Set VRP cluster using overrides parameter

502
2
02-22-2023 07:29 PM
DimasPutrapangestu
New Contributor II

Hi, I have a question about VRP, namely how does parameter overrides work in the arcgis REST API?. There are 2 ways to set overrides that I know of, namely "RSPPenaltyFactor" and "optimize for local order". I know that "optimize for local order" has some specific rules that maybe this is not suitable for me to use because the VRP parameters are homogeneous. But for "RSPPenaltyFactor" I can't find out more, especially how to find out the required value. Maybe there is information that can help with this, I would appreciate any opinion given. Thank You

Tags (3)
0 Kudos
2 Replies
HeatherMoe
Esri Contributor

Hello,

We did a presentation on the use of those two overrides. You can see the slides here and the presentation here. As for the input for the REST API you would pass {"RSPPenaltyFactor": "10"} with the curly brackets and everything inside. 

From your question I am guessing you are working on a high density VRP problem. I would love to know more about your use case. 

DimasPutrapangestu
New Contributor II

Thanks for the source, I've read that before and have watched the video "Creating a High Density Route with a VRP Solver". But I have difficulty understanding the RSP value, especially the 10 value presented as an example. I have tried several values ranging from 1 to 200 and there is a significant difference between the cluster results that are generated when I download the output layer via the REST API.

Some case information:
- I have 10 vehicles with an outlet spread of up to 10,000
- Currently order outlets are high, reaching more than 300
- The vehicle that I have has only one specialty but the number of delivery orders between outlets is different
- I have used the zone route, although the results are effective, in my opinion, it is necessary to consider the number of orders and the needs of vehicles in a certain zone. Therefore I want to use parameter overrides

Expected output:
I want to have good cluster results in order to minimize costs


Please correct me if the RSP in question is "Restricted Set Partitioning". Thank You

0 Kudos