Problemsheet 04 S
Problemsheet 04 S
Problemsheet 04 S
Material covered
□ Definition of a function 𝑓 ∶ 𝐴 → 𝐵 and composites, domain, codomain and image/range of a function;
□ Injective, surjective, and bijective functions; inverse functions.
□ The concept of natural domain of a real valued function of a real variable.
□ The graph of a function, and the horizontal line test for injectivity.
□ The hyperbolic sine and cosine functions sinh 𝑥 and cosh 𝑥.
Outcomes
After completing this tutorial you should
□ understand the concepts of domain, codomain and image/range of functions;
□ be able to calculate the image/range of various functions;
□ be able to prove whether given functions are injective, surjective or bijective and compute inverse functions;
□ identify the natural domain of real valued functions of a real variable;
□ work with the hyperbolic cosine and sine functions, and prove identities involving them.
Summary of essential material
The hyperbolic sine and cosine. The hyperbolic co- 𝑦
sine and hyperbolic sine functions are defined by cosh(𝑥)
𝑒𝑥 + 𝑒−𝑥 sinh(𝑥)
cosh 𝑥 =
2
𝑒𝑥 − 𝑒−𝑥 1 𝑒−𝑥
sinh 𝑥 = 𝑒𝑥
2 2
2
for all 𝑥 ∈ ℝ. They share many properties with the 0 𝑥
cosine and sine functions as shown in some questions 𝑒−𝑥
−
below. 2
The graph of the hyperbolic cosine function is the
shape of a hanging cable or chain attached at two ends.
Functions. Let 𝐴 and 𝐵 be sets. A function 𝑓 ∶ 𝐴 → 𝐵 is a rule which assigns exactly one element of 𝐵 to
each element of 𝐴. We write 𝑥 → 𝑓 (𝑥) to indicate the value 𝑓 (𝑥) assigned to 𝑥. The set 𝐴 is called the domain
of 𝑓 , the set 𝐵 the codomain of 𝑓 . The image or range of 𝑓 is im(𝑓 ) = {𝑓 (𝑎) ∣ 𝑎 ∈ 𝐴} ⊆ 𝐵.
The function 𝑓 is surjective or onto if im(𝑓 ) = 𝐵. To show that 𝑓 is surjective one has to show that for
every 𝑦 ∈ 𝐵 there exists 𝑥 ∈ 𝐴 such that 𝑓 (𝑥) = 𝑦.
The function 𝑓 is injective or one-to-one if every point in the image comes from exactly one element in the
domain. To show a function is injective prove
( )
𝑥1 , 𝑥2 ∈ 𝐴 and 𝑓 (𝑥1 ) = 𝑓 (𝑥2 ) ⇐⇒ 𝑥1 = 𝑥2
(the converse is obvious by definition of a function). The above means for all choices of 𝑥1 , 𝑥2 with 𝑓 (𝑥1 ) =
𝑓 (𝑥2 ) the implication has to be true.
The function 𝑓 is bijective or invertible if it is both injective and surjective. In that case there exists an
inverse function is the function 𝑓 −1 ∶ 𝐵 → 𝐴 defined by
( )
𝑓 −1 (𝑦) = the unique element 𝑥 ∈ 𝐴 such that 𝑓 (𝑥) = 𝑦 .
In practice, to find 𝑓 −1 we solve the equation 𝑦 = 𝑓 (𝑥) for 𝑥 ∈ 𝐴.
𝑦= 𝑏 𝑥
− 𝑏 𝑎
𝑎 𝑥 𝑦=
𝑎 𝑥
2
(c) Explain, using the graphs, why sinh ∶ ℝ → ℝ and cosh ∶ [0, ∞) → [1, ∞) are bijective. Sketch
the graphs of the inverse functions.
Solution: We can deduce injectivity using the horizontal line test (note that we are restricting the
domain of cosh, and so we are only considering the part of the graph with 𝑥 ≥ 0). Surjectivity is
also clear from the graphs. The graphs are obtained by reflection at the line 𝑦 = 𝑥, giving:
𝑦 𝑦
1 𝑥 𝑥
3. Let 𝐴 = {𝑧 ∈ ℂ ∣ Re(𝑧) ≥ 2 and −𝜋 < Im(𝑧) ≤ 𝜋}, and let 𝐵 be the image of 𝐴 under 𝑓 (𝑧) = 𝑒𝑧 .
(a) Sketch 𝐴 and 𝐵, and show that 𝑓 ∶ 𝐴 → 𝐵 is bijective.
Solution:
𝑖ℝ
𝑖ℝ
2 + 𝜋𝑖
Im 𝑧 = 𝜋
𝐴
ℝ 𝑒2 ℝ
Im 𝑧 = −𝜋
2 − 𝜋𝑖
𝑧-plane 𝑤-plane
This function is necessarily surjective (because the codomain is defined to be the range for this
function). To prove injectivity, suppose that 𝑧1 , 𝑧2 ∈ 𝐴 with 𝑒𝑧1 = 𝑒𝑧2 . Writing 𝑧1 = 𝑧1 + 𝑖𝑦1
and 𝑧2 = 𝑧2 + 𝑖𝑦2 we have 𝑒𝑥1 𝑒𝑖𝑦1 = 𝑒𝑥2 𝑒𝑖𝑦2 . Equating the modulii of these complex numbers and
using that |𝑒𝑖𝑦 | = 1 gives 𝑒𝑥1 = 𝑒𝑥2 , and so 𝑥1 = 𝑥2 . Hence 𝑒𝑖𝑦1 = 𝑒𝑖𝑦2 and thus 𝑦1 = 𝑦2 + 2𝑘𝜋
for some 𝑘 ∈ ℤ, however since −𝜋 < 𝑦1 ≤ 𝜋 and −𝜋 < 𝑦2 ≤ 𝜋 we have 𝑘 = 0, and thus 𝑦1 = 𝑦2 .
