activity based route choice solutions

i wonder how activity based aggregate models could be useful?...Activity based models are data intensive; we dont have that kind of data available...similarly, even if some region in the united states used activity based models, my question to them would be how much of a difference it made to their projections as compared to when they used regular gravity model or any other combination of route choice model. If thedifference is not much, then how do they justify activity based models?

My personal thought on this issue is that Activity based models are only good in combined traffic assignment and control framework (CTAC), and that too only if we use dynamic assignment process. Otherwise, activity based models will get stale soon and the researchers will have to come up with some new idea to do projections.