Hello, I am trying to solve a routing problem in Network Analysis by using VRP solver. I have many points to serve per each road segment so I am dealing more with an Arc Routing Problem than Vehicle Routing Problem, but there is no ARP solver in Network Analyst. So what I want to do is create a two mid-points for each segment that has some demand, one for each side of street (my problem has side dependency) and than sum the demand on each side of the street to this point.
Have someone ever try this with success?
I found that this approach is mentioned here: GIS high density vehicle routing soloution by jsandhu-esristaff on 25.6.2013 15:41
But there are not many details how to do this. My question is: Is there a way how to do this NOT manually. Some geoprocessing using the model builder, maybe? I know I will need Feature-to-point tool to get my mid points for each segment and then some sort of selection by location, but how to define that I want to place my new mid-point on the side of the segment and how to define the selection of present demand points from the same side of segment?
Any suggestions are strongly welcome.
Thank you.
If you want to load mid-points on both sides of a road then you can load them twice. Use the Add Locations GP tool and then in the Field Mappings, there is a curb approach property, choose Right Side of Vehicle and load them first. Then run the tool again but choose Left Side of Vehicle. Now you will have two points loaded in the same location but they have to be approached from a particular side of the road. Also for the route layer, make sure to set one of the analysis property of U-Turns at Junctions to be "Allowed Only at Dead Ends" to get paths which do not make u-turns at each stop and visit the other side!
Jay Sandhu
Thank you Jay Sandhu, I will try to do this. I have one other question. In the discussion I cited above you mentioned that there is "no guarantee that you will get the best possible route, but it will be close" What sort of not optimal solution it could mean? If I run my VRP solver on a high number of orders that are close together it happens that some orders on a one street segment are skipped and the truck returns for them latter.
Can this imperfection of the solution be caused by the VRP solver itself?
I used the workaround "Preserve route and relative sequence" for this you mentioned in the GIS high density vehicle routing soloution on 25.6.2013 15:17
"By adjust I mean you can take an existing VRP layer and for the Orders, set the Assignment rule to be "Preserve route and relative sequence" instead of the default "Override". Now you can update the sequence numbers and re-solve. This way you can take a solved layer and adjust some stops manually."
And it gave me a slightly better solution. I want to make sure that this is caused by that the VRP solver is not designed for high density tasks.
Thank you
Karel Kus