4. Modeling Problems
Abstract Level
Definition-1: A model M ={ G, F, R}
G= goal(s) for modeling
F=Candidate (observable or controllable)features or properties from the real system
R= Possible Relations between elements of F
Assumption: We observe and construct objects based on properties.
Application-Domain specific level
Technical level
Abstract level
5. Modeling Problems
Abstract Level
Definition-1: A model M ={ G, F, R}
G= goal(s) for modeling
F=Candidate (observable or controllable)features or properties from the real system
R= Possible Relations between elements of F
Assumption: We observe and construct objects based on properties.
Definition-2: A dynamic model M is a model, with varying element(s).
Definition-3: Model Complexity: ???
Definition-4: Model sufficiency:???
Definition-5: Model accuracy:???
Definition-6: Model validation and model-ability measure :???
Definition-6: Model ???
• Selection of a model should be based on the above mentioned measures from definition 2,
onwards
So in the next level, we will possibly have several classes of technical modeling
problems and modeling approachesbased on different combinations of these
elements
Application-Domain specific level
Technical level
Abstract level
8. Modeling Problems
Application-Domain specific leve
Technical level
Abstract level
Then, Application-Domain Specific problems can be
considered as a combination of one or more categories of
technical problems.
Building Management System
Design
Optimization
A few examples
9. Modeling approaches
(How we approach the problems?)
• Linearity Vs. Non-linearity
• Specific Vs. Pre-specific (explicit models vs.
environment for possible/probable explicit
models)
• Model separate than Real phenomenon vs.
Model as a Part of the real system. (e.g.
simulated agents or real connected actors)
11. It is the Story of Elephant in the Dark
RoomWhich Model is the best?
(e.g. : which approach is better for
modeling urban environment?)
Single Disciplinary Methods. (e.g. Analytical)
Multidisciplinary and Interdisciplinary models (Stocks and Flows, Urban
Metabolism)
12. Model as a part of Reality
Modeling
Reality
• We can’t Model the elephant.
Since the elephant, itself is
evolving!!
• We no longer model Agents
explicitly. We let them play.
• To Provide an environment for
real agents instead of simulated
agents!
• And this is happening with
emerging technologies in ICT:
e.g. Crowd sourcing, Pervasive
Data Collection, Eigen-Behavior
Monitoring,…
Model is a part of the everyday life.
We just collect Data through social
Media
We Process data to find evolving
patterns, vs, rigid Categories.
We inform stakeholders And they will
evolve themselves.
13. Model as a part of Reality
Modeling
Reality
And the main Elements will be:
Coding, Communication and Data Modeling.
Network Theory
Data Processing
and Modeling
Algebra
Associative Networks
(SOM)
Structured DBs
Un-Structured
DBs (Text, Audio,
Video, Image,…)
What we need is Mathematics as the art of
learning :
14. So, What next?
• Developing a Formalized Mathematical model
for “modeling” to
– Be able to categorize the real world problems,
based on modeling point of view
– Advise about model complexity, model-ability,…
• Developing a practical toolbox for modeling
based on the categories in the technical level.
15. Some points
• We are no-longer dependent to classic
academic disciplines, since we have a new
cut!?
• Theory-Driven Vs. Data Driven Modeling