
3 Second day
i Modeling with Stella ®
4 Dissonance in Nikunau
4 Detailed explanation of the model and corresponding simulations
i The war on drug
4 System Dynamics in action
i Exercise

SD Page 10
SD Stéphane Copin
« System Dynamics lesson »
 The concept of “force” and of “state”

3 Every leader of SMALL AND MEDIUM

SIZED FIRM will understand easily the
difference between the concept of
“force" and "state "
Factor of Movement
(search for new customers)
3 A good management, a solid
insurance contract, a healthy
Faucet
finance, constitute force which
protect from the disaster

3 But only the search for

customers, for new contracts
establish the force, the push which
insures a development of the
company
STATE
(Finance…)
Bathtub
Tap
3 And we know well that the search
for new customers (active force) is
influenced by the financial situation
(state of the system) of the
company : this search made easier
when the financial situation is good

SD Page 11
SD Stéphane Copin
« System Dynamics lesson»
 The concept of “force” and of “state”

A force acting on the state of the

Reality system always results from the
comparison between two states:
the target and the reality

It is the difference (gap) between

Gap Force target and reality that grows to
the action, which generates one
(or) force (s) which will have the
effect of modifying later the
reality
Target
And later even the target

This notion is essential and very useful to

start the modelling of a system

SD Page 12
SD Stéphane Copin
« System Dynamics lesson»
 Feedback loop

3 We speak about negative relation when a variation

(increase or decrease) of the variable creates an
effect in the opposite direction
i We symbolize this effect by a negative polarity

-
3 We speak about positive relation when a variation
(increase or decrease) of the variable creates an
effect in the same direction
i We symbolize this effect by a positive polarity

+
SD Page 13
SD Stéphane Copin
« System Dynamics lesson»
 Feedback loop

3 An increase in A makes B greater; a decrease in A makes B

less

+ B
A

3 An increase in C makes D less; a decrease in C makes D

greater

- D
C

3 Definitions of Link Polarities

i All words and phrases are expressed as quantities that have a clear
sense of increase or decrease
i No verbs — the action is in the arrows

SD Page 14
SD Stéphane Copin
« System Dynamics lesson»
 Feedback loop

Parental Support Higher test scores mean

Test Scores for School
+ Administration
more support; lower
scores, less support

More teachers mean smaller

Teachers
– Class Size Student classes; smaller classes mean
– Learning more learning. Fewer teachers
would mean larger classes and
less learning
“Ceteris paribus”... (Toutes choses étant égales par ailleurs) (Assuming all else is held constant)
All other influences held constant as we assign polarities

SD Page 15
SD Stéphane Copin
« System Dynamics lesson»
 Counteracting feedback loop

General behavior of a
counteracting feedback loop
without delay between cause and
effect Counteracting
Intensity
Force Feedback loopState

Stabilization

3 A counteracting or
Time
balancing feedback loop,
Intensity
will tend to bring in a
asymptotic way the
corresponding structure
towards a state of
Stabilization
balance: we can speak
about homeostatic
Time
behavior
SD Page 16
SD Stéphane Copin
« System Dynamics lesson»
 Counteracting feedback loop

The cyberneticists* use the notion of negative feedback loop which is

characterized by an odd number of negative relations between variables

Example
A stock results directly from the
level of production

The more there will be of available

stock the less we produce
Production
- Stabilizing effect
Stabilizing effect Stocks
An increase in stock resulting from a
+ strong production will reduce it
(Stabilizing effect)

* Cybernetics was created by Norbert Wiener in the 40s. It is used in all engineering

SD Page 17
SD Stéphane Copin
« System Dynamics lesson»
 Counteracting feedback loop

Example

Stock
Stabilization of stock - Stocks

Counteracting
Counteracting
Feedbackloop
Feedback loop

Time
Production
+
If the stock grows then the
production decreases what
stabilizes the system

SD Page 18
SD Stéphane Copin
« System Dynamics lesson»
 Counteracting feedback loop

Other example

The more the population

is important, the more the
- Stabilizing
deaths increase
Stabilizing
Population effect
effect Deaths
+
The deaths decrease the
population and the
regulating nature of this
feedback loop has the
effect of stabilizing the
population

SD Page 19
SD Stéphane Copin
« System Dynamics lesson»
 Counteracting feedback loop

Other example
Price A price increase of a product will have
- an opposite effect on its consumption
(slowing down)
Stabilizing
Stabilizing -
effect
effect Consumption Hence an increase in available stocks
Stocks
- And a lower increase in prices

Price
The increase in prices increases profits
-
Stabilizing
Stabilizing
+ Profits tend to make the production
effect
effect Profits
Production grow
+ The growth of the production regulates
the prices later

SD Page 20
SD Stéphane Copin
« System Dynamics lesson»
 Counteracting feedback loop

