Corollary 1.1. Hermitian squares on the two-torus
Notebook numbering. The input theorem is Corollary 3.4 in Scheiderer (2006).
Statement. Let be a scalar Laurent polynomial with complex coefficients and finite support . Assume that takes real, nonnegative values on . There are finitely many Laurent polynomials such that No finiteness or nondegeneracy assumption is imposed on the zero set of . Each factor can also be multiplied by a monomial to make both its exponents nonnegative.
Proof of the torus corollary
Put The two equations define a nonsingular integral affine surface over . Indeed, their gradients have disjoint coordinate supports and are nonzero everywhere on the complex variety. The complexified coordinate ring is via , . The real locus of the original surface is the compact product of two circles.
Here is an explicit first operation on the input coefficients. In each term of , replace by and by , expand, and take the real part. The resulting class agrees with on the real locus because is real-valued there. Thus is nonnegative on that locus.
Scheiderer’s surface result gives with . For each , choose a real polynomial representative and substitute The resulting is Laurent polynomial and real-valued on the torus. Therefore evaluates to . Multiplication of by a suitable clears negative exponents while preserving its modulus on the torus. This proves the claim. At every zero of , every summand must vanish. ∎
Where the coefficients are needed
Both changes of variables are explicit. The existence step supplies elements of the coordinate ring. The cited statement does not specify their coefficients, their number, or a Fourier support to search. The divisors and in the inverse substitution are constants. No input-dependent polynomial denominator occurs in the final factors.
In his survey the author derives the surface theorem from a local SOS result in dimension two and a local-to-global membership criterion. The latter requires localizations at relevant maximal ideals, which may include nonreal points. Applying those two theorems certifies existence globally. The passages examined give no procedure that takes the Fourier coefficient array and returns those global summands.
Sources and scope
- Claus Scheiderer, “Sums of squares on real algebraic surfaces,” manuscripta mathematica 119 (2006), 395–410. The publisher’s abstract and bibliographic record were read. The numbered statement and its proof were traced through the author’s survey; the full 2006 proof was not available for line-by-line comparison. Published article
- Claus Scheiderer, “Positivity and sums of squares: a guide to recent results,” author’s PDF. Theorem 3.3.1 explicitly identifies the 2006 result as [Sch7], Corollary 3.4. The proof dependencies appear in Theorem 3.1.12 ([Sch7], Corollary 2.10) and Theorem 4.7 (citing the earlier local-ring paper). Page numbers below refer to the author’s PDF. Surface theorem, p. 28 · Localization, p. 25 · Local rings, p. 31
- Maria Charina, Mihai Putinar, Claus Scheiderer and Joachim Stöckler, “A Real Algebra Perspective on Multivariate Tight Wavelet Frames,” Oberwolfach Preprints 2012-11, §4.1, proof of Theorem 4.1. The authors apply Corollary 3.4 to the nonnegative defect of a normalized 2D mask satisfying the sub-QMF condition; the SOS factors of the defect and the resulting UEP filters play different roles. Preprint
Proof provenance. The surface theorem is an input from Scheiderer. The two coordinate substitutions and the monomial shift above give the notebook’s derivation of the torus specialization. This turn covers scalar input and existence over real/complex coefficients. Exact coefficient recovery, symmetry constraints on filters, and the OEP construction are separate questions.
작성 · 노트 작성자
독립 검토 · 독립 theorem reviewer · 저자 survey와 2012년 preprint, 토러스 적용 증명, 영점·분모·개수·계수 산출 범위 및 인용 판본을 대조했다. 2006년 출판본에서는 공개 초록과 서지만 확인했다.