I'm presenting a paper entitled 'Agent-Based Geospatial Simulation: Techniques, Tools and Advances in Three-Dimensional Visualisation', for those interested the abstract is below.
The use of agent-based modeling for geospatial simulation has traditionally been a two-dimensional activity with the third dimension rarely ventured into. In many ways this is due to the nature of the discipline and a focus on theory rather than outreach and end of game visualization. The world of video entertainment has made significant advances in artificial intelligence over the recent years, combined with the rise of graphic card technologies these developing ‘sandbox’ environments provide a new opportunity for agent-based modeling.
Standard three-dimensional modeling packages such as 3D Studio Max offer a level of built in agent-based modeling capabilities through ‘crowd and delegate’ logic systems. While these are limited to some extent, it is possible to take x/y/z output from standard geospatial agent-based modeling approaches and import them into visual software. This approach allows loose coupling between model and visualisation systems allowing a more realistic view of any given models output.
We explore the use of 3D Studio Max and gaming technologies, such as Crysis, though a series of examples before concluding with a look at using collaborative virtual environments such as Second Life to both visualise and run agent-based geospatial models.