Question 8.T.10: Let f : [a, b] → [c, d] be Riemann integrable. If φ : [c, d]...

Let f : [a, b] → [c, d] be Riemann integrable. If φ : [c, d] → \mathbb{R}  is continuous, then φ ◦ f is Riemann integrable on [a, b].

The blue check mark means that this solution has been answered and checked by an expert. This guarantees that the final answer is accurate.
Learn more on how we answer questions.

Let ε > 0 be given. We shall prove the existence of a partition P ∈ \mathcal{P} (a, b) such that

U (φ ◦ f ) − L (φ ◦ f ) < ε.

Since φ is continuous on the compact interval [c, d] , it is bounded and uniformly continuous. Consequently there is a real constant K such that

|φ(t)| ≤ K  for all t ∈ [c, d] ,        (8.12)

and if we set ε^{\prime} =\frac{ε}{2K + (b − a)}, we know that there is a δ > 0 such that

s, t ∈ [c, d] , |t − s| < δ ⇒ |φ (t) − φ (s)| < ε^{\prime}.        (8.13)

On the other hand, since f ∈ \mathcal{R}(a, b) , there is a partition P = \left\{x_{0}, x_{1}, …, x_{n}\right\} such that

U (f, P ) − L (f, P ) < ε^{\prime}δ.        (8.14)

Let

M_{i} = \sup \left\{f (x) : x ∈ [x_{i}, x_{i+1}]\right\}

M^{∗}_{i} = \sup \left\{φ (f (x)) : x ∈ [x_{i}, x_{i+1}]\right\}

m_{i} = \inf \left\{f (x) : x ∈ [x_{i}, x_{i+1}]\right\}

m^{∗}_{i} = \inf \left\{φ (f (x)) : x ∈ [x_{i}, x_{i+1}]\right\} .

We then have

U (φ \circ f ) − L (φ \circ f) = \sum\limits_{i=0}^{n−1}{(M^{∗}_{ i} − m^{∗}_{i} ) (x_{i+1} − x_{i})}

= \sum\limits_{i∈J_{1}}{(M^{∗}_{ i} − m^{∗}_{i} ) (x_{i+1} − x_{i})}

+ \sum\limits_{i∈J_{2}}{(M^{∗}_{ i} − m^{∗}_{i} ) (x_{i+1} − x_{i})} ,        (8.15)

where

J_{1} = \left\{i ∈ \left\{0, 1, . . . , n − 1\right\} : M_{i} − m{i} < δ\right\}

J_{2} = \left\{i ∈ \left\{0, 1, . . . , n − 1\right\} : M_{i} − m_{i} ≥ δ\right\} .

If i ∈ J_{1}, then

|f (x) − f (y)| < δ  for all x, y ∈ [x_{i}, x_{i+1}] ,

and therefore, using (8.13),

|φ (f (x)) − φ (f (y))| < ε^{\prime}  for all x, y ∈ [x_{i}, x_{i+1}] ,

which implies that M^{∗}_{i} − m^{∗}_{i} ≤ ε^{\prime}. Therefore

\sum\limits_{i∈J_{1}}{(M^{∗}_{ i} − m^{∗}_{i} ) (x_{i+1} − x_{i})} ≤ \sum\limits_{i∈J_{1}}{ε^{\prime} (x_{i+1} − x_{i}) ≤ ε^{\prime} (b − a)} .

Now from (8.14) we see that

ε^{\prime}δ >\sum\limits_{i=0}^{n−1}{(M_{i} − m_{i}) (x_{i+1} − x_{i})}

≥ \sum\limits_{i∈J_{2}}{(M_{i} − m_{i}) (x_{i+1} − x_{i})}

≥ δ \sum\limits_{i∈J_{2}}{(x_{i+1} − x_{i})} ,

hence

\sum\limits_{i∈J_{2}}{(x_{i+1} − x_{i}) < ε^{\prime}} .

Since M^{∗}_{ i} − m^{∗}_{i} ≤ 2K, we must have

\sum\limits_{i∈J_{2}}{(M^{∗}_{ i} − m^{∗}_{i} ) (x_{i+1} − x_{i})} < 2Kε^{\prime}.

Using these inequalities in (8.15) yields

U (φ \circ f ) − L (φ \circ f ) < ε^{\prime} (b − a) + 2Kε^{\prime} = ε.

Related Answered Questions

Question: 8.5

Verified Answer:

Since the function f (x) = x − x^{2}[/latex...
Question: 8.T.7

Verified Answer:

Let P = \left\{x_{0}, x_{1}, ..., x_{n}\rig...