I am working on a Sales Person Deployment project and quite often come across requests for calculating centroid of a territory where work is distributed across 1000s of ZIP codes. Increasingly we get requests to calculate centroids based on travel distance and not straight line (or crow fly) distance.
- What is the most optimal way to calculating centroid based on travel time (i guess it is similar if not same to point of minimum distance)?
I have attempted to calculate distance (using Google API) between all the ZIP codes with workload and then take the approach of travelling Sales Man problem to come up with a starting point. But there are couple of challenges here:
- A sales person may cover multiple targets if s/he goes to a cluster and
- Using this method is time and resource intensive
Excuse any digressions from Forum rules as I am a novice here and catching up with the protocols. Looking fwd to be part of this interesting community.