Maximize Points in n-D Space with Distance Threshold

  • Thread starter Thread starter EvoG
  • Start date Start date
  • Tags Tags
    Point
AI Thread Summary
The discussion centers on maximizing the number of points in a finite n-dimensional space constrained to [0,1], ensuring that no two points are closer than a specified threshold t. For example, in two dimensions, the maximum number of points varies based on the distance threshold, with configurations resembling square corners or opposite corners of a square. The problem is likened to sphere packing, where the radius is set to t/2, but there is debate over whether spheres or cubes provide a better packing solution. The author seeks an approximation for large dimensions (n > 10,000) and acknowledges that a perfect formula is not necessary. The conversation highlights the complexity of the problem and its relation to existing mathematical theories.
EvoG
Messages
1
Reaction score
0
I have a problem and I'm not sure if it has been formulated and/or solved before.

Suppose you have a finite n-dimensional space in which each dimension is constrained to [0,1]. I need the maximum number of points p that can be put in this space such that no two points are closer than some threshold t.

For instance, if n = 2 and t = 1, then p = 4. Think of the corners of a square with sides of length 1.
If n = 2 and t = 1.4, then p = 2, since opposite corners of a square with sides of length 1 are > 1.4 distance apart.

In general, my n is going to be large (>10000) and my t will certainly be greater than 1. I don't need a perfect formula. An approximation that gets me to within a factor of about 2 of the actual solution would be good enough.

I suspect this problem may be reducible to the sphere packing problem which has only been solved for up to 8 dimensions, but I thought I'd check with the experts to see.
 
Mathematics news on Phys.org
It does look like a sphere packing problem with radius = t/2.
 
mathman said:
It does look like a sphere packing problem with radius = t/2.
Except for at the corners and edges of the enclosing "box".
 
But since the "t" is minimum possible distance, the corners not really relevant. Using spheres of radius t/2 should give the same result as using cubes with side length t.
 
HallsofIvy said:
But since the "t" is minimum possible distance, the corners not really relevant. Using spheres of radius t/2 should give the same result as using cubes with side length t.

I don't think so. Using cubes will fill up the box neatly. Spheres will pack more tightly.
 
Suppose ,instead of the usual x,y coordinate system with an I basis vector along the x -axis and a corresponding j basis vector along the y-axis we instead have a different pair of basis vectors ,call them e and f along their respective axes. I have seen that this is an important subject in maths My question is what physical applications does such a model apply to? I am asking here because I have devoted quite a lot of time in the past to understanding convectors and the dual...
Insights auto threads is broken atm, so I'm manually creating these for new Insight articles. In Dirac’s Principles of Quantum Mechanics published in 1930 he introduced a “convenient notation” he referred to as a “delta function” which he treated as a continuum analog to the discrete Kronecker delta. The Kronecker delta is simply the indexed components of the identity operator in matrix algebra Source: https://www.physicsforums.com/insights/what-exactly-is-diracs-delta-function/ by...
Back
Top