B4.11

Probability and Measure
Part II, 2003

Let f:[a,b]Rf:[a, b] \rightarrow \mathbb{R} be integrable with respect to Lebesgue measure μ\mu on [a,b][a, b]. Prove that, if

Jfdμ=0\int_{J} f d \mu=0

for every sub-interval JJ of [a,b][a, b], then f=0f=0 almost everywhere on [a,b][a, b].

Now define

F(x)=axfdμ.F(x)=\int_{a}^{x} f d \mu .

Prove that FF is continuous on [a,b][a, b]. Show that, if FF is zero on [a,b][a, b], then ff is zero almost everywhere on [a,b][a, b].

Suppose now that ff is bounded and Lebesgue integrable on [a,b][a, b]. By applying the Dominated Convergence Theorem to

Fn(x)=F(x+1n)F(x)1n,F_{n}(x)=\frac{F\left(x+\frac{1}{n}\right)-F(x)}{\frac{1}{n}},

or otherwise, show that, if FF is differentiable on [a,b][a, b], then F=fF^{\prime}=f almost everywhere on [a,b][a, b].

The functions fn:[a,b]Rf_{n}:[a, b] \rightarrow \mathbb{R} have the properties:

(a) fnf_{n} converges pointwise to a differentiable function gg on [a,b][a, b],

(b) each fnf_{n} has a continuous derivative fnf_{n}^{\prime} with fn(x)1\left|f_{n}^{\prime}(x)\right| \leqslant 1 on [a,b][a, b],

(c) fnf_{n}^{\prime} converges pointwise to some function hh on [a,b][a, b].

Deduce that

h(x)=limn(dfn(x)dx)=ddx(limnfn(x))=g(x)h(x)=\lim _{n \rightarrow \infty}\left(\frac{d f_{n}(x)}{d x}\right)=\frac{d}{d x}\left(\lim _{n \rightarrow \infty} f_{n}(x)\right)=g^{\prime}(x)

almost everywhere on [a,b][a, b].