Taking a dataset of points and iteratively merging the two closest points. Repeat until dataset has been whittled down to the number of clusters you want.

• Educba

Applications

One idea I have for this is related to an Erlang whitepaper I once read. They used load balancing proxies players connected to in an MMO setting. Load balancers then connected players with a collection of world servers which ran the actual world simulations. On occasion the positions of each player across the logical game world were analyzed to determine how the physical servers workloads were to be partitioned. Areas of low density would be given a single physical server to run those whole areas, while areas of high density received more physical servers for less logical space.