Thus 𝑧1 = 𝑧2 and the function is injective.
(b) Find a formula for the inverse function 𝑓 −1 ∶ 𝐵 → 𝐴.
Solution: If 𝑤 = 𝑒𝑧 = 𝑒𝑥 (cos 𝑦 + 𝑖 sin 𝑦) then |𝑤| = 𝑒𝑥 and Arg(𝑤) = 𝑦 (because −𝜋 < 𝑦 ≤ 𝜋).
Thus 𝑥 = ln |𝑤| and 𝑦 = Arg(𝑤), and so the inverse function is
4. A function 𝑓 ∶ ℝ → ℝ is called strictly increasing if 𝑥1 < 𝑥2 implies that 𝑓 (𝑥1 ) < 𝑓 (𝑥2 ).
(a) Show that if 𝑓 is strictly increasing then 𝑓 is injective.
Solution: If 𝑥1 ≠ 𝑥2 then either 𝑥1 < 𝑥2 or 𝑥2 < 𝑥1 . In the first case we have 𝑓 (𝑥1 ) < 𝑓 (𝑥2 ), and
in the second case 𝑓 (𝑥2 ) < 𝑓 (𝑥1 ). In both cases we have 𝑓 (𝑥1 ) ≠ 𝑓 (𝑥2 ), and so 𝑓 is injective.
(b) Show that if 𝑓 ∶ ℝ → ℝ and 𝑔 ∶ ℝ → ℝ are strictly increasing, then the composition 𝑔◦𝑓 ∶ ℝ →
ℝ is strictly increasing. Deduce that 𝑔◦𝑓 is injective.
3
Solution: Let 𝑥1 < 𝑥2 . Since 𝑓 (𝑥) is increasing we have 𝑓 (𝑥1 ) < 𝑓 (𝑥2 ), and since 𝑔(𝑥) is
increasing, we have 𝑔(𝑓 (𝑥1 )) < 𝑔(𝑓 (𝑥2 )). Thus 𝑔◦𝑓 is increasing. Thus 𝑔◦𝑓 is injective by the
previous part.
(c) Using the result of the previous part, and the fact that 𝑒𝑥 is strictly increasing, prove that cosh ∶ [0, ∞) →
ℝ is strictly increasing, and hence injective.
−1
Solution: By definition, cosh 𝑥 = 𝑔(𝑒𝑥 ) where 𝑔(𝑥) = 𝑥+𝑥2 . To apply the result of the previous
part, we take 𝐸 = [0, ∞). Certainly 𝑒𝑥 is increasing on 𝐸, and when restricted to this domain its
range is 𝐷 = [1, ∞). So all we need to check is that 𝑔(𝑥) is increasing on 𝐷. In other words, we
need to show that if 1 ≤ 𝑥1 < 𝑥2 , then
𝑥1 + 𝑥−1
1
𝑥2 + 𝑥−1
2
< .
2 2
After multiplying by 2 and rearranging, this inequality becomes
which is true because both factors on the left-hand side are strictly positive. So we have proved
that cosh 𝑥 is increasing on [0, ∞).
Remark: You could also use some calculus to check that cosh 𝑥 ∶ [0, ∞) → ℝ is strictly increasing,
because 𝑓 ′ (𝑥) = sinh 𝑥 > 0 for 𝑥 > 0. This implies that cosh 𝑥 ∶ [0, ∞) → ℝ is strictly increasing
(the proof of this fact will come when we discuss the Mean Value Theorem in a few weeks!).
5. Each formula below belongs to a function 𝑓 ∶ 𝐴 → 𝐵 where we take 𝐴 ⊆ ℝ to be the natural domain
of 𝑓 , and we take the codomain 𝐵 to be the image of the natural domain under 𝑓 . Thus each function is
automatically surjective. In each case find 𝐴, and decide if the function 𝑓 ∶ 𝐴 → 𝐵 is a bijection. If so,
find a formula for the inverse function.
𝑥−2
(a) 𝑓 (𝑥) = ,
𝑥+2
Solution: The natural domain (also known as the domain of definition) of 𝑓 is ℝ ⧵ {−2}, that is,
the set of all real numbers except −2. Observe that for any 𝑥, 𝑥−2
𝑥+2
≠ 1, and so 1 is not in the range.
𝑥−2
The number 𝑦 is in the range if the equation 𝑦 = 𝑥+2
has at least one solution for 𝑥 in the domain.
Rearranging this equation gives 𝑥 = 2(1+𝑦)
1−𝑦
, showing that for any 𝑦 ≠ 1, there is one and only one
2(1+𝑦) 𝑥−2
𝑥, namely 𝑥 = 1−𝑦 , such that 𝑦 = 𝑥+2 . Hence the range of 𝑓 is ℝ ⧵ {1} and the function
𝑥−2
𝑓 ∶ ℝ ⧵ {−2} → ℝ ⧵ {1}, 𝑓 (𝑥) =
𝑥+2
is a bijection. The inverse function is
2(1 + 𝑦)
𝑓 −1 ∶ ℝ ⧵ {1} → ℝ ⧵ {−2}, 𝑓 −1 (𝑦) = .
