Latest Contributions by JaySandhu

I will suggest that rather than trying to make a m by n network, use a grid. A M by N network can become very densely connected if M and N start becoming large. For example a 50 by 100 means there will be a 100 lines originating from each of the 50 locations. The path finding code which is used to compute the OD Cost Matrix on this resulting network will start to run slower. So I suggest that you use the GP tool Fishnet to create an appropriately dense or sparse grid that encompasses your point locations and build a network and solve VRP on it. This will more resemble a street network! f you still need a M by N directly connected network, I suggest that once you have this grid network, simply use the OD Cost Matrix tool on your M by N points and it will create the straight lines between all your points as part of the output. Jay Sandhu   ... View more

We added a concept of a Travel Mode to group a number of settings to use when performing network analysis. You can use the network dataset property page to set up a travel mode. For example you can have a Car travel mode, a Truck travel mode, a bicycle travel mode. Each will have some settings of the impedance attribute, restriction attributes and evaluator/parameter settings. ArcMap UI cannot use these travel modes, but Catalog can be used to set them. They are used by Pro and Enterprise if you publish logistic services using these network datasets. You can read more about travel modes here: Travel modes—Help | Documentation  Jay Sandhu ... View more

And here is the output from ArcMap with the Service Area Index and same trim settings. It matches Pro. ... View more

Thanks for including the data. So the issue is that Service Area polygon generation code was improved quite a bit both in performance as well as output polygon quality some years ago. To enable this new behavior, you need to turn on the service area index and rebuild the network. In ArcMap, this is on the network dataset property page under the optimizations tab. If you do this, then ArcMap will generate the same polygons you are seeing in Pro. In Pro, we enabled these optimizations by default. So even if the network dataset does not have the service area index built, it will build it on the fly and use it. That is why you are seeing different results from ArcMap and Pro.  Now, in your screen shot, the quality of the output polygons does not look nice in Pro (and will not look nice in ArcMap after the index is added). This is because by default the polygon trim is set to 100 meters and you have service area breaks in much larger numbers like 10000 meters. So in ArcMap, change the trim to 1000 meter and solve it again. You will get better and more realistic output. In Pro, the trim can be set when you first make the layer (it is not in the UI yet). So use the GP tool Make Service Area Analysis Layer and set the trim to 1000 meters in the output geometry section (you can see it in the screenshot below with the updated polygon output).   Jay Sandhu ... View more

You can look at how the tutorial network dataset's like Paris have the costs and restrictions setup for examples of setting up evaluators. Jay Sandhu ... View more

There is no pre-assignment of a demand point to a facility point. However you could simply remove that demand location from the problem and reduce the capacity of its intended facility by that amount. That way the rest of the demand will be allocated normally. Jay Sandhu ... View more

It could be few different reasons. The first is that the settings are not identical. They appear to be. But most likely the network dataset settings such as restrictions could be different. Pro works with a travel mode (custom in your case) and ArcMap works on some defaults if they are set. So check these settings in Pro. Another is that Pro has a slightly different snapping model. So the locations could be snapped to a different edge if there are restrictions present in that area. You could open the attribute table of the demand points and look at the network location fields (e.g. SourceOID) and see if they are the same between ArcMap and Pro, assuming you are using the same network dataset. Jay Sandhu ... View more

What kind of network dataset are your trying to build? That is, what role should the z values in the points play? Jay Sandhu ... View more

We do not impose any limits on the number of facilities or demand you can load into a location-allocation layer IF you are using your own local network dataset. But the online logistic services has limits as documented in the Usage Limits section here: Location-Allocation service—ArcGIS REST API: Directions and Routing Services | ArcGIS for Developers  So you can either use a local network dataset with desktop software like ArcGIS Pro, or you could stand your own services using Enterprise and change the limits for these services. Jay Sandhu ... View more

A network dataset models a road with both sides of travel. Based on the direction of digitization, cost of travel can be specified on the From-To and To-From directions. Normally when you set up a cost attribute, you specify the two fields on the feature class that provide these two costs. As long as you can set up the values for the uphill/downhill costs, they can be loaded into the network and the Service Area solver will be able to use them. When you solve the Service Area you can specify the travel direction as From the facility or To the facility. That will use the appropriate direction of the slope costs into account. The setting up of the data and the network dataset can be done with the Pro 2.5 using the property pages of the network dataset.  The main problem in setting up these cost fields will be loading the two fields (From-To and To-From) correctly from your DEM. You may need to do some scripting for that aspect. Jay Sandhu ... View more

In your script you are loading stops_home and stops_work which point to Blocks_CentroidOrigin and Blocks_CentroidDestination.  What do these two feature classes contain? Do they contain the same set of 1000 locations? If yes, then you are mapping A to A, B to B, C to C, etc., based on the routename property and thus getting only those pairs of paths. What you need to do is load A, 1000 times with routename 1 and then load A to A1000 locations with the same routename. That is how Route works. So if you cannot load in this way then use Closest Facility as it will solve 1 to many. Jay Sandhu ... View more

The best way is to use the Closest Facility solver. You make the CF layer. Then load your 1000 locations as Facilities and also as Incidents. Set the property to solve all 1000 routes. And solve. it will solve the 1000 by 1000 routes, i.e. a million routes and the shortest path and cost will be in the routes sub layer. If you need the individual links that make up each route, you can run the Copy Traversed Source Features. I suggest that you try on a 10 by 10 case to make sure you have the process correct before trying it on all 1000. As for your other issue: >I was trying to find the route using Make Route Layer, however, the output feature class did not contain any values. I understand that the route details from "A" to "A" is zero (0) but there should be info for "A to B" and "A to C", "A to D"....etc.  I do not follow. What stops did you load into the route layer? Two stops A and B? After solve, did you not see an output route displayed on the map? Or if you open the attribute table for the routes sub layer, do you not see any row for the route you solved? If you cannot get ANY route to solve then perhaps the network is not set up correctly.  Jay Sandhu ... View more

After you solve a route analysis layer, you can run the GP tool Copy Traversed Source Features that creates a new feature class for the edges that made up the route. Also creates feature classes for the junctions and turns. I did not understand your question on running a thousand times and NOT exporting it each time. What are your trying to do? More info about the tool: Copy Traversed Source Features—ArcGIS Pro | Documentation  Jay Sandhu ... View more

Maximize attendance is probably not the best problem type for container locations. It is meant for retail locations like fast food and has a linear decay in "attendance" up to the specified Cutoff.  >i actually want to find the least amount of containers to serve the whole area. I would say use a bracketing approach to run the Location-Allocation solver with the Maximize Capacitated Coverage a few times to narrow down the minimum number of facilities needed to serve your demand.  Jay Sandhu  ... View more

In the multi-modal network tutorial, one of the costs that is added is called PedestrianTime. This cost defines how fast you can walk on a road and how fast you can transfer through the metro stations to get on the metrolines. If you choose this cost to solve a route, it will find a path using the streets, and rail. You have set up a similar cost attribute in your network and use it to solve your routes. Jay Sandhu ... View more