Linearity in the integrator (Riemann–Stieltjes)

Integrability and the integral are linear with respect to linear combinations of integrators
Linearity in the integrator (Riemann–Stieltjes)

Let f:[a,b]Rf:[a,b]\to\mathbb{R} and let α,β:[a,b]R\alpha,\beta:[a,b]\to\mathbb{R} be functions of . Suppose the abfdα\int_a^b f\,d\alpha and abfdβ\int_a^b f\,d\beta exist. Let c,dRc,d\in\mathbb{R} and define a new integrator γ=cα+dβ. \gamma=c\alpha+d\beta.

Proposition: The integral abfdγ\int_a^b f\,d\gamma exists and abfd(cα+dβ)=cabfdα+dabfdβ. \int_a^b f\,d(c\alpha+d\beta)=c\int_a^b f\,d\alpha+d\int_a^b f\,d\beta.

This is the “integrator-side” linearity of the Riemann–Stieltjes integral. Together with integrand-side linearity, it makes fdα\int f\,d\alpha bilinear in (f,α)(f,\alpha) (within the class where the integral exists).

Proof sketch: For a tagged , the Riemann–Stieltjes sums satisfy $ \sum f(t_i)\bigl(\gamma(x_i)-\gamma(x_{i-1})\bigr)

c\sum f(t_i)\bigl(\alpha(x_i)-\alpha(x_{i-1})\bigr) +d\sum f(t_i)\bigl(\beta(x_i)-\beta(x_{i-1})\bigr). IftherighthandsumsHAHAHUGOSHORTCODE695s3HBHBasthemeshtendsto If the right-hand sums as the mesh tends to 0$, then so does the left-hand sum, with the stated limit. A careful argument uses the definition of Riemann–Stieltjes integrability via control of upper/lower sums or via the for these sums.