1−𝑦
(The graph of 𝑓 is a hyperbola with vertical asymptote 𝑥 = −2 and horizontal asymptote 𝑦 = 1.)
√
(b) 𝑓 (𝑥) = 2 + 5𝑥,
Solution: The natural domain is [− 25 , ∞). The range of 𝑓 is [0, ∞). For any 𝑦 ≥ 0, there is one
√
and only one 𝑥 such that 𝑦 = 2 + 5𝑥, namely 𝑥 = 15 (𝑦2 − 2). Hence
𝑓 ∶ [− 25 , ∞) → [0, ∞)
4
(c) 𝑓 (𝑥) = 𝑥|𝑥| + 1.
Solution: The natural domain is ℝ. When 𝑥 ≥ 0, 𝑓 (𝑥) = 𝑥2 + 1 and 𝑓 (𝑥) takes all values in
[1, ∞). When 𝑥 ≤ 0, 𝑓 (𝑥) = 1 − 𝑥2 and 𝑓 (𝑥) takes all values in (−∞, 1]. Hence the range of 𝑓
is ℝ. The function is injective as it is an increasing function on each of the intervals [0, ∞) and
(−∞, 0], and therefore increasing on the whole of ℝ. Hence 𝑓 ∶ ℝ → ℝ is a bijection and its
inverse function is 𝑓 −1 ∶ ℝ → ℝ, where
{ √
− 1 − 𝑦, 𝑦 < 1,
𝑓 −1 (𝑦) = √
𝑦 − 1, 𝑦 ≥ 1.
(b) The function cosh ∶ (−∞, 0] → [1, ∞) is also a bijection. Find a formula for its inverse function.
√
Solution: If 𝑥 ∈ (−∞, 0), we would take the negative sign, to obtain cosh−1 𝑥 = ln(𝑥− 𝑥2 − 1).
7. For what values of the constants 𝑎, 𝑏, 𝑐 (with 𝑏 ≠ 0) is the function with formula
𝑥−𝑎
𝑓 (𝑥) = and domain {𝑥 ∈ ℝ ∣ 𝑥 ≠ 𝑐∕𝑏}
𝑏𝑥 − 𝑐
equal to its own inverse? (Hint: It may help to draw the graph.)
Solution: Using the fact that 𝑏 ≠ 0, we can rewrite the formula for 𝑓 (𝑥) as follows:
𝑐−𝑎𝑏
1 2
𝑓 (𝑥) = + 𝑏 𝑐 .
𝑏 𝑥−
𝑏
Notice that if 𝑐 = 𝑎𝑏, then this becomes the constant function 𝑓 (𝑥) = 1𝑏 , which is clearly not injective
and hence has no inverse. So we must have 𝑐 ≠ 𝑎𝑏. Then the graph of 𝑓 (𝑥) is a hyperbola, with vertical
asymptote at 𝑥 = 𝑏𝑐 , and horizontal asymptote at 𝑦 = 1𝑏 . So the domain of 𝑓 is ℝ ⧵ { 𝑏𝑐 }, and the range of
𝑓 is ℝ ⧵ { 1𝑏 }. Setting 𝑦 = 𝑓 (𝑥) and solving for 𝑥, we get a formula for the inverse function:
𝑐−𝑎𝑏
𝑐 2
𝑓 −1
(𝑦) = + 𝑏 1 ,
𝑏 𝑦−
𝑏
which is another hyperbola, this time with domain ℝ ⧵ { 1𝑏 } and range ℝ ⧵ { 𝑏𝑐 }. In order to have 𝑓 = 𝑓 −1 ,
the domain of 𝑓 must equal the domain of 𝑓 −1 (which is the range of 𝑓 ), so we must have 𝑐 = 1.
Conversely, if 𝑐 = 1 then the two formulas become the same, so 𝑓 does equal its own inverse. To sum
up, the answer to the question is that 𝑓 is equal to its own inverse whenever 𝑏 ≠ 0, 𝑐 = 1, and 𝑎 ≠ 1∕𝑏.
5
(a) cosh(𝑥 + 𝑦) = cosh 𝑥 cosh 𝑦 + sinh 𝑥 sinh 𝑦
Solution: We have
cosh 𝑥 cosh 𝑦 + sinh 𝑥 sinh 𝑦
[ ][ ] [ ][ ]
= 12 (𝑒𝑥 + 𝑒−𝑥 ) 21 (𝑒𝑦 + 𝑒−𝑦 ) + 12 (𝑒𝑥 − 𝑒−𝑥 ) 12 (𝑒𝑦 − 𝑒−𝑦 )
[ ]
= 41 (𝑒𝑥+𝑦 + 𝑒𝑥−𝑦 + 𝑒−𝑥+𝑦 + 𝑒−𝑥−𝑦 ) + (𝑒𝑥+𝑦 − 𝑒𝑥−𝑦 − 𝑒−𝑥+𝑦 + 𝑒−𝑥−𝑦 )
= 14 (2𝑒𝑥+𝑦 + 2𝑒−𝑥−𝑦 )
= 12 (𝑒𝑥+𝑦 + 𝑒−(𝑥+𝑦) )
= cosh(𝑥 + 𝑦)
(b) sinh(𝑥 + 𝑦) = sinh 𝑥 cosh 𝑦 + cosh 𝑥 sinh 𝑦.
