ENGINEERING AND DESIGN SCIENCE
METHODOLOGIES
INHOUD
Introduction..................................................................................................................................................................... 2
A1 – On Science and Paradigms ....................................................................................................................................... 3
A2 – The Science of Engineering ....................................................................................................................................... 8
B1 – The Design of Artifacts ............................................................................................................................................ 14
Technique 1: Abstraction: Modeling ........................................................................................................................... 15
Technique 2: Logic: Optimization ............................................................................................................................... 17
Technique 3: Realization: Mapping............................................................................................................................. 18
Technique 4: Validation: Statistics ............................................................................................................................. 21
B2 – Architecting the Artificial ......................................................................................................................................... 23
C1 – Systems and Control Theory ................................................................................................................................... 27
C2 – Combinatorics and Entropy .................................................................................................................................... 37
D1 – Hevner: Design Science in Information Systems Research ...................................................................................... 40
D2 – Van Aken: Mgt Research based on Paradigm Design Sciences ................................................................................. 41
E1 – Guest lecture: Biometics as a Design Science Paradigm to build complex adaptive systems .................................... 42
E2.1 – NST on Modular Architectures: Artifact Production as Part of Modular Design ....................................................... 47
E2.2 – NST on Modular Architectures: Design Directives Grouneded in Combinatorics .................................................... 53
E2.3 – NST on Modular Architecture: Cross-Cutting Concern Integration ARchtectures ................................................... 59
E2.4 – NST on Modular Architectures: Integrated Elements and the Integration Design Matrix.......................................... 64
1
,INTRODUCTION
ON PARTICIPANTS, CONTENTS AND EXPECTATIONS
• About me (Prof. dr. ir. Herwig Mannaert)
o Electronics engineer, PhD in computer vision
o Teaching on engineering and software architectures, Normalized Systems Theory
o Entrepreneurial innovator
• On the Course Settings
o Course/Lecturer biases
▪ Lecturer is an engineer
• Believes in importance of engineering sciences
▪ Lecturer feels uneasy about some trends
• Overemphasis on statistics in science
• Trusting “experts” without questioning
o Master course, i.e., opinions and discussions allowed / desired
• On the Course Outcomes
o To obtain a broader perspective on science and its methodologies, enabling you to consider and apply
design science techniques
▪ in traditional technological environments
▪ in other settings where it might be appropriate
o This outcome requires some knowledge of
▪ scientific methodologies
▪ design science techniques
▪ various perspectives and approaches
• On the Course Contents
A – Some Key Essentials A1 – On Science and Paradigms
A2 – The Science of Engineering
B – Basic Design Techniques B1 – Designing the Artificial
B2 – Architecting the Artificial
C – Engineering Foundations C1 – Systems and Control Theory
C2 – Thermodynamics and Entropy
D – Some More Perspectives D1 – Design Science in Information Systems – Alan Hevner
D2 – Management and Design Sciences – Joan Van Aken
E – Some Interesting Approaches E1 – Biomimetics as a Design Strategy – Herbert Peremans
E2 – NST on Modular Architectures – Herwig Mannaert
• On the Course Materials
o Blackboard: PowerPoints
o Blackboard: Blog/Video Hyperlinks
o Parts of book: Mannaert, H., Verelst, J., De Bruyn, P., Normalized Systems Theory, Koppa, 2020.
