W09 Simulation Modeling

Question 1

When not to simulate?

Answer 1

experiments are easily and cheaply realized in real systems

desired indicators can be calculated analytically

Question 2

simulation run procedure

Answer 2

behavior of real phenomena and real environmental conditions
are sufficiently approximated by simulated phenomena

then
simulations can plausibly represent real-world behaviors

else we have to highligth conditions that are not fulfilled as explicit and implicit assumptions

Question 3

Simulation Types

Answer 3

System
constant or dynamic
State Changes
continous or discrete
Input Type
deterministic or stochastic

Question 4

Discrete Event Simulation

Answer 4

-events occur at time points and change system state; simulated time does not pass smoothly

Question 5

Concepts of Discrete Event Simulation

Answer 5

Entity
Entity Queue
Events
Event list
Simulation Clock
Statistics Indicator
Entity State (idle or busy)

Question 6

Agents can..?

Answer 6

communicate
perceive system states
act rationally or not
influence system states
evolve

Question 7

Agents?

used for?

Answer 7

interconnected heterogeneous agents generate emergent effects

Question 8

System-Dynamic Simulation

Answer 8

qualitatively model effects of continous influence factors

• holistic approach
• feedback loops

Question 9

Building a model:

Transition from real system to model by?

Answer 9

reduction
abandonment of unimportant components
abstraction
generalization of specific system characteristics

detailed as necessary, asimple as possible
every detail requires data, rules and costs and intorduces complexity

Question 10

Mistakes by novice modellers

Answer 10

over-reliacne and available data

taking shortcuts

insufficient use of abstract variables and relationsihps

ineffective self-regulation

overuse of brainstorming

Question 11

Real world problems

Answer 11

uncertainty
multifacetted
many points of view
assumptions
messy
ambiguity and disagreement
always changing

-> no single and permanent solution

Question 12

Models require

Answer 12

empirical knowledge
theoretical thesis
imagination (assumptions)

Question 13

Conceptional Model

Answer 13

• software independent description of model

- should be in place prior to coding

Question 14

Conceptional modelling frameowrk

Answer 14

determine Inputs
are outputs achieved?
do objectives correspond to problem situation?

Question 15

Model Evaluation

Answer 15

Results

• scope of output
• accuracy
• understanding

Future Use
-portability

Confidence

• Verification
• Validation
• Credibility

Resources

• Build time
• Run time
• hardware requriements

Question 16

Simple Models

Answer 16

easier, faster, less data required

probably less validity

->want simplest valid model!

Question 17

Assumptions should be assessed for?

Answer 17

confidence? how sure are we?

impact? how large is effect of assumption?

Question 18

How to deal with low confidence, high impact assumptions?

Answer 18

• worst case
• sensitivity analysis
• decision variable
• more information
• highlight in results
• stop project

Question 19

Simplifications

Answer 19

asess by impact

Types:
leave things out
group things
restrict value range
scale slow changes as constant

Question 20

Simplifying and Existing Model

Answer 20

+increase understanding
+easier experimentation and analysis of results
+cross validation and verification

• time consuming
• danger of misleading results
• reducing experimental frame