Solution: We have
sinh 𝑥 cosh 𝑦 + cosh 𝑥 sinh 𝑦
[ ][ ] [ ][ ]
= 12 (𝑒𝑥 − 𝑒−𝑥 ) 12 (𝑒𝑦 + 𝑒−𝑦 ) + 12 (𝑒𝑥 + 𝑒−𝑥 ) 12 (𝑒𝑦 − 𝑒−𝑦 )
[ ]
= 14 (𝑒𝑥+𝑦 + 𝑒𝑥−𝑦 − 𝑒−𝑥+𝑦 − 𝑒−𝑥−𝑦 ) + (𝑒𝑥+𝑦 − 𝑒𝑥−𝑦 + 𝑒−𝑥+𝑦 − 𝑒−𝑥−𝑦 )
= 14 (2𝑒𝑥+𝑦 − 2𝑒−𝑥−𝑦 )
= 12 (𝑒𝑥+𝑦 − 𝑒−(𝑥+𝑦) )
= sinh(𝑥 + 𝑦)
10. Let 𝐴 = {𝑧 ∈ ℂ ∣ Re(𝑧) < 1 and 2𝜋 < Im(𝑧) ≤ 4𝜋}, and let 𝐵 be the image of 𝐴 under 𝑓 (𝑧) = 𝑒𝑧 .
(a) Sketch 𝐴 and 𝐵, and show that 𝑓 ∶ 𝐴 → 𝐵 is bijective.
Solution:
𝑖ℝ
1 + 4𝜋𝑖 𝑖ℝ
𝐴
𝐵
1 + 2𝜋𝑖
𝑒 ℝ
ℝ
𝑧-plane 𝑤-plane
6
This function is necessarily surjective (because the codomain is defined to be the range for this
function). To prove injectivity, suppose that 𝑧1 , 𝑧2 ∈ 𝐴 with 𝑒𝑧1 = 𝑒𝑧2 . Writing 𝑧1 = 𝑧1 + 𝑖𝑦1 and
𝑧2 = 𝑧2 + 𝑖𝑦2 we have 𝑒𝑥1 (cos 𝑦1 + 𝑖 sin 𝑦1 ) = 𝑒𝑥2 (cos 𝑦2 + 𝑖 sin 𝑦2 ). Equating the modulii of these
complex numbers gives 𝑒𝑥1 = 𝑒𝑥2 , and so 𝑥1 = 𝑥2 . Also, 𝑦1 = 𝑦2 + 2𝑘𝜋 for some 𝑘 ∈ ℤ, however
since 2𝜋 < 𝑦1 ≤ 4𝜋 and 2𝜋 < 𝑦2 ≤ 4𝜋 we have 𝑘 = 0, and thus 𝑦1 = 𝑦2 . Thus 𝑧1 = 𝑧2 and the
function is injective.
(b) Find a formula for the inverse function 𝑓 −1 ∶ 𝐵 → 𝐴.
Solution: If 𝑤 = 𝑒𝑧 = 𝑒𝑥 (cos 𝑦 + 𝑖 sin 𝑦) with 𝑧 = 𝑥 + 𝑖𝑦 ∈ 𝐴 then |𝑤| = 𝑒𝑥 . Moreover,
Arg(𝑤) = 𝑦−3𝜋, because 𝑦 is an argument of 𝑤, and since 2𝜋 < 𝑦 ≤ 4𝜋 we have −𝜋 < 𝑦−3𝜋 ≤ 𝜋,
thus 𝑦 − 3𝜋 is the principal argument of 𝑤. Thus 𝑥 = ln |𝑤| and 𝑦 = Arg(𝑤) + 3𝜋, and so the
inverse function is
𝑓 −1 (𝑤) = ln |𝑤| + 𝑖(3𝜋 + Arg(𝑤)).
You should compare the with Question 3.
11. Let 𝑓 (𝑥) = 𝑥3 , considered as a function 𝑓 ∶ 𝐴 → 𝐵 for the various 𝐴 and 𝐵 listed below. In each case
decide whether 𝑓 is injective and weather 𝑓 is surjective.
(a) 𝑓 ∶ ℝ → ℝ
Solution: Bijective.
(b) 𝑓 ∶ ℤ → ℤ
Solution: Injective, but not surjective.
(c) 𝑓 ∶ ℚ → ℚ
Solution: Injective, but not surjective.
(d) 𝑓 ∶ {−1, 0, 2} → {−1, 0, 8}
Solution: Bijective.
(e) 𝑓 ∶ [0, 1] → [−1, 1]
Solution: Injective, not surjective.
(f) 𝑓 ∶ [0, ∞) → [0, ∞)
Solution: Bijective.
12. Explain why the functions given by the formulas and domains below are injective. Find their ranges and
formulas for their inverses.
(a) 𝑓 (𝑥) = 𝑥2 + 𝑥, 𝑥 ≥ − 21 .
Solution: Because 𝑥 → 𝑥2 is an increasing function on [0, ∞), 𝑓 (𝑥) = (𝑥 + 21 )2 − 14 is an
increasing function on the domain [−1∕2, ∞). It is therefore injective. As 𝑥 runs over [−1∕2, ∞),
(𝑥 + 12 )2 runs over [0, ∞), so 𝑓 (𝑥) runs over [−1∕4, ∞). Thus the range is [−1∕4, ∞).
