��������������������������
�������� : Relationship between real world issue and abstracted model . Each step moves the model representation further from the system being modeled , but sufficient realism must be maintained to allow generalizations from the behavior of the model to be applied to the expected behavior of the system . This realism relies on identifying and retaining the relevant influences throughout the abstraction process .
The second level of abstraction is designing the model ; selecting and documenting the features and relationships within the target system that are to be included in the final model . Simplifications at this level are conscious decisions about which physical , social , environmental , and technological factors within the target system are significant for the model question , and how to specify those factors , their interactions and their behavior . At this stage it is sufficient to document that there is some relationship , without directly stating its form .
For the case study , relevant factors include rainfall , population size , use of water efficient equipment , and water price . Each of these factors affects water volume in some way . The relationship could be described with words ; for example , ‘ greater rainfall directly increases water ’ or ‘ higher price reduces consumer demand which makes more water available ’. Alternatively , a
diagram could be used with arrows between factors to indicate the direction of influence .
The model design explicitly represents the features and relationships and can be considered a model in itself , but the design is typically translated into a more detailed model implementation , which involves a final level of abstraction . This is the stage at which the relationships between factors are fully specified . There are many different techniques available , including detailed diagrams , mathematical equations , games , and computer simulations . In the case study , the final model was a System Dynamics model , a type of computer simulation .
Once constructed , the model is used to answer the original question . Typically such use examines the model output under various inputs , and if the abstraction is sufficiently accurate , the output can be generalized to the target system in the expectation
63