The Cosine-Integral Gram Integral
Abstract
Positive dilations of the real cosine-integral tail have an absolutely integrable product and an exact Gram integral.
Theorem 1.1 (All positive scales, including equal scales).
Proof. Machine-checked in Lean as D5/S3/Fourier/Asymptotics/CosineIntegralGram.result (✓ std3). ∎
Source. Repository-derived.
Commentary.
For every pair of positive real numbers a and b, the product Ci(a|z|)Ci(b|z|) is Lebesgue integrable on the real line and its integral is pi/max(a,b). In particular, the square at scale a has integral pi/a. Ci is the existing real cosine integral, defined for positive x as sin(x)/x minus the absolutely convergent integral of sin(t)/t^2 from x to infinity.D5/S3/Fourier/Asymptotics/CosineIntegralLattice.cosineIntegral
On the positive half-line set S(x)=sinc(x) and T(x)=Ci(x)-S(x). The sine-tail representation and the fundamental theorem of calculus give T’(x)=S(x)/x. Consequently B(x)=x T(ax)T(bx) has derivative Ci(ax)Ci(bx)-S(ax)S(bx).
The tail is bounded by 1/x. Near zero, the normalized finite cosine-sum estimate at N=1 gives |Ci(x)|<=1+2K+|log(x)| for one K>0 and 0<x<=1. Thus sqrt(x)T(x) tends to zero. These estimates show that B tends to zero at both endpoints of the positive half-line.D5/S3/Fourier/Asymptotics/CosineNormalizedRemainder.result
Add the integral of S(at)S(bt) from zero to x to B(x), obtaining a primitive P with derivative Ci(ax)Ci(bx). For a=b this derivative is nonnegative. The finite limit of P at infinity and continuity at zero therefore establish square integrability by the nonnegative form of the fundamental theorem of calculus. The L2 product inequality then establishes integrability for arbitrary a and b, allowing the same primitive to transfer the integral to the sinc product.
The existing integral of (sin(x)/x)^2 is pi. Scaling gives the integral of (sin(cx)/x)^2 as pi|c|, with c=0 treated directly. Put p=(a+b)/2 and q=(a-b)/2. The sine-product identity expresses S(ax)S(bx), away from zero, as the difference of the squared sine quotients at p and q divided by ab. Its integral is therefore pi(p-|q|)/(ab)=pi/max(a,b). Reflection transfers the positive half-line calculation to the real line; the singleton at zero has Lebesgue measure zero.
References
- Truth anchor:
D5/S3/Fourier/Asymptotics/CosineIntegralGram.result - Truth anchor:
D5/S3/Fourier/Asymptotics/CosineIntegralLattice.cosineIntegral - Truth anchor:
D5/S3/Fourier/Asymptotics/CosineNormalizedRemainder.result - Dependency: D5/S3/Fourier/Asymptotics/CosineNormalizedRemainder
- Dependency: D5/S3/TotalVariation/Asymptotics/StatLeanFourierCore