Hi,
I think you can now install QGIS without administration password
if you used the advanced OSGeo4w installer and choose (only me).
I you have access to Python, you could modify the following
algorithm found in
https://gitlab.com/njacadieux/upstream_downstream_shortests_path_dijkstra.
That currently makes and "all point" to "all point" Dijkstra but I
am in the process of making a Closest Facility solver using the
actual network as input and output (not strait lines). I
currently have other projects so this is not a top priority for me
(I could be convinced to place it on top of the list if you have a
budget...) but for now, I have a PHD to finish! This algorithm
does not take traffic rules into consideration. It was built for
rivers. You can see this module in action here
https://www.youtube.com/watch?v=qQrHcKtmr3o. Keep in mind that I
used this video to show a float point problem I had that has since
been solved.
The other problem is that it has no tolerance for network error (like disconnected edges in the network). I therefore just finish a (free) QGIS model that will find and correct network file errors. Look for "Fix Directional Network models" 1 to 3. Use QGIS 3.20. (https://plugins.qgis.org/models/)
This YouTube video explains what the 3 models do.
https://youtu.be/v61PafSByvM
Nicolas
I had a problem similar to what the Closest Facility solver in Network Analyst solves, i.e., for every point in an "incident" feature class, find the closest point in a "facility" feature class and find the shortest route between the two. Not having access to Network Analyst, though, I tried to do this using NetworkX, with a road network extracted from OpenStreetMap by means of OSMnx. However, the shortest routes returned by NetworkX consist of straight segments from one node of the road network to the next, not following the curves in the road. This may or may not be good enough for my purposes; I am yet to find out. Nicolas Cadieux has also suggested QNEAT3 in QGIS for finding shortest routes; based on my past experience, I may or may not have the authorization to install QGIS on my computer. What other alternatives to the Closest Facility solver in ArcGIS are there?
--
You received this message because you are subscribed to the Google Groups "networkx-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to networkx-discu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/networkx-discuss/20945446-dff1-4e9d-8f35-b3d67fddf41an%40googlegroups.com.
-- Nicolas Cadieux https://gitlab.com/njacadieux
Le 24 sept. 2021 à 13:32, Dmitriy Tarasov <dmitriy...@uconn.edu> a écrit :
Thank you! I've now found an answer about how to have NetworkX return the true geometry of the underlying network (rather than straight lines from node to node); there is a description at https://stackoverflow.com/a/62179444/9728072 .
To view this discussion on the web visit https://groups.google.com/d/msgid/networkx-discuss/db1be8ea-4e57-4495-9e27-5c5b826b740bn%40googlegroups.com.
Le 27 sept. 2021 à 11:51, Dmitriy Tarasov <dmitriy...@uconn.edu> a écrit :
The script I'm writing is similar to that example, except that there are multiple origins and destinations and, for every origin, the script finds the nearest destination and saves the path to it. I am now trying to incorporate the tie_outside_node() function into it, so that it can trace routes all the way to and from the points.
To view this discussion on the web visit https://groups.google.com/d/msgid/networkx-discuss/2244646f-9220-46c8-8c8f-0d09fdd9142fn%40googlegroups.com.
Le 8 oct. 2021 à 11:51, Dmitriy Tarasov <dmitriy...@uconn.edu> a écrit :
Am I right that the tie_outside_node() function returns a dictionary of all nodes, both the old one and the new, formerly outside nodes that have been tied into the graph?
To view this discussion on the web visit https://groups.google.com/d/msgid/networkx-discuss/54eddbf1-ab17-480b-81e4-338327f1e742n%40googlegroups.com.