√
Solving the equation 𝑦 = 𝑥2 + 𝑥 for 𝑥 gives 𝑥 = − 12 ± 𝑦 + 14 . As we are only interested in the
case that 𝑥 ≥ − 21 , we take the positive square root. Thus we get the following rule for the inverse
function: √
𝑓 −1 (𝑦) = − 21 + 𝑦 + 14 .
√
(b) 𝑔(𝑥) = 4 𝑥, 𝑥 ≥ 0.
Solution: 𝑔(𝑥) is injective by definition, because it is defined to be the inverse of the bijection
𝑓 ∶ [0, ∞) → [0, ∞) given by 𝑓 (𝑦) = 𝑦4 . (To spell out the proof of injectivity for this particular
√ √
case, if we have 4 𝑥1 = 4 𝑥2 then we must have 𝑥1 = 𝑥2 , just by raising both sides to the fourth
power.) So the range of 𝑔(𝑥) is [0, ∞) and its inverse is
𝑔 −1 (𝑦) = 𝑦4 .
7
1 + 𝑒𝑥
(c) ℎ(𝑥) = , 𝑥 ≠ 0.
1 − 𝑒𝑥
2
Solution: Note that ℎ(𝑥) = −1 + 1−𝑒 𝑥
. Because 𝑥 → 𝑒𝑥 is an increasing function, 𝑥 → 1 − 𝑒𝑥 is a
decreasing function, so ℎ is increasing on (−∞, 0) and also on (0, ∞). It is not, however, increasing
on the whole domain ℝ ⧵ {0}: for example, ℎ(−1) > ℎ(1). Nevertheless, ℎ is injective, because it
takes positive values on (−∞, 0) and negative values on (0, ∞), so no horizontal line cuts the graph
more than once. In fact, as 𝑥 runs over (−∞, 0), ℎ(𝑥) takes all values in (1, ∞); and as 𝑥 runs over
(0, ∞), ℎ(𝑥) takes all values in (−∞, −1). So the range of ℎ is (−∞, −1) ∪ (1, ∞) = ℝ∖[−1, 1].
2 2
To find the inverse function, set 𝑦 = ℎ(𝑥) = −1 + 1−𝑒 𝑥
. Rearranging gives 𝑒𝑥 = 1 − 𝑦+1 , so
2
𝑥 = ln(1 − 𝑦+1
) = ln( 𝑦−1
𝑦+1
). Thus, for 𝑦 < −1 or 𝑦 > 1, we get:
( )
𝑦−1
ℎ−1 (𝑦) = ln .
𝑦+1
√
(d) 𝑓 (𝑥) = ln(3 + 𝑥 − 4), 𝑥 ≥ 5.
Solution: 𝑓 is injective because it is strictly increasing on the given√ domain;this is because
√
𝑥 is an increasing function and so is ln(𝑥). As 𝑥 runs over [5, ∞), 𝑥 − 4 runs over [1, ∞),
√
so ln(3√+ 𝑥 − 4) runs over [ln(4), ∞). Thus the range is [ln(4), ∞). Solving the equation 𝑦 =
ln(3 + 𝑥 − 4) for 𝑥 gives 𝑥 = (𝑒𝑦 − 3)2 + 4, so the formula for the inverse function is
13. Is the following statement true or false? “A function 𝑓 ∶ ℝ → ℝ is injective if and only if 𝑓 is either
strictly increasing or strictly decreasing.” If you think it is true, give a proof. If you think it is false, give
a counterexample.
Solution: It is true to say that if 𝑓 is strictly increasing or strictly decreasing, then 𝑓 is injective. For
if 𝑥1 and 𝑥2 are distinct elements in the domain such that 𝑥1 < 𝑥2 , then either 𝑓 (𝑥1 ) > 𝑓 (𝑥2 ) (if 𝑓 is
decreasing) or 𝑓 (𝑥1 ) < 𝑓 (𝑥2 ) (if 𝑓 is increasing). In both cases, 𝑓 (𝑥1 ) ≠ 𝑓 (𝑥2 ). This shows that distinct
inputs produce distinct outputs. However, the converse is not true. As a counter-example, consider the
function 𝑓 with domain ℝ given by the formula
⎧
⎪𝑥, 𝑥 < 0,
⎪1, 𝑥 = 0,
⎪
𝑓 (𝑥) = ⎨𝑥, 0 < 𝑥 < 1,
⎪0, 𝑥 = 1,
⎪
⎪𝑥, 𝑥 > 1.
⎩
This is an injective function by the horizontal line test, but is not an increasing function, as 0 < 1 but
𝑓 (0) ≮ 𝑓 (1).
14. Give an example of functions 𝑓 ∶ 𝐴 → 𝐵 and 𝑔 ∶ 𝐵 → 𝐶 such that 𝑔 is surjective yet the composition
function 𝑔◦𝑓 ∶ 𝐴 → 𝐶 is not surjective.
Solution: Let 𝑓 ∶ ℝ → ℝ be the function 𝑓 (𝑥) = 𝑥2 , and let 𝑔 ∶ ℝ → ℝ be the function 𝑔(𝑥) = 𝑥. Then
𝑔 is surjective, yet 𝑔◦𝑓 ∶ ℝ → ℝ is the function (𝑔◦𝑓 )(𝑥) = 𝑥2 is not surjective.
