Chapter 1 Flashcards
Define a model
A model is an imitation of a real world system or process - captures key features of the system
Define a deterministic model
Variable we model (output) is unqieuly determined by the input variables
Define a stochastic model
The output is not uniquely determined by the input variables it is an RV. Outputs will have a range of probabilities
Describe sensitivity analysis
The effects of changing certain input parameters can be studied before a decision is made to implement the changes in the real world
What are the four different types of models you can have in terms of state space and time
Continuous state space with continuous time
Continuous state space with discrete time
Discrete state space with continuous time
Discrete state space with discrete time
How do we know what model type to use?
Purpose of the model determines what model type
Give summary of 11 steps for building a model
- set well defined objectives
- Plan how to validate model
- Define relationship between inputs and outputs
- collect and analyse data
- involve real world experts and get feedback
- decide how to implement model
- write and debug computer program
- test reasonableness and analyse output
- test sensitivity of output
- communicate and document results
- monitor changes and update the model
What are the advantages of models
Framework that’s structured to update our knowledge as a model only is superseded when a better one is made
They help study complex systems and allow consequences of certain actions to be assessed
What are the pitfalls of modelling
Assumptions can be unrealistic
Requires alot of expertise time and money
Model is only as good as the data inputs
Complex models that look good can lead to overconfidence
Models can become obsolete with changes
Communicating results and interpreting model for non experts can be difficult
Limitations of the model have to be understood
