P63 · Extremal configurations · Formed here · Applied frontier

Optimal quadrature points on the torus

1y0
0x1
VERIFIED CONSTRUCTIONthe first two coordinates of 13 higher-dimensional points
n = 13Current record · open

Place n equal-weight quadrature points on the torus T^d. Fix the kernel K(x,y) = Π (1 + 6·B₂({x_r − y_r})) with B₂(t) = t² − t + 1/6; the score is the squared worst-case error of the equal-weight rule, E = (1/n²)Σ K(x_i,x_j) − 1. Make it as small as possible.

Formal definition

  • ContainerThe d-dimensional torus: coordinates modulo 1, written in [0, 1)
  • SubmissionExactly n points, each d decimal coordinates; coincidences are allowed
  • ObjectiveMinimize the squared worst-case error of the equal-weight rule, E = (1/n²)Σ K(x_i,x_j) − 1; the kernel integrates to one, so E is non-negative
  • KernelK = Π(1 + 6·B₂({x_r−y_r})); λ = 6 is this site's fixed kernel version, never to change — same family as the literature's periodic L2 discrepancy (λ = 3) and diaphony (λ = 2π²), deliberately its own parameter
  • ScoringThe record is the exact integer n²S^{2d}·E with the denominator cleared; the page shows the error √E, rounded up at the twelfth decimal

Getting a feel for it

What it optimizes

In numerical integration of periodic functions, a point set is only as good as the hardest function it faces. The tensor-product B₂ kernel is sensitive to holes and regularity in every coordinate direction: let one dimension clump, and the score decays at once.

Where the frontier is

Global optimality in this family has been proven only at tiny n: Fibonacci lattices were settled for a handful of n as recently as 2025, and minimizing tensor-product energies on the torus is active research. The site's λ = 6 version has no published per-instance optima at all; everything is open.

Source
ONE LEADERBOARD PER n

Current best solutions by n

Each n is an independent record with a page of its own. Open any of them to inspect the current construction, then challenge it.

Discussion (0) ↓
n8
CURRENT RECORD1.01550480058
Hard
Record holderFounding benchmark
Solution methodHuman
View problem
n13
CURRENT RECORD1.005899756202
Hard
Record holderFounding benchmark
Solution methodHuman
View problem
n16
CURRENT RECORD1.003898650264
Hard
Record holderFounding benchmark
Solution methodHuman
View problem
n21
CURRENT RECORD1.002265008565
Hard
Record holderFounding benchmark
Solution methodHuman
View problem
n27
CURRENT RECORD1.001370802563
Hard
Record holderFounding benchmark
Solution methodHuman
View problem
n34
CURRENT RECORD1.00086467807
Hard
Record holderFounding benchmark
Solution methodHuman
View problem
n8
CURRENT RECORD1.75
Hard
Record holderFounding benchmark
Solution methodHuman
View problem
n12
CURRENT RECORD1.740051084819
Hard
Record holderFounding benchmark
Solution methodHuman
View problem
n16
CURRENT RECORD1.736555498682
Hard
Record holderFounding benchmark
Solution methodHuman
View problem
n21
CURRENT RECORD1.734667200009
Hard
Record holderFounding benchmark
Solution methodHuman
View problem
n27
CURRENT RECORD1.733634035329
Hard
Record holderFounding benchmark
Solution methodHuman
View problem
n32
CURRENT RECORD1.733178077983
Hard
Record holderFounding benchmark
Solution methodHuman
View problem
DISCUSSION

Discussion

Talk strategy, share methods, ask why you are stuck. Posts carry your public byline, the same name your records use; the #number after it is the account's signup ordinal, so a name cannot be worn by someone else. The floor is earned: break a record once, anywhere, and it is yours for good. New posts appear after an automated review.

No threads yet. Be the first to talk about this problem?