15. Last week you proved a closed formula for 1 + 2 cos 𝑥 + 2 cos 2𝑥 + ⋯ + 2 cos 𝑛𝑥. Find a corresponding
‘hyperbolic’ version for 1 + 2 cosh 𝑥 + 2 cosh 2𝑥 + ⋯ + 2 cosh 𝑛𝑥.
8
𝑒𝑥 + 𝑒−𝑥
Solution: Since cosh 𝑥 = we have, using the formula for the summation of a geometric series,
2
∑
𝑛
∑
𝑛
∑
𝑛
∑
𝑛
−𝑘𝑥
1+2 cosh 𝑘𝑥 = 1 + (𝑒 𝑘𝑥
+𝑒 )=1+ 𝑥 𝑘
(𝑒 ) + (𝑒−𝑥 )𝑘
𝑘=1 𝑘=1 𝑘=1 𝑘=1
𝑒𝑥 − 𝑒(𝑛+1)𝑥 𝑒−𝑥 − 𝑒−(𝑛+1)𝑥
=1+ + ,
1 − 𝑒𝑥 1 − 𝑒−𝑥
where we have used the geometric sum formula. In the first fraction we multiply the numerator and
denominator by 𝑒−𝑥∕2 and in the second fraction we multiply the numerator and denominator by 𝑒𝑥∕2 ,
giving
1 1
−𝑒𝑥∕2 + 𝑒(𝑛+ 2 )𝑥 + 𝑒−𝑥∕2 − 𝑒−(𝑛+ 2 )𝑥
1 + 2 cosh 𝑥 + ⋯ + 2 cosh 𝑛𝑥 = 1 +
𝑒𝑥∕2 − 𝑒−𝑥∕2
1 1
𝑒(𝑛+ 2 )𝑥 − 𝑒−(𝑛+ 2 )𝑥
= 𝑥∕2 −𝑥∕2
(𝑒 − 𝑒 1 1
)
1
2
𝑒(𝑛+ 2 )𝑥 − 𝑒−(𝑛+ 2 )𝑥 sinh(𝑛 + 12 )𝑥
1 ( 𝑥∕2 )
= = .
𝑒 − 𝑒−𝑥∕2 sinh 𝑥2
2
This identity if valid for all 𝑥 ≠ 0. Alternatively, noting that cosh 𝑥 = cos(𝑖𝑥) we could have used the
formula from last week to derive the above identity.
16. Show that if 𝑓 ∶ 𝐴 → 𝐵 is bijective, then the inverse function 𝑓 −1 ∶ 𝐵 → 𝐴 is also bijective.
Solution: To see that 𝑓 −1 ∶ 𝐵 → 𝐴 is injective, suppose that 𝑓 −1 (𝑏1 ) = 𝑓 −1 (𝑏2 ). Then 𝑓 (𝑓 −1 (𝑏1 )) =
𝑓 (𝑓 −1 (𝑏2 )), and so 𝑏1 = 𝑏2 (since 𝑓 (𝑓 −1 (𝑥)) = 𝑥). To see that 𝑓 −1 ∶ 𝐵 → 𝐴 is surjective, note that if
𝑎 ∈ 𝐴 then taking 𝑏 = 𝑓 (𝑎) we have 𝑓 −1 (𝑏) = 𝑓 −1 (𝑓 (𝑎)) = 𝑎, and so 𝑓 −1 is surjective. Thus the inverse
function is bijective.
9
Solution: If 𝐴 has the same cardinality as 𝐵 then there is a bijection 𝑓 ∶ 𝐴 → 𝐵, and if 𝐵 has the
same cardinality as 𝐶 then there is a bijection 𝑔 ∶ 𝐵 → 𝐶, and hence 𝑔◦𝑓 ∶ 𝐴 → 𝐶 is a bijection
(by Question 17), and hence 𝐴 has the same cardinality as 𝐶.
(c) Show that if 𝐴 and 𝐵 have finitely many elements then 𝐴 and 𝐵 have the same cardinality if and
only if 𝐴 and 𝐵 have the same number of elements.
Solution: Suppose that 𝐴 and 𝐵 are finite sets, and write |𝐴| and |𝐵| for the number of elements
in each set.
If 𝐴 and 𝐵 have the same cardinality, then there is a bijection 𝑓 ∶ 𝐴 → 𝐵. Since 𝑓 ∶ 𝐴 → 𝐵 is
injective we have |𝐴| ≤ |𝐵| (because the elements 𝑓 (𝑎) with 𝑎 ∈ 𝐴 give |𝐴| distinct elements of
𝐵). Since 𝑓 ∶ 𝐴 → 𝐵 is surjective, we have |𝐴| ≥ |𝐵| (because the elements 𝑓 (𝑎) with 𝑎 ∈ 𝐴
give all of the elements of 𝐵). Thus |𝐴| = |𝐵|.
On the other hand, suppose that |𝐴| = |𝐵| = 𝑛, say. Let the elements of 𝐴 be 𝑎1 , … , 𝑎𝑛 and the
elements of 𝐵 be 𝑏1 , … , 𝑏𝑛 . Define 𝑓 ∶ 𝐴 → 𝐵 by 𝑓 (𝑎𝑖 ) = 𝑏𝑖 for 𝑖 = 1, … , 𝑛. This is a bijection,
and so 𝐴 and 𝐵 have the same cardinality.
