using a linear solver to get a "centre of gravity" of a set of points

المشرف العام

Administrator
طاقم الإدارة
As a programmer, if I have a linear solver available. How can I formulate the following problem as a set of constrains / decisions / goals for the linear solver?

I have a matrix of say 10*10, so 100 points (potential depots).

I randomly select 10 of those points as "Customers".

Now I would like to "solve" the point(s) totalling the shortest distance from the 10 "Customers" and I would like the coordinates of that point.

One "depot" point is so easy it doesn't need a solver, however solving the two "depot" points that combine to give the shortest distances to all the "customers" is tough, so needs a solver.

I can use the K-means to do this, however I would like to know if and how I can use a linear solver to do the same thing.



أكثر...
 
أعلى