Open akashShanmugraj opened 1 month ago
Having multiple systems at the optimal possible smaller scale, and routing new request for clustering based an additional condition can theocratically reduce the search time.
Like if you have a system for India and another system for Bangladesh, and suppose if you have many clusters in both systems, you need not check min distance of the thousands of clusters inside India if your request is coming from Dhaka.
Once again we run into the edge case: If Kolkata needs something and its available in both Mumbai and Dhaka, the multiple system wont even consider the nearest relief (Dhaka)
So its really a double edged sword
Seperate systems for country / sub continent / continent level? or one common system?