19. We say that a set 𝐴 has the same cardinality as the set ℕ of natural numbers if there is a bijection 𝑓 ∶
ℕ → 𝐴. In this case we say that 𝐴 is countable. This means that we can write all of the elements of 𝐴 in
a list in which every element occurs exactly once:
𝑎0 , 𝑎1 , 𝑎2 , … ,
where 𝑓 (𝑗) = 𝑎𝑗 is a bijection 𝑓 ∶ ℕ → 𝐴. Thus, morally, 𝐴 has the “same size” as ℕ, because the
elements of 𝐴 are paired-up bijectively with the elements of ℕ.
(a) Show that ℤ is countable.
Solution: We list the elements of ℤ as
0, 1, −1, 2, −2, 3, −3, 4, −4, 5, −5, … .
Thus we have a bijection 𝑓 ∶ ℕ → ℤ given by 𝑓 (0) = 0, 𝑓 (1) = 1, 𝑓 (2) = −1, 𝑓 (3) = 2,
𝑓 (4) = −2, and so on. The general formula is
{
−𝑛∕2 if 𝑛 is even
𝑓 (𝑛) =
(𝑛 + 1)∕2 if 𝑛 is odd.
(b) Show that the set ℕ × ℕ = {(𝑚, 𝑛) ∣ 𝑚 ∈ ℕ and 𝑛 ∈ ℕ} is countable.
Solution: We list the elements of ℕ × ℕ according to the sum of their entries:
(0, 0)
(0, 1) (1, 0)
(0, 2) (1, 1) (2, 0)
(0, 3) (1, 2) (2, 1) (3, 0)
(0, 4) (1, 3) (2, 2) (3, 1) (4, 1)
⋮ ⋮ ⋮ ⋮ ⋮
Every element of ℕ × ℕ will occur in this list exactly once (specifically, the element (𝑚, 𝑛) will
occur as the (𝑚 + 1)-th entry in the (𝑚 + 𝑛 + 1)-th row). This gives a bijection 𝑓 ∶ ℕ × ℕ → ℕ by
setting:
𝑓 (0, 0) = 0
𝑓 (0, 1) = 1 𝑓 (1, 0) = 2
𝑓 (0, 2) = 3 𝑓 (1, 1) = 4 𝑓 (2, 0) = 5
𝑓 (0, 3) = 6 𝑓 (1, 2) = 7 𝑓 (2, 1) = 8 𝑓 (3, 0) = 9
𝑓 (0, 4) = 10 𝑓 (1, 3) = 11 𝑓 (2, 2) = 12 𝑓 (3, 1) = 13 𝑓 (4, 1) = 14
⋮ ⋮ ⋮ ⋮ ⋮
10
(c) Show that if 𝐴 and 𝐵 are countable then the set 𝐴 × 𝐵 = {(𝑎, 𝑏) ∣ 𝑎 ∈ 𝐴 and 𝑏 ∈ 𝐵} is also
countable.
Solution: If 𝐴 and 𝐵 are countable then there is a bijection 𝑓 ∶ 𝐴 → ℕ and 𝑔 ∶ 𝐵 → ℕ. Let
ℎ ∶ 𝐴 × 𝐵 → ℕ × ℕ be the function
0∕1
0∕2 1∕2
0∕3 1∕3 2∕3
0∕4 1∕4 2∕4 3∕4
0∕5 1∕5 2∕5 3∕5 4∕5
0∕6 1∕6 2∕6 3∕6 4∕6 5∕6
0∕7 1∕7 2∕7 3∕7 4∕7 5∕7 6∕7
⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮
Unlike in part (b), there are repetitions in this list resulting from the fact that the representation of
rational numbers is not unique, for example 12 = 24 = 36 = 48 = ⋯. Crossing out these duplicates
we obtain a list in which every rational number 𝑟 with 0 ≤ 𝑟 < 1 occurs exactly once:
0∕1
1∕2
1∕3 2∕3
1∕4 3∕4
1∕5 2∕5 3∕5 4∕5
1∕6 5∕6
1∕7 2∕7 3∕7 4∕7 5∕7 6∕7
⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮
11
(e) Deduce that the set ℚ of all rational numbers is countable.
Remark: This is rather surprising, since intuitively ℚ feels a lot “bigger” than ℕ.
Solution: Let 𝑋 be as in the previous part. We first observe that the function
Thus
𝑚 𝑟
= 𝑞 + = 𝑓 (𝑞, 𝑟∕𝑛).
𝑛 𝑛
Thus ℤ × 𝑋 and ℚ have the same cardinality. Since both 𝑋 and ℤ are countable (by parts (a) and
(d)) it follows from part (c) that ℚ is also countable.
(f) So perhaps every infinite set is countable? No: Show that the set of real numbers in the interval
[0, 1] is not countable.
Note: This is tough if you haven’t seen something like it before!
Solution: Suppose, for a contradiction, that [0, 1] is countable. Thus we can list all of the num-
bers on [0, 1] as 𝑟1 , 𝑟2 , 𝑟3 , 𝑟4 , … such that every number in [0, 1] appears exactly once in this list.
