Hi,
I have a set of locations on earth (100 - 1,000), given with latitude and longitude values. I want to create a network out of them by connecting those locations to additional "hubs". The overall distance in this network should be minimal. I want to use the haversine function as distance. Any suggestions how to infer the best locations for these additional hubs? How could I do this within reasonable compute time?
Thanks in advance,
Stefan
http://mathworld.wolfram.com/Geodesic.html
https://en.wikipedia.org/?title=Geodesic
Computation time/expense not sure !!
Read it when I was processing WGS-84 sentences.
Hi Aparna,
thanks for your suggestions, but I understand those geodesics just as the shortest connection between two points on a sphere. My haversine function is an approximation of that. However, the question is not so much to find a minimal distance between two points, but to minimize the distance of all paths in the network.
Thus, my question remains open.
Thanks,
Stefan
I guess , as said -you have to build a graph ! Assuming hub equates to nodes with a weight obtained from the haversine function value. Again with m # of path between source and destination the matrix will have n nodes. The adjacency matrix has to be built.This adjacency matrix can be used to find the shortest path ,time tested algorithms exists in graph theory (Dijkstra's algorithm).
As far as computation cost is considered it is nxm , choosing an algorithm may optimize this complexity.
- Badges
- Science topic
- Similar topics
- Biological Science
- Bioecology
- Systems Ecology
- Landscape Ecology
- Connectivity