▪ Chapters 8-12, 12-15 (, 16-19)
• On the Course Exams
o Written
o Some predictable open questions
o Well-structured answers
2
,A1 – ON SCIENCE AND PARADIGMS
ON THE BASICS OF SCIENCE AND THE SCIENTIFIC METHOD, AND THE PITFALLS OF SELF -FULFILLING
MODELS AND RUSTED PARADIGMS
“Trust is the antithesis of the scientific method” - Tyler S. Farley
You should critically validate things // questioning the existence of science, experts, …
The Scientific Method (= the method how you do science, the empirical method)
1. Observe a phenomenon
2. Find patterns in observations
3. Develop fitting descriptions and/or equations: these will be called models or hypotheses
o Models or hypotheses that needs to be validated
4. Conduct experiments to verify to what extent the models are able to predict future observations
o Crucial! Not only explain the past, also predict the future
o You can create a polynomial curve that goes perfectly through 5-10-20-… points → you can always fit a
model perfectly to existing data
o A model becomes valid when it starts to predict future observations
5. If the model/hypothesis predicts multiple observations successfully, it will become a law or scientific theory
Characteristics of Models
o Are a description, a simplification of reality
▪ Do not detail every aspect
o Fundamental laws (of physics) describe, do not explain reality
▪ For example gravitation, some things are so ‘normal’ you don’t need an explanation
▪ Does not explain how
o Appeal preferably to intuition
o As simple as possible, i.e., Ockham’s razor
▪ The lowest possible set of elements
o Need to remain stable with respect to new data
▪ You can always make the model more complex to explain previous data, but in order to be a valid
hypothetical model, you need to be able to predict future data
o Are in general valid within certain boundaries
o Should be able to predict future observations, both through extrapolation and interpolation
Observation and Modeling
o Fit model to data, but not overfitting
o Overfitting: if you make your model complex enough, you can always explain the past, fit the model to
past data, but you need to make it as simple as possible → higher chances of predicting future points
3
, Modeling
o Models are a simplified representation of reality, E.g., price car ~ weight car
o Models are never a perfect representation of reality
▪ Having exceptions and deviations
▪ Valid within certain boundaries
• E.g. truck curve: different curve (more expensive & steeper)
Laws of Nature
o Are fundamental models
▪ They describe and do not explain
▪ Fundamental: they cannot be derived from other models
▪ They do not explain how the earth is attracting you through gravitation
o Very thoroughly tested
▪ Can be (partially) falsified, never fully verified
▪ They should be tested as much as possible, people need to try to falsify it as much as possible
▪ Not falsified, it can remain a law of nature
o Are only valid within certain boundaries
▪ Newton’s Gravitation Law → General Relativity Theory
o Can be superseded by ‘better’ models that
▪ provide more accurate predictions
▪ remain valid in a broader range or scope
o Are essentially differential equations
▪ Relationships between various parameters of natures are always valid at a certain point of time
o Some Examples
▪ Newton’s Law → F = m.a
• Gravitation Law: two objects attract each other proportional with the masses of both
objects, inversional proportional to the square of the distance
▪ Electricity
• Ohm’s Law
• Coulomb’s Law
▪ Chemistry
• Mass conversation
• Energy conservation
Philosophy of Science
o Scientific explanation and prediction ~ empirical verification
▪ Verification of predictions: actual test in predicting the future
o Problem of induction → engineering
▪ It’s not because you’ve only seen white swans in your life, that there are no black swans
▪ At the same time; engineering heavily relies on induction
▪ Maybe not 100% scientific
o Ockham’s razor: simplest solution
▪ Simples solution is the best of other things being equal
o Theory and observation ~ Thomas Kuhn: paradigms
o Demarcation problem ~ Karl Popper: falsifiability
▪ It’s only science if you do falsified
4
METHODOLOGIES
INHOUD
Introduction..................................................................................................................................................................... 2
A1 – On Science and Paradigms ....................................................................................................................................... 3
A2 – The Science of Engineering ....................................................................................................................................... 8
B1 – The Design of Artifacts ............................................................................................................................................ 14
Technique 1: Abstraction: Modeling ........................................................................................................................... 15
Technique 2: Logic: Optimization ............................................................................................................................... 17
Technique 3: Realization: Mapping............................................................................................................................. 18
Technique 4: Validation: Statistics ............................................................................................................................. 21
B2 – Architecting the Artificial ......................................................................................................................................... 23
C1 – Systems and Control Theory ................................................................................................................................... 27
C2 – Combinatorics and Entropy .................................................................................................................................... 37
D1 – Hevner: Design Science in Information Systems Research ...................................................................................... 40
D2 – Van Aken: Mgt Research based on Paradigm Design Sciences ................................................................................. 41
E1 – Guest lecture: Biometics as a Design Science Paradigm to build complex adaptive systems .................................... 42
E2.1 – NST on Modular Architectures: Artifact Production as Part of Modular Design ....................................................... 47
E2.2 – NST on Modular Architectures: Design Directives Grouneded in Combinatorics .................................................... 53
E2.3 – NST on Modular Architecture: Cross-Cutting Concern Integration ARchtectures ................................................... 59
E2.4 – NST on Modular Architectures: Integrated Elements and the Integration Design Matrix.......................................... 64
1
,INTRODUCTION
ON PARTICIPANTS, CONTENTS AND EXPECTATIONS
• About me (Prof. dr. ir. Herwig Mannaert)
o Electronics engineer, PhD in computer vision
o Teaching on engineering and software architectures, Normalized Systems Theory
o Entrepreneurial innovator
• On the Course Settings
o Course/Lecturer biases
▪ Lecturer is an engineer
• Believes in importance of engineering sciences
▪ Lecturer feels uneasy about some trends
• Overemphasis on statistics in science
• Trusting “experts” without questioning
o Master course, i.e., opinions and discussions allowed / desired
• On the Course Outcomes
o To obtain a broader perspective on science and its methodologies, enabling you to consider and apply
design science techniques
▪ in traditional technological environments
▪ in other settings where it might be appropriate
o This outcome requires some knowledge of
▪ scientific methodologies
▪ design science techniques
▪ various perspectives and approaches
• On the Course Contents
A – Some Key Essentials A1 – On Science and Paradigms
A2 – The Science of Engineering
B – Basic Design Techniques B1 – Designing the Artificial
B2 – Architecting the Artificial
C – Engineering Foundations C1 – Systems and Control Theory
C2 – Thermodynamics and Entropy
D – Some More Perspectives D1 – Design Science in Information Systems – Alan Hevner
D2 – Management and Design Sciences – Joan Van Aken
E – Some Interesting Approaches E1 – Biomimetics as a Design Strategy – Herbert Peremans
E2 – NST on Modular Architectures – Herwig Mannaert
• On the Course Materials
o Blackboard: PowerPoints
o Blackboard: Blog/Video Hyperlinks
o Parts of book: Mannaert, H., Verelst, J., De Bruyn, P., Normalized Systems Theory, Koppa, 2020.