Writing the decimal expansions of these numbers we have a list:
We need to be a tiny bit careful here, because decimal expansions are not quite unique, for example
0.1499999999 ⋯ = 0.150000000 ⋯
(prove this!). So in the decimal expansions above we stipulate that we do always take the represen-
tative that ends in ⋯ 99999 ⋯, and so, for example, we take 0.1499999 ⋯ rather than 0.1500000 ⋯.
With this proviso, the decimal expansions given above for 𝑟1 , 𝑟2 , … are uniquely determined, and
by the assumption that [0, 1] is countable, every decimal expansion 0.𝑎1 𝑎2 𝑎3 𝑎4 ⋯ with no trailing
zeros must occur on our list.
Now we construct a number 𝑟 ∈ [0, 1] and show that it is not on our list. This contradicts the
assumption that our list was complete, and hence [0, 1] and ℕ do not have the same cardinality.
We construct 𝑟 as
{
𝑎𝑖𝑖 − 1 if 𝑎𝑖𝑖 ∈ {1, 2, 3, 4, 5, 6, 7, 8, 9}
𝑟 = 0.𝑎1 𝑎2 𝑎4 𝑎4 ⋯ where 𝑎𝑖 =
𝑎𝑖𝑖 + 1 if 𝑎𝑖𝑖 = 0.
In other words: We construct a number 𝑟 = 0.𝑎1 𝑎2 𝑎3 𝑎4 ⋯ such that the 𝑖-th digit in its decimal
expansion is different from the number 𝑎𝑖𝑖 , where 𝑎𝑖𝑖 is the 𝑖-th digit in the decimal expansion of
𝑟𝑖 . This number should appear somewhere on our list, since we assumed the list is complete. But
it doesn’t! To see this, note that 𝑟 ≠ 𝑟𝑖 , because 𝑟 and 𝑟𝑖 differ in their 𝑖-th digits (by construction
of 𝑟). Thus [0, 1] is not countable.
12
(g) The power set of a set 𝐴 is the set (𝐴) = {𝐵 ∣ 𝐵 ⊆ 𝐴} consisting of all subsets of 𝐴. For example,
if 𝐴 = {1, 2, 3} then = {∅, {1}, {2}, {3}, {1, 2}, {1, 3}, {2, 3}, {1, 2, 3}}, where ∅ = {} is the
‘empty set’. Show that for any set 𝐴 the set (𝐴) does not have the same cardinality as 𝐴. Hence
deduce that there is a set ‘bigger’ than ℝ, and that in fact there is an infinite number of growing
‘sizes’ of infinite sets.
Solution: Suppose that there is a bijection 𝑓 ∶ 𝐴 → (𝐴). Thus to each 𝑎 ∈ 𝐴 the function 𝑓
gives a subset 𝑓 (𝑎) of 𝐴. Consider the set
𝐵 = {𝑎 ∈ 𝐴 ∣ 𝑎 ∉ 𝑓 (𝑎)}.
Since 𝐵 is a subset of 𝐴 we have that 𝐵 ∈ (𝐴). Thus, since 𝑓 ∶ 𝐴 → (𝐴) is surjective, there
is an element 𝑏 ∈ 𝐴 such that 𝐵 = 𝑓 (𝑏). Is the element 𝑏 in 𝐵? If 𝑏 ∈ 𝐵 then by the definition of
𝐵 we have 𝑏 ∉ 𝑓 (𝑏), but since 𝑓 (𝑏) = 𝐵 this says that 𝑏 ∉ 𝐵, a contradiction. On the other hand,
if 𝑏 ∉ 𝐵 then since 𝐵 = 𝑓 (𝑏) we have 𝑏 ∉ 𝑓 (𝑏), and thus by the definition of 𝐵 we have 𝑏 ∈ 𝐵,
again a contradiction!
Thus no bijection 𝑓 ∶ 𝐴 → (𝐴) exists, and so 𝐴 and (𝐴) do not have the same cardinality.
Hence the cardinality of (𝐴) is a ‘larger’ cardinality than 𝐴 (it certainly is not ‘smaller’, because
there is an injective function 𝑓 ∶ 𝐴 → (𝐴) given by 𝑓 (𝑎) = {𝑎}).
Applying this to ℝ we see that the power set (ℝ) has a strictly ‘larger’ cardinality than ℝ. So
(ℝ) is a very very large set indeed! But it does not stop here, we can take the power set of the
power set of ℝ (that is, ((ℝ))), and then the power set of the power set of the power set of ℝ,
and so on. Each is larger in cardinality than the last, and so we get a growing chain of larger and
larger cardinalities. Infinity certainly is an interesting beast.
Remark: The topics covered in this exercise were first discovered by the German mathematician
Georg Cantor around 1870. He is generally considered as the father of set theory, and as the
first person to really understand the complexity of infinity. His work shook the foundations of
mathematics. Indeed many prominent mathematicians at the time rejected his ideas – for example,
Henri Poincaré said that Cantor’s theory was a “grave disease” infecting mathematics. However
over time people had to accept the strangeness of infinity. In fact the complexity of infinity has
lead to a far more beautiful world of mathematics. Sadly the hostility directed towards Cantor
during his life only exacerbated the severe bouts of depression that plagued his later years, and he
died in relative poverty in 1918 before the end of World War I.
There are some nice books in which you can discover more. For example, Brian Clegg’s A Brief
History of Infinity is a great read. You can also go directly to the source and have a go at reading
Cantor’s surprisingly accessible book Contributions to the Founding of the Theory of Transfinite
Numbers.
13