how does the added cost of a barrier get calculated when generating service areas?

[SharynHickey1]

Hi

I am currently creating service areas at 1600m. I want to add a point and line barrier as an added cost. I have done this but I am unsure of the calculation that is being used in the service area tool, i.e., if i say that the line barrier increases distance by 100m and I put 100m in the cost field, does this get multiplied by the line length?

I have looked through the help and forums but I can't find an answer to how the added cost is calculated with the service area anywhere.

Any help or direction would be great!!

thanks,

Sharyn

[PatrickStevens]

Hello!

Here is the 10.1 Desktop help documentation on Barriers:

http://resources.arcgis.com/en/help/main/10.1/#/Barriers/004700000056000000/

Additive point barriers *ADD* cost and scaled polyline barriers *MULTIPLY* cost when they are traversed.  Service Area usually works by accumulating costs of travel in each direction until the cutoff is reached.  If you have an additive point barrier that is traversed, then the extra cost is added into travel at that point.  If you have a scaled polyline barrier that is traversed, then the cost of the scaled edge (original cost * scale value) is included in the travel.

For your case, what do the additive barriers represent?  What do the scaled polylines represent?  What information are you trying to find out through service area analysis?

[SharynHickey1]

thanks Patrick, that is good to know!

I am trying to model a walkable area from a person's house. I want to add time costs for stopping at road intersections and also for terrain.

The network data I have doesn't have elevation but I do have a DEM that covers the area. I was hoping to change walking speed/or add time impediments to roads depending on their slope. I attributed the road lines with a slope value (i got this from creating slope from the DEM and using zonal statistics as a table with the road lines as the boundary and then joining back to the road lines based on an ID). So I was then hoping to be able to say for example, if the slope is 10% add 2 seconds to the journey, but since the lines multiply the added cost I don't think will work. Do you have any suggestions?

I did consider adding elevation into the network but since the data didn't come with it, I thought that this might bring in connectivity issues, i.e., different elevations might occur and mean the road lines will no longer connect?? Also I need to be able to add a cost for the terrain. At the moment I am thinking I could add the cost at the start of the line segment as a point, the problem with this is that the whole line may not be traversed but the cost for the whole line being traversed is being added. 

The points are the intersections, and as this cost is added, it should work, say if 1 sec is added each time a point is passed.

thanks for your help! 🙂

sharyn

[PatrickStevens]

Hello, Sharyn!

For the cost at intersections, perhaps you would be better off using the Global Turn Delay Evaluator, rather than additive point barriers.  The GTDE should save you the trouble of making a point for every intersection.  It will also allow for the ability to vary the intersection cost depending on the type of roads and the direction of the turn.  For example, going straight on a local road across a major road should have a greater delay than turning right from a local road onto a local road.

Here is some documentation about the delay evaluator to help you get started:

About global turns
Establishing global turns


I'll get back to you about your slope questions.  In the meantime, here is an article that discusses slope and network datasets:


http://www.esri.com/news/arcuser/0708/files/burnaby_1.pdf

[SharynHickey1]

thanks! I will definitely take a look!!

[RobertGarrity]

Hello, Sharyn.
Could you achieve what you want by using a more common approach of scaling instead of adding? For example, if the grade is 10%, multiply the travel time by 1.1; if it's -10%, multiply by .95, .94, etc. Essentially, multiply by some number that matches, more or less, the values you are trying to create by adding 2 seconds.

If you can do that, I would also suggest creating a new cost attribute in your network dataset that has the original travel time scaled by the effect of slope. Would that work for you?

Robert