AoThesis-54

Technical Research |Establishing Model Methodology

Problem formulation: It is important to begin this process by first stating the problem. This was already established in Part 1:Introduction and Theory by investigating the emergence of dynamic architecture in Chapter 1.1, the inadequacy of current visualization methods in Chapter 1.2, and the advent of game engines in Chapter 1.3.

Setting of objectives and overall project plan: This step aims to indicate the questions that can be answered with the simulation. In the context of this thesis, the objective would then be finding a way to simulate and visualize human crowd movements within a dynamic architectural space.

Model conceptualization: This step is the most technical, as it deals with establishing the various components within the simulation model, as well as defining their interactions within the system. This is what this current chapter (2.2) will investigate and what the next chapter (2.3) will establish. This model will then be updated as the thesis moves on from the processing prototype to the game engine environment.

Data Collection: This step focuses on the collection of input data for the model, and as such, constant interplay will be present between this step and the last. In the context of this thesis, Chapter 2.3 will be investigating various properties of human behaviors and how that can be utilized and represented within the system model. The initial data from this step, as well as the established methodology from the initial model conceptualization, will be invaluable when re-establishing the model within the game engine environment.

Model Translation: This step deals with the technical translation of this simulation model into machine logic. The Java-based program processing will be utilized first to prototype this methodology as a way to validate the concept of emergent complex systems within the context of crowd simulations. Once this processing prototype is validated, the model will then be translated into a game engine environment to utilize its higher abstraction tools.

Verification: This step mainly focuses on debugging to ensure the software operates according to the model. Fixing certain software bugs can take minutes to days, therefore, this step arguably provides the greatest unknown in terms of time expenditure due to the inherent complexity and uncertainty within the debugging process.

Validation: The aim of this step is to determine whether the simulation model is an accurate representation of the real system. As such, this step may be repeated until the resulting model accuracy is judged acceptable. In the context of this thesis, the processing prototype will first need to be validated to confirm that the concept of emergent behaviors from autonomous agents is enough to convey the movements of human crowds. Then, once the model is re-updated for the game engine environment, this step will need to be revisited to confirm this new model within the context of human crowds.

Experimental design: This step determines the various alternative systems that can be simulated within this model. This thesis will be investigating this in Part 4, where it will be utilizing the developed simulation tool within a variety of architectural scenarios to determine the usability of this tool within alternative spatial applications.

Production runs and analysis: This step utilizes the simulation to estimate measures of performances within these alternative systems. This step can be utilized to analyze the architectural scenarios from the experimental design step, which can then be used to measure their effectiveness in visualizing spatial typologies.

More runs: Additional runs may or may not be required depending on the analysis and updates of the previous runs.

Documentation and reporting: The documentation of the methodologies and concepts obtains from creating this model will allow potential upgrades to the system model in order to provide increased accuracy and performance metrics in the future.

Implementation: At this point, it is possible to establish a rough workflow within architectural visualization that can benefit with the utilization of this crowd simulation tool.

As shown by the various steps of this framework, the next stage in creating this crowd simulation—as well as the focus of this and the next few chapters— is developing a model for this simulation system. A model as defined by Banks et al. is “an abstract representation of a system, usually containing structural, logical, or mathematical relationships which describe a system in terms of state, entities and their attributes, sets, processes, events, activities, and delays.”[3] He remarks, “Just as the components of a system were entities, attributes, and activities, Models are represented similarly. However, the model contains only those components that are relevant to the study.”[4] Going by this, it is then important to first simplify architectural spaces into its fundamental relevant components. The purpose of this model, then becomes to define the interactions between classes of entities within the system, which can be defined as the active human agents and the passive (or active in the case of dynamic spaces) architectural objects.