���������������������������������������������������������������
2006 ; Voinov & Bousquet , 2010 ), which are very similar for computer and mathematical models . While diagram models are developed with the same broad process , some of the detailed steps are not required in the Build and Test phases and the phases may also be more fluid . Games fall between ; more sophisticated virtual environments require the same detailed steps as other computer simulations , but board games can be modified as required by participants while being used .
���������������������
The Design phase has two outputs : the
model design ( or conceptual model ) and the project plan . These are both oriented to delivering against a purpose , but the type of purpose is different for each . The purpose for the model design is the question the model is to answer , and the terminology ‘ purpose ’ will be reserved for this meaning . The objective for the project plan is to develop a model that delivers the required policy development functionality , or to answer ‘ What is intended to be achieved by building a model ?’.
Planning tasks include inviting appropriate experts and stakeholders to participate in particular steps of the modeling process , obtaining funding and other resources , identifying available datasets , determining how the model is to be evaluated , and establishing who is to make what types of decisions . These are all important and should be discussed between the policy developer and the modeler . However , only appropriate participation substantially contributes to the functionality delivery , and other planning elements are not discussed further .
Focusing on production of the model design , the Design phase encompasses all of the abstraction required to move from the real world system to the model . That is , the purpose or policy question must be converted to a modeling perspective ; typically by the modeler asking a series of questions to assist the policy analyst to gradually refine the model design . From Figure 1 , the steps of pre implementation are :
• define the purpose , and consequently the scope and boundary of the target system ;
• describe the relevant features and relationships , and hence create the model design ; and
• select the modeling technique to be used for the model implementation .
The first step is to clarify the question the model is to answer . This question is essential to enable scoping of the model and setting the boundary of what is to be included or excluded .
The model design is then developed in response to the purpose , to describe the relevant features of the target system and their relationships . This includes information about what factors influence decisions at various points in a process , or knowledge that a change in one part of the system may induce specific changes in other parts . Some relationships may be drawn from scientific theory or existing models while others may be based on experience . Not all identified features and relationships will be able to be included in the design , and the designers will be required to decide which are most important . The objective is to present a coherent description of the combined relevant understandings about the real world system from all participants . The model design is often developed as diagrams showing the influences between factors and text , logical or mathematical rules that describe the basic operation of those influences .
66