ABM in GISolve2

Modeling scenarios

  • Basic: for first-time user/students. we need good examples/tutorials to bring in users
  • Intermediate: user modifies parameters and feeds their data to do ABM computation.
  • Advanced: modeler needs to use ABM for his own model development: coding is needed

Primitive thinking

  • Basic: we provide web content: examples/tutorials and user interface to invoke ABM code. Usability is the focus: examples must be straightforward and must work; user interface is informative
  • Intermediate: support parameter-sweeping based experimental analysis.
  • Advanced: use VM. Like Amazon EC2, we ask user to give a list of current analysis applications he needs for his model development. We then customize a VM by installing bin packages of these apps. What we give to user is a ready-to-work machine VM. User will work his way out to integrate ABM, develop code to change model behavior (elk→bear, yellowstone→rocky mountain, ...). He will test his code and validate results. Then we have a set of publishing tools/services to put it in GISolve for community use: examples, user interfaces, web service interfaces, CI resource engagement, etc.