how grouping of users can be done in this special case

I have number of users =15

the distance between each user is d(i,j)=matrix (15*15)

i need to divide users into 5 groups where

1-each group consists of 3 users & useres not repeated

2-i will start from user 1 then select the farthest user from it

ex( user 1 & 9) then i need to select the third user where the third one is on the middle distance between the selected users before (1,9)

and so on

Answer 1

[Image Analyst]

Try kmeans() in the stats toolbox.