POST
|
What is the best way to offset the bus routes as well as stops? Thanks, Chris Pollard It depends on how you have created your bus route edges. I wrote a Python script that creates bus lines from points, so I simply shifted all of my points by 10m North, checked if any were still coincident, and moved these 10m East. If any were sitll coincident, I could have shifted them again. However I kept a copy of the bus stops as they were when they were coincident with the streets. Then I used the Point to Line tool, using a bus stop ID as a case field: there is now one line between each pair of points. I then ran my Python script to create bus lines, and now my network looks just like the diagram above. The points coincident with the streets I call 'BusEntrances'. The points offset from the streets, but connected to BusEntrances by a line, I call 'BusStops'. The lines between Stops and Entrances contain a wait time penalty, which you can set different FromTo and ToFrom time penalties for (two fields), representing getting on and getting off a bus, respectively. However I still don't know how to model tranfers from bus line to bus line, where the lines share stops (both Entrances and Stops), such as along a bus corridor/network spine/major transfer station. Any help?
... View more
12-05-2012
01:24 PM
|
0
|
0
|
731
|
POST
|
I am setting up the network to model transfers and have set the elevation field values for the bus routes to the bus route number. I currently have two junction features, one for accessing a bus route from the street and one for bus-to-bus transfers, the first of which connects a bus route to the street center line file and the second which connects bus lines at a given stop. The transfer junction has a penalty of ten minutes whether an agent is transferring "up" or "down". I'm pretty much doing the same thing as you are. I have solved your problem by splitting the wait time penalty in half: one half is applied when 'entering' a bus route, and one half is applied when 'exiting' a bus stop, with the time cost in the same field and directionality not considered. For instance, for a 5 minute wait penalty, 2.5 minutes applies when you enter, and 2.5 when you exit. You could of course define WaitTimeFromTo (entering) and WaitTimeToFrom (exiting) fields with an x minute time penalty for the FromTo, and a 0 minute time penalty for the ToFrom. The latter is a better option and I'm not sure why I'm still using the former, but the difference is minor and not applicable to what I'm using a network for. My network is structured, though, with 'BusEntrances' coincident with the streets, and spatially offset 'BusStops' (offset by a trivial distance), with an edge between each pair: these edges contain my wait times, which you can configure to vary based on direction, as recommended above. My bus routes have end vertices at BusStops, but not BusEntrances. However I am stuck at the point of creating a transfer penalty. I have elevation values of the same nature as yours (equal to the bus service number), but whenever I include them in my network, the bus lines lose all connectivty from the roads and from each other (as though they are in separate, non-intersecting planes, when really every route has at least one stop it shares with another route, and all routes connect to the streets). Ignoring the bus-elevation, my network conectivity is fine, but there is of course no penalty for 'hopping' between routes, causing some weird choices and travel times to arise. What I want is what you've already achieved here: a penalty for a 'vertical turn' between bus routes (but not applied when going from buses to streets). Can you please explain to me how you achieved this?
... View more
12-05-2012
12:49 PM
|
0
|
0
|
414
|
POST
|
Is it in the same coordinate system and projection as the rest of your layers?
... View more
01-18-2012
04:22 PM
|
0
|
0
|
141
|
POST
|
Have you checked your pixel type? Sounds like your mosaic raster is attempting to become 32-bit unsigned. Given your 16-bit inputs, make your output 16-bit too.
... View more
01-17-2012
04:38 PM
|
0
|
0
|
385
|
Online Status |
Offline
|
Date Last Visited |
11-11-2020
02:24 AM
|