▪ Chapters 8-12, 12-15 (, 16-19)
• On the Course Exams
o Written
o Some predictable open questions
o Well-structured answers
2
,A1 – ON SCIENCE AND PARADIGMS
ON THE BASICS OF SCIENCE AND THE SCIENTIFIC METHOD, AND THE PITFALLS OF SELF -FULFILLING
MODELS AND RUSTED PARADIGMS
“Trust is the antithesis of the scientific method” - Tyler S. Farley
You should critically validate things // questioning the existence of science, experts, …
The Scientific Method (= the method how you do science, the empirical method)
1. Observe a phenomenon
2. Find patterns in observations
3. Develop fitting descriptions and/or equations: these will be called models or hypotheses
o Models or hypotheses that needs to be validated
4. Conduct experiments to verify to what extent the models are able to predict future observations
o Crucial! Not only explain the past, also predict the future
o You can create a polynomial curve that goes perfectly through 5-10-20-… points → you can always fit a
model perfectly to existing data
o A model becomes valid when it starts to predict future observations
5. If the model/hypothesis predicts multiple observations successfully, it will become a law or scientific theory
Characteristics of Models
o Are a description, a simplification of reality
▪ Do not detail every aspect
o Fundamental laws (of physics) describe, do not explain reality
▪ For example gravitation, some things are so ‘normal’ you don’t need an explanation
▪ Does not explain how
o Appeal preferably to intuition
o As simple as possible, i.e., Ockham’s razor
▪ The lowest possible set of elements
o Need to remain stable with respect to new data
▪ You can always make the model more complex to explain previous data, but in order to be a valid
hypothetical model, you need to be able to predict future data
o Are in general valid within certain boundaries
o Should be able to predict future observations, both through extrapolation and interpolation
Observation and Modeling
o Fit model to data, but not overfitting
o Overfitting: if you make your model complex enough, you can always explain the past, fit the model to
past data, but you need to make it as simple as possible → higher chances of predicting future points
3
, Modeling
o Models are a simplified representation of reality, E.g., price car ~ weight car
o Models are never a perfect representation of reality
▪ Having exceptions and deviations
▪ Valid within certain boundaries
• E.g. truck curve: different curve (more expensive & steeper)
Laws of Nature
o Are fundamental models
▪ They describe and do not explain
▪ Fundamental: they cannot be derived from other models
▪ They do not explain how the earth is attracting you through gravitation
o Very thoroughly tested
▪ Can be (partially) falsified, never fully verified
▪ They should be tested as much as possible, people need to try to falsify it as much as possible
▪ Not falsified, it can remain a law of nature
o Are only valid within certain boundaries
▪ Newton’s Gravitation Law → General Relativity Theory
o Can be superseded by ‘better’ models that
▪ provide more accurate predictions
▪ remain valid in a broader range or scope
o Are essentially differential equations
▪ Relationships between various parameters of natures are always valid at a certain point of time
o Some Examples
▪ Newton’s Law → F = m.a
• Gravitation Law: two objects attract each other proportional with the masses of both
objects, inversional proportional to the square of the distance
▪ Electricity
• Ohm’s Law
• Coulomb’s Law
▪ Chemistry
• Mass conversation
• Energy conservation
Philosophy of Science
o Scientific explanation and prediction ~ empirical verification
▪ Verification of predictions: actual test in predicting the future
o Problem of induction → engineering
▪ It’s not because you’ve only seen white swans in your life, that there are no black swans
▪ At the same time; engineering heavily relies on induction
▪ Maybe not 100% scientific
o Ockham’s razor: simplest solution
▪ Simples solution is the best of other things being equal
o Theory and observation ~ Thomas Kuhn: paradigms
o Demarcation problem ~ Karl Popper: falsifiability
▪ It’s only science if you do falsified
4
