Wiki
Wiki

Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.

Updated

Exact repeated-simplex lattice diagnostics

../


The checker supplies finite author diagnostics for L271. Its proof, including the inherited analytic interfaces, is the warrant for the unbounded statement. No numerical Gaussian sum, independent verification or finite reduction of that statement is claimed.

Run from the repository root:

sh
uv run --no-sync python wiki/theory/integer_sequences/L271_repeated_simplex_extraction/evidence/main.py

Write q=2bq=2^b, d=bkd=bk, and let F\mathcal F consist of Boolean frequencies missing at least one whole block. The fixed domain checks:

  • Ten full lattices: (b,k)=(1,1),…,(1,5)(b,k)=(1,1),\ldots,(1,5), (2,1),…,(2,3)(2,1),\ldots,(2,3), and (3,1),(3,2)(3,1),(3,2). The product of the Smith invariants of the Walsh rows is their index in the saturated integer lattice, giving its exact covolume. Independently, fraction-free determinants of the monomial Gram matrix GS,T=2d−∣S∪T∣G_{S,T}=2^{d-|S\cup T|} give the square covolume. Both are compared with log⁡2det⁡K=bk(q−1)k−1/2\log_2\det K=bk(q-1)^{k-1}/2.
  • All 254 nonempty spectral subfamilies at (b,k)=(1,3),(2,2)(b,k)=(1,3),(2,2). Their Smith covolumes are compared exactly, after dividing the logarithm by rank, with the full lattice. This domain contains non-downsets.
  • All nonempty downsets inside F\mathcal F at (b,k)=(1,3),(1,4),(2,2),(3,2)(b,k)=(1,3),(1,4),(2,2),(3,2): respectively 18, 166, 25 and 361 downsets. Their mean degrees are compared with the full frequency family by integer cross-multiplication.

The code uses exact integers and fractions, the standard library, repository tools, and the retained Smith-form helper. Inputs are fixed in the code. No private file, cached success result or external solver is required. The run takes under thirty seconds and less than 64 MiB; --quick has the same coverage. A failed comparison or nonexact arithmetic step exits nonzero.