Price
- +
Stock
-
- -
Consumption
Profit

+
-
Production
+

Naturally previous both feedback loop In reality, we find that the

act simultaneously complexity of temporal behavior is
due to the interweaving of
What complicates the problem hundreds of feedback loops acting
more or less simultaneously

SD Page 21
SD Stéphane Copin
« System Dynamics lesson»
 Reinforcing feedback loop

General behavior of a reinforcing

feedback loop without delay
between cause and effect
Reinforcing
Intensity Force feedback loop State
If the cause grows then
Explosion
the effect grows faster
and faster, which
gives an explosive
behavior …

Time A reinforcing loop is one

in which an action
Intensity
produces a result which
influences more of the
same action thus
…or implosive resulting in growth or
decline. The popular
Implosion
naming is the vicious
Time circle

SD Page 22
SD Stéphane Copin
« System Dynamics lesson»
 Reinforcing feedback loop

The cyberneticists use the notion of positive feedback loop which is

characterized by an even number of negative relations between variables

Example
Potentially explosive loop in which behavior is
verified today by the world evolution of the human
population

+ Higher population, higher birth rate

Amplification
Amplification
effect
effect The more the births increase, the
+ more the population increases

SD Page 23
SD Stéphane Copin
« System Dynamics lesson»
 Reinforcing feedback loop

Other Examples

Whatever the interest rate (not null) is, a

Amplification
Amplification
+Capital capital we don’t touch grows in a
effect
Profit effect
exponential way (but not necessarily as
+ fast as inflation

Prices
A fast growth in the prices leads to
- buying hedge (An investor might use a buying hedge if
Stock
Amplification
Amplification
+
Anticipated he/she expects to buy a certain amount of the commodity in the future,
effect but is worried about price fluctuations)
effect price
- These purchases empty the reserves
Buying + (stocks)
hedge
The prices increase all the more
justifying more buying hedge

SD Page 24
SD Stéphane Copin
« System Dynamics lesson»
 Reinforcing feedback loop

Other Examples
Motivation The more we act the more we have
- chances to obtain a reward
Frustration
This is compared with the desire that
+ R + we have to obtain the reward
Action
Gap Inhibition If the reality is below our aspiration, we
Loop of can develop a feeling of frustration
+ - H. Laborit
This frustration decreases the
Reward
+ motivation, thus the action and the
chances to obtain a reward, the
Desired
frustration becomes even stronger
reward

SD Page 25
SD Stéphane Copin
« System Dynamics lesson»
 To sum up this chapter

3 Reinforcing loops 3 Counteracting loops

i self-reinforcing i balancing
i growth producing (Process of i goal seeking (Obtain a desired
growing) input)
i destabilizing i stabilizing
i accelerating i compensating
i even number of minus i odd number of minus
3 Symbolized by 3 Symbolized by

R + B C –

SD Page 26
SD Stéphane Copin
« System Dynamics lesson »
 Delay

3 The behavior which we have just described contains in

fact one or several delays between variables
3 The relations of influence, when they are not of purely
accounting nature …

3 …contain almost always delays which represent the

non spontaneity of cause and effect relationship

SD Page 27
SD Stéphane Copin
« System Dynamics lesson »
 Delay

Illustrations
Delay
Delay
It is not immediately that the evolution
of a stock influences the corresponding

Stocks
- production
Production
+ There is always a response time,
corresponding generally to a delay of
consideration of the realities and the
corresponding information

Price
Delay Delay
Delay
Delay
-
Stocks
-
Consumption
It is only after a while that a variation of
stock can have an opposite effect on the
- price

SD Page 28
SD Stéphane Copin
« System Dynamics lesson»
 Delay

Delay

Stocks
-
Production
+

3 The causal relations don’t have all the same

meanings
3 The relation between production and stock is a physical
relation
i A produced quantity fills automatically a stock
i In that case the delay depends mainly on the speed of the machine
3 The relation between stock and production is a
relation at once of information and decision, it is not
an automatic relation
i We can decide to continue to produce whatever is the level of stocks
i In that case, there is always a delay between the information and the
decision which has to result from it
i In companies these delays are often known with a very big imprecision

SD Page 29
SD Stéphane Copin
« System Dynamics lesson»
 Delay

Feedback loop
IDAR
The decision-maker receives an
Information information
This information arrives with certain delay
Delay 4 Delay 1
Delay 4 Delay 1 Information
Delay n
Delay n
Feedbackloop
loop The decision-maker starts an action
Feedback
Result IDAR
IDAR Décision Again, between the decision and the
Delay
Delay
action, there is delay
Delay 2
Delay 3
Delay 3 Delay 2 Information
The action leads to a result
Action Between the action and the result it is
necessary certain time
IDAR is established by multiple similar
interconnected feedback loop (any
The result leads to a new information …
decision being generally taken on the
basis of several information)

SD Page 30
SD Stéphane Copin
« System Dynamics lesson»
 Delay

Feedback loop IDAR : degradation of

information
Feedback
information In the feedback loop IDAR the information,
before or during its use, can be degraded

