Linear programming, smallest enclosing ball and center point
The computational geometry flavored LP programs here solves LP-type
(aka Generalized Linear Programming) problems, which are not
necessarily linear. They might also be faster for low dimensional
problems with many constraints. For traditional LP programs,
which can probably
handle much larger problems more robustly, see the
Linear programming FAQ.
Finding the smallest ball containing a family of balls is another LP-type problem. Very nice for manipulating objects surrounded by bounding balls.
A center point is kind of like a d-dimensional median; any hyperplane through a center point cuts the input point set into two roughly equal pieces. The truely hard-core will realize that center points belong on this page because the dumb way to find one is by linear programming.
By Mike Hohmeyer, then at U.C. Berkeley.
Available by ftp from ICEMCFD, Mike's company.
See the Web page for more info on the nifty algorithm and a link to the code.
By Ken Clarkson, Bell Labs.
Take a look at the Web page for more details and pointers to the code.
By David White, U.C.S.D.
Code by Ken Clarkson, of Bell Labs.
Comments to: firstname.lastname@example.org
Created: May 31 1995 --- Last modified: Thu Jun 1 14:29:20 1995