Reputation: 307
I have a data set of N
nodes and K
distribution centers. If I know their respective (x,y) coordinates
What algorithm can I use to cluster these nodes keeping the balance between the distance node-center and number of nodes assigned to the centers?
For example, I could use as criteria, assign the nodes to their nearest center of distribution but this could produce distribution centers with many nodes, and I want to avoid this. Any help?
Upvotes: 0
Views: 52
Reputation: 77454
Formulate this as an optimization problem.
E.g. average distance, and a penalty factor when too many N are assigned to the same K.
Then use a standard optimization technique to find a solution.
Don't expect anything out of the box to support vague constraints like yours. You have to decide on what a good solution is, then have an optimizer search a good solution.
Upvotes: 1