Result Information J It can have errors, distortions there

J The information arrives or is used with delay

and is rarely used continuously (There is rather
Degradation of
sampling of the information: daily, monthly
information dashboards …)

Action Use of « AIRBUS » Project

information Distortion of information
« The very lively pressure exercised from
the height very fast aroused behavior
excessively optimistic. Each minimized the
Decision met troubles and made go back up a vision
embellished by the marker poles for which
he was responsible» Le monde 20/06/06

SD Page 31
SD Stéphane Copin
« System Dynamics lesson»
 Interaction between two feedback loop with delays

Intensity Intensity Intensity

Explosive
behavior
Stabilizing Explosion then
behavior stabilization

Time Time Time

Intensity Intensity Intensity

Explosive
oscillations
Rise,
Rise, oscillations
overtaking of
and stabilization
objective and stabilization

Time Time Time

Intensity

Permanent
oscillations

Time

SD Page 32
SD Stéphane Copin
« System Dynamics lesson»
 Principle of nonlinearity

In the reality of the world which

surrounds us, most of the relations
of influence between variables
contain an element of nonlinearity
between cause and effect

SD Page 33
SD Stéphane Copin
« System Dynamics lesson»
 Principle of nonlinearity

Example

Birthrate

In case of overpopulation, the birthrate

tends to fall

NNo
onn
r RReela
eaar In case of subpopulation, there is an
ine

la tio effect of compensation ( Baby Boom)

liln

tio n
n

But if the population decreases too

much, the birthrate begins falling
dangerously (Risk of extinction of
species)
Population

SD Page 34
SD Stéphane Copin
« System Dynamics lesson»
 Principle of nonlinearity

Change of polarity

Favorable zone
(Counteracting feedback
loop)
-
Actual Security Unfavorable zone
(Reinforcing
+ N
N on
o n li
lin ne
feedback loop)
ea ar
ri ity The relation between number of test
ty r
re ela
la ti
ti on
and actual security …
on
Explosion
Explosion Transform a negative loop (Intention /
stabilization of the incidents) …
Number of security test
-
Incident New security rules …In a positive loop (increase of
+ incidents)

Result against intuitive and potentially

catastrophic

SD Page 35
SD Stéphane Copin
« System Dynamics lesson»
 The structural evolution

3 During a detailed analysis of a system, we notice fast that

the relations between variables and feedback loops
revealing are not all activate at the same moment
i According to the evolution of the system and the state in which it is
at the given moment
4 Such loop can be active or not
4 Dominant or ineffective
3 It results from it a possibility of change of behavior
i Sometimes radical
i Sometimes very fast
3 This structural change is thus predictable, because
contained implicitly in the structure

SD Page 36
SD Stéphane Copin
« System Dynamics lesson»
 The structural evolution

Example
If the market is "a salesman" …
Intensity the production being lower than the demand …

« Salesman » market « Buyer » market the prices are going to increase …

Capacity of what will generate at once a fall in demand and

production …
Price
an increase of the investments to satisfy the
demand …
Demand

It is this crossing between demand and

production …
Production
After a certain delay of appreciation (delay in the
consideration of the information) …
0 10
Years Which will make pass of “salesman's state” in that
Delay of appreciation of “of buyer”
change in the situation

SD Page 37
SD Stéphane Copin
« System Dynamics lesson»
 Summary of Structure and Behavior

3 Reinforcing loops are responsible for

accelerating growth, accelerating decline, or
traps due to inertia
3 Counteracting loops are much more prevalent but
often less obvious; they balance and adapt
3 Counteracting loops can compensate for policy
initiatives and defeat intended effects
3 Combinations of reinforcing and counteracting
loops are responsible for all the complex
dynamics we create in our complex systems

SD Page 38
SD Stéphane Copin
« System Dynamics lesson »
 John Sterman model

SD Page 39
SD Stéphane Copin
« System Dynamics lesson »
 John Sterman model

3 As an illustration of the use of system dynamics, imagine

an organization that plans to introduce an innovative new
durable consumer product
3 The organization needs to understand the possible market
dynamics in order to design marketing and production
plans
3 There are two feedback loops in this diagram
i the positive reinforcement (labeled R) loop on the right indicates that the
more people have already adopted the new product, the stronger the
word-of-mouth impact. There will be more references to the product,
more demonstrations, and more reviews. This positive feedback should
generate sales that continue to grow
i the second feedback loop on the left is negative reinforcement (or
"balancing" and hence labeled B). Clearly growth can not continue
forever, because as more and more people adopt, there remain fewer
and fewer potential adopters
3 Both feedback loops act simultaneously, but at different
times they may have different strengths. Thus one would
expect growing sales in the initial years, and then declining
sales in the later years

SD Page 40
SD Stéphane Copin
« System Dynamics lesson »