As of this writing there is not a current QRS II tutorial available. QRS II and GNE are written to be comparatively user-friendly, but with all complex software, there is at least a moderate learning curve to surmount. However, nobody getting into the subject of travel forecasting is a dummy, so I am guessing that most people don’t need a lot of hand-holding to get going with the software. So here are my top 10 things-to-know for QRS II newbies.
- All QRS II runs need at least one highway network, and typically only one highway network. Highway networks are drawn with GNE. Networks consist of nodes, links, and maybe polygons. In a network, links always connect to two nodes at either end, the A-node and the B-node. All link operations in GNE are node-to-node. That is, to do anything with a link (draw it, open it, etc.) click one end of the link, then the other end of the link. Nodes are usually connected to links, but nodes can also be free standing. For one-way streets, traffic always flows in the A to B direction.
- QRS II looks for a “Project” folder that contains most inputs for a particular QRS II run. The project folder can be set on QRS II’s main window. Networks can be conveniently placed in the project folder, but they can be placed elsewhere. I recommend that all QRS II outputs be written to separate folders, and not mixed in with inputs. Hence, it is a good idea to create two new subfolders, “Reports” and “Temp” within the project folder. Then you can direct QRS II to place its outputs in those folders via the Run Controls dialog box. When QRS II is installed it creates one project folder, QRSIIProject64, on your computer, but you can create as many copies of this folder as you like for other QRS II runs.
- Every network drawn in GNE needs an “application schema”, which is a template containing all the definitions of links, nodes, polygons, and their attributes. Before inputting data into a GNE network, you must first open an application schema file. Look in the QRSII9 folder under your Program Files folder for application schemas. There are lots of these files supplied with QRS II to support various versions. They range from minimal to really complex. For educational purposes, I recommend the QRSEasy.dta application schema. The QRSExtensiveS.dta application schema (downloadable from AJHAssoc.com) supports all version 9.1 features, but it is very complicated.
- Some new users will draw a network without first opening an application schema file, then discover they are unable to input any attributes. If you are in this situation, you can fix the problem by first saving the draft network to any valid name, say Draft.dta, then opening an application schema file, say QRSEasy.dta. Finally, “append” the draft network to the existing application schema.
- All QRS II application schemas come in pairs. Those schemas with “UK” in their name are for places with left-hand drive, such as England and Australia. If you are drawing a network for one of those places, you must also tell QRS II about this fact on the Assignment Options dialog box.
- You can draw a network completely by hand, but these days it is usually easiest to import a network from a GIS source with GNE’s ArcGIS Shapefile Import tool. This process can be tricky and may require several tries. The most important thing to remember is that the GNE coordinate system is in pixels, while most GIS databases use long-lat, although they don’t have to be this way. Thus, it is important to make sure GNE converts the coordinate system faithfully. GNE can do most of the work, but you must help. First load an application schema file into GNE and go to the ArcGIS Shapefile Import tool. Pick a projection method, then set a scale for the network in pixels per mile (apologies to most of the world who use the metric system). Click the Suggest button to set the transformation parameters, then select your GIS line (SHP) file. Once the parameters are set, do the import. For simple networks, you usually only want to import a line file, not a point file. If you import both, you will likely end up with many duplicate nodes, since GNE creates nodes it needs during the line file import. You might also want to import a polygon file, say for TAZs. Be sure to use the exact same transformation parameters for all files being imported, so that they align correctly. GNE retains the original long-lat information for all nodes, either imported or created.
- QRS II has many parameters. Some of them can be found on QRS II’s menus. Others are found in text files. QRS II can also input a variety of “Add” files, which provide, perhaps, large quantities of supplementary information. Most often, when you want to supply something special to QRS II, you must prepare that information in the correct format (see the QRS II Reference Manual for details), then you must also check a box on QRS II parameter menus to tell QRS II to look for that information. All parameters and add files must be placed in the QRS II “project” folder. For example, the most often seen Add file is AddVTrips.txt, which contains a vehicle trip table.
- Polygons are not free standing. Each polygon has an owner node, usually located near the middle of the polygon. For TAZs, it is most helpful for those polygons to have the centroid as its owner node. When you import polygons from a GIS shape file, GNE automatically creates an owner node at the exact mathematical center of gravity for the polygon. For TAZs, it is a good idea to shift the owner node from GNE’s newly created node to the centroid for that TAZ. GNE has a “Sketch Pad mode” for doing just this.
- It is usually poor practice to have multiple nodes at exactly the same pixel location on a network. When GNE detects this situation, it displays the top node in a cyan color. Use node move to separate those multiple nodes. If the multiple nodes require merging into a single node, GNE has a Sketch Pad mode for that purpose, as well. If your network has many cyan nodes, there may have been a problem with how the network was imported. It might be easier to redo the import than to merge all the nodes. Incidentally, GNE has a feature, called Calculate, that enables broad manipulation of the whole network. One Calculate function merges all nodes within a certain proximity of each other, so sometimes it is possible to short-cut this process with Calculate. A little knowledge of Calculate can really speed up network preparation. BTW, GNE does not warn of duplicate links or links that lie exactly under another link, but these situations are often accompanied by duplicate (cyan) nodes.
- QRS II checks for errors as it performs the calculations. It may take many QRS II runs to find and fix all the errors. When an error occurs, QRS II stops and displays a warning. In most cases, QRS II will also display the internal sequence number of the offending node, centroid or link. Much of the time that network element is to blame for the error, but not always. QRS II stops when it cannot do a required calculation, and that moment may not be perfectly indicative of where the fix should be applied. GNE lists a node’s sequence numbers on its Node Properties dialog box. Centroid numbers may be matched to centroid names through the output file NodeLabl.txt. However, there is no quick way to find links by sequence number. To help the search, use the FindLink.pgm program within GNE’s Calculate (See the QRS II Reference Manual for more details.)
I realize this is much more than 10 things. Be sure to first check the FAQ on AJHAssoc.com should QRS II do something you do not expect.
Alan Horowitz, Whitefish Bay, March 27, 2022
,