home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky sci.math:17504 rec.puzzles:8165
- Path: sparky!uunet!portal!cup.portal.com!lordSnooty
- From: lordSnooty@cup.portal.com (Andrew - Palfreyman)
- Newsgroups: sci.math,rec.puzzles
- Subject: Optimisation Problem With Groups
- Message-ID: <72617@cup.portal.com>
- Date: Wed, 30 Dec 92 06:38:04 PST
- Organization: The Portal System (TM)
- Distribution: world
- Lines: 29
-
- I am interested in finding an efficient algorithm for solving
- the following problem:
-
- System
- ------
- A bounded Euclidean 2-D surface with discrete coordinates (0<=x<=XMAX,
- 0<=y<=YMAX) is assigned numbers z(x,y) at each point. Nothing is known
- _a priori_ about this assignment. The range of z (0<=z<=ZMAX) is far
- higher than the total number of points;
- i.e. XMAX*YMAX / (dx*dy) << ZMAX/dz, where d(i) is the granularity.
-
- Problem
- -------
- Given a fixed range of z-values RZ (RZ > dz), the problem is:
- group contiguous points within the range such that, after covering
- every point on the surface, the final grouping minimises the total
- number of groups.
-
- Notes
- -----
- * "Contiguous" is taken to mean +/- dx or dy of the point under
- consideration, but not both (i.e. no diagonal connections).
- * The minimum size of a group is one point.
- * Note that RZ is a range = zmax-zmin, and we are free to choose the
- absolute value of one of {zmax, zmin} for any group.
- --------------------------------------------------------------------------
- | lord snooty @the giant | Would You Like Space Potatoes With That? |
- | poisoned electric head | andrew_-_palfreyman@cup.portal.com |
- --------------------------------------------------------------------------
-