X hits on this document





16 / 28

Clay and Johnston had run the model with the GUI many times, and so had a good idea of how various inputs affected outputs. At this time, Clay, working overtime with the students in the lab, discovered that the GUI was not reading the right files, when getting certain inputs. This explained the pattern of mistaken outputs, as this problem only occurred when you overwrote a previous policy set. So, the students did a heroic job of explaining wrong outputs in some of the papers, for some of the scenarios and were congratulated for their determination. Clay and the programmer corrected this last set of bugs and the GUI then ran properly.


For Week Eight, the class went beyond MEPLAN to a discussion of PECAS, which has a superior measure of producer surplus and many other improvements (similar to those in UrbanSim) and should be more useful for policy analysis.  The innovations in the UrbanSim modeling framework, including: microsimulation of households, small grid cells, development time lags, and an explicit floorspace developer were presented and discussed, as well as the ability of UrbanSim and PECAS to be linked to the next generation of travel models (tour-based and microsimulation).   The students were not assigned the recent technical papers by Hunt and Abraham on PECAS-type modeling or the technical papers on UrbanSim by Waddell and others.  Discussions were kept general and focused on model frameworks and practical applications. The students were assigned the PECAS overview talk by Hunt (http://www.odot.state.or.us/tddtpau/3symposium.htmt) and a paper on using GIS and economic models together at SACOG (Johnston and Garry, 2003).

Document info
Document views91
Page views91
Page last viewedMon Jan 23 10:47:53 UTC 2017