You are on page 1of 3

Midterm exam Solutions, MATH 515, Spring 11

Problem 1) (50 points) Explain the mathematical meaning of the word nearly
in the three Littlewood Principles of Lebesgue Integration. Solution: In all three
principles it refers to the fact that a certain property or condition holds for all points
in a measurable set except on a set of measure with measure as small as we want. The
first principle is that for each measurable set there is a finite union of intervals such
that the measure of the symmetric difference can be made as small as we want. In
Egoroff’s theorem it is uniform convergence except on a set with measure as small
as we want. In Lusin’s theorem it says that a measurable function is continuous
except on a set of measure as small as we want. Thus more mathematically stated
nearly means that for each ε > 0 we find a set E0 (depending on ε) such that the
corresponding property fails to hold only for x in a set E0 with m(E0 ) < ε. It does
not mean that the property holds a. e.˙!

Problem 2) (25 points) Let the function f be defined on a measurable set E.


(i) Show that the collection of sets A with the property that f −1 (A) is measurable
is a σ-algebra. Solution: Compare Solution to Problem 7 from Chapter 3. (ii) Is
the collection of sets {f (A)|A ⊆ E and f (A) is measurable} a σ-algebra? Justify
your answer! Solution: The problem did not specify σ-algebra of subsets of what
set? Without specification it could be interpreted as σ-algebra of subsets of R.
But of course R 6= R if f is not onto, and then we do not have a σ-algebra. Now
it also certainly cannot be a σ-algebra of subsets of f (E), which is the naturally
choice, because we know examples with E measurable but f (E) not measurable.
But if we assume f (E) is measurable then f (A) measurable will imply f (E) \ f (A)
is measurable and f (Ai ) measurable for i ∈ N will imply ∪i f (Ai ) is measurable
because countable unions of measurable sets are measurable.

Problem 3) (25 point) Suppose that the function f is defined on a measurable


set E and has the property that {x ∈ E|f (x) > c} is measurable for each rational
number c. Is f measurable? Justify your answer! Solution: Because for each
d ∈ R:
{x ∈ E|f (x) ≥ d} = ∩c∈Q,d>c {x ∈ E|f (x) > c},
if {x ∈ E|f (x) > c} is measurable for all rational numbers c then, because count-
able intersections of measurable sets are measurable, f is measurable.

Problem 4) (100 points) (Each correct answer counts 10 points, each wrong
answer counts -5 points; the total number of points on this problem is ≥ 0.) Just
mark by true if the statement is true, and by false if the statement is not true:
1. The inverse image of a measurable set under a measurable function is measur-
able. Solution: False. The function ψ in Ch. 21 maps a measurable set to
a non-measurable set. Its inverse is continuous and thus measurable and thus
has an inverse image of a measurable set, which is non-measurable.

2. There exists an increasing non-measurable function with domain an interval.


Solution: False. Chapter 4, Proposition 3.

3. The characteristic function of the Cantor set is a simple function. Solution:


True. A characteristic function χA is measurable if and only if A is measur-
able. The Cantor set is measurable.

4. For each nonempty subset of R the outer measure of the set is equal to its
measure. Solution: False. There exist non-measurable sets for which the
measure is not even defined.

5. Each set of measure zero is a Borel set. Solution: False. Chapter 2, Propo-
sition 22

6. RThere exists a nonnegative measurable function a. e. zero on a set E with


E
f > 0. Solution: False. Chapter 4, Proposition 9.

7. A sequence (fn ) of bounded measurable Rfunctions on a set of finite measure E


has a measurable limit function f and E f = limn→∞ fn . Solution:
R False.
First of course because of the typo, it had to be limn→∞ E fn . But even
without the typo because the sequence has to be uniformly bounded. Coun-
terexample page 78 top.

8. Each measurable function is a uniform limit of simple functions. Solution:


False. The limit in the Simple Approximation Theorem is not necessarily
uniform. Just take the function f (x) = x on R and notice that a simple
function can take at most finitely many values. Compare the Case 1/Case
2 discussion on page 63. Also compare Problem 13 on the same page where
uniform convergence is shown if simple is replaced by semisimple.

9. Each nonempty subset of R contains a non-measurable subset. Solution:


False. All subsets of sets of outer measure zero (and there are many nonempty
ones) are measurable.

10. Each measurable set is the union of an open set and a set of measure zero.
Solution: False. The correct statement is that each measurable set is the
union of a Gδ -set and a set of measure 0. For an explicit counterexample
consider the set of irrationals in [0, 1]. This set does not contain any open
subset but has measure 1, so it cannot be the union of an open set and a set
of measure 0.

You might also like