Establishing Model Methodology

Now that the foundational principles of complex system simulation have been investigated, and the plan of utilizing a bottom-up simulation methodology has been established, it is possible then to dive deeper into the creation of this simulation. Within Discrete-Event System Simulation, Jerry Banks et al. determined a set of steps to facilitate a thorough simulation study to aid in creating a simulation model.[1] (Fig. 2.2.1) He breaks this down into four phases:

“The first phase, consisting of steps 1 (Problem Formulation) and 2 (Setting of Objective and Overall Design), is a period of discovery or orientation. […] The second phase is related to model building and data collection and includes steps 3 (Model Conceptualization), 4 (Data Collection), 5 (Model Translation), 6 (Verification), and 7 (Validation). […] The third phase concerns running the model. It Involves steps 8 (Experimental Design), 9 (Production Runs and Analysis), and 10 (Additional Runs). […] The fourth phase, implementation, involves steps 11 (Documentation Reporting) and 12 (Implementation).”[2]

These phases describe a very logical approach, as such they line up nicely with the chapter structure of this thesis. The first phase of discovery and orientation is covered in Part 1: Introduction and Theory, where the proposed framework was established; the second phase of model building and data collection is covered within this current part (Part 2: Technical Research), as well as Part 3: Tool Creation for when the simulation model is updated for the game engine environment; The third phase concerning running the model is covered within Part 4: Simulation Applications, where the simulation will be utilized within a variety of applications to review their effectiveness in architectural crowd simulation; The fourth phase of implementation is finalizing the resultant simulation as a tool within visualization applications. Since the initial scope of this thesis was meant to provide a foundational framework for this tool, the full completion of this phase may be beyond the scope of this thesis and would be suited to further development at a later date.

As it can be seen, this relatively simple framework provides a good starting point and guideline in creating a custom model for crowd simulation. Therefore, by applying this further, it is possible to provide a more thorough breakdown of what each step may entail in relation to this thesis.