���������������������������������������������������������������
����������Phases of the modeling process and their potential to deliver functionality . Not all functions are available for all modeling techniques . The general flow of the modeling process follows the large arrows from Design to Build to Test to Use , with the potential for iteration between Build and Test through increasingly more complete model versions for techniques such as computer simulation . In addition , models may go through redesign as requirements change or later phases highlight errors or gaps .
uate five techniques for similar functions as described in this paper . Ghorbani et al . ( 2014 ) also assess the suitability of five techniques , but instead focus on 18 potential tasks within an options comparison policy process . While these analyses both refer to specific techniques ( with System Dynamics and Agent-Based Modeling included in both ), much of the suitability arises from the trade-off between accessibility and scenario testing and which broad group the technique sits within ( such as computer simulation ). Other differences are related to the accuracy considerations discussed in the next section .
How then can a modeling project include the full range of knowledge synthesis and unknowns management if ‘ what if ’ analysis is also required ? To overcome the communication limitations of mathematical and computer models , the modeling process for these techniques usually use other models to develop the design . For example , Causal Loop Diagrams are typically used when developing System Dynamics models ( Sterman , 2000 ). Similarly , developers of Agent Based Models may use UML diagrams ( Bommel et al ., 2014 ) or games Lynam et al . ( 2002 ) as intermediate steps in model development . Similarly , the Cooum River diagram was used to help develop a simulation model for one of the identified policy options ( Bunch , 2003 ).
75