MAT237 Multivariable Calculus with Proofs

Problem Set 4
Due Thursday November 27, 2025 by 23:59 ET

Instructions
This problem set is based on Module E: Inverse and implicit functions (E1 to E7). Please read the Problem Set
FAQ for details on submission policies, collaboration rules, and general instructions.
• Tutorials on Tuesday November 25, 2025 will help you with Problem Set 4. You will work with peers
and get help from TAs. Before attending, seriously attempt these problems and prepare initial drafts.
• Submissions are only accepted by Gradescope. Do not send anything by email. Late submissions beyond
your grace budget are not accepted under any circumstance. Please read the Problem Set FAQ for details.
• Submit your polished solutions using only this template PDF. You will submit a single PDF with your
full written solutions. If your solution is not written using this template PDF (scanned print or digital) then
you will receive zero. Do not submit rough work. Organize your work neatly in the space provided.
• Show your work and justify your steps on every question, unless otherwise indicated. Put your final
answer in the box provided, if necessary.
We recommend you write draft solutions on separate pages and afterwards write your polished solutions here.
You must fill out and sign the academic integrity statement below; otherwise, you will receive zero.

Academic integrity statement

Full Name:

Student number:

Full Name:

Student number:

I confirm that:
• I have read and followed the policies described in the Problem Set FAQ.
• I have read and understand the rules for collaboration on problem sets described in the Academic Integrity
subsection of the syllabus. I have not violated these rules while writing this problem set.
• I understand the consequences of violating the University’s academic integrity policies as outlined in the
Code of Behaviour on Academic Matters. I have not violated them while writing this assessment.
By signing this document, I agree that the statements above are true.

Signatures: 1)

2)

Created by C. Davies, M. Khaqan, and A. Qiu. Last updated November 14, 2025 at 10:47:55.
Problems
1. Let U, V ⊆ Rn be open sets and let ϕ : U → V be a diffeomorphism. Let S ⊆ U be a set. Prove by definition
that S has non-empty interior (i.e., S o ̸= ∅) if and only if ϕ(S) has non-empty interior (i.e., ϕ(S)o ̸= ∅).

2. Let A be an m × n matrix. Define B = AT A and let λ1 , . . . , λn denote the eigenvalues of B. Define the function
f : Rn → R given by f (x) = ∥Ax∥2 .
(2a) Justify that f (x) attains a maximum M subject to the constraint ∥x∥ = 1. Hint: Use PS2 Q1.

MAT237 Problem Set 4 - Page 2 of 7 November 27, 2025

 (2b) Use the chain rule to show that ∇ f (x) = 2B x.

(2c) Define g : Rn → R by g(x) = ||x||2 . Prove that x ∈ Rn and λ ∈ R is a solution to the system
∇ f (x) = λ∇g(x), g(x) = 1 (1)
if and only if x is a normalized eigenvector of B with eigenvalue λ.

(2d) Use the previous parts to conclude that M = max f (x) = max{|λ1 |, . . . , |λn |}.
∥x∥=1

MAT237 Problem Set 4 - Page 3 of 7 November 27, 2025

3. Define the function F : R3 → R by F (x, y, z) = x 2 − y 2 z.
(3a) Show by definition that F (x, y, z) = 0 locally defines z as a C 1 function of (x, y) at the point (1, −1, 1).

MAT237 Problem Set 4 - Page 4 of 7 November 27, 2025

 (3b) Reprove the previous part using the implicit function theorem.

(3c) Larry is attempting to show that F (x, y, z) = 0 locally defines y as a C 1 function of (x, z) at (1, −1, 1).

x
1. Let V = (0, ∞)2 ⊆ R2 and W = (−∞, 0). Define ϕ : V → W by ϕ(x, z) = − p .
z
2. Notice that the partial derivatives

∂ϕ 1 ∂ϕ x
= −p , = 3/2
∂x z ∂z 2z

exist and are continuous for all (x, z) ∈ V , so ϕ is C 1 on its domain.

3. Furthermore,
x 2 x2
 ‹
F (x, ϕ(x, z), z) = x 2 − − p z = x2 − z=0
z z
for every (x, z) ∈ V .

4. By definition, the equation F (x, y, z) = 0 locally defines y as a C 1 function of (x, z) at (1, −1, 1).

In one or two complete sentences, explain the key error in Larry’s proof.

MAT237 Problem Set 4 - Page 5 of 7 November 27, 2025

 2
4. The set Mn of n × n matrices can
• be viewed
2
˜ as n -dimensional Euclidean space R
n×n
= Rn . For example, we
x x2
identify each 2 × 2 matrix X = 1 ∈ M2 with an element (x 1 , x 2 , x 3 , x 4 ) ∈ R4 .
x3 x4
(4a) Define the square map f : R4 → R4 by f (X ) = X 2 via matrix multiplication. That is,

f (x 1 , x 2 , x 3 , x 4 ) = (x 12 + x 2 x 3 , x 1 x 2 + x 2 x 4 , x 1 x 3 + x 3 x 4 , x 2 x 3 + x 42 ).

Show that D f (I2 ) = 2I4 where I n is the n × n identity matrix.

(4b) Prove that there exists an open set U containing I2 such that if A ∈ U, then there exists a matrix X ∈ M2
with f (X ) = A, that is, A has a square root. Hint: Use the inverse function theorem.

MAT237 Problem Set 4 - Page 6 of 7 November 27, 2025

 (4c) Writing X = [x i j ]i, j , for i, j ∈ {1, . . . , n}, the determinant map det : Rn×n → R is a real-valued polynomial
map. By linear algebra, the cofactor expansion of the determinant along the i-th row is given by
n
X
det(X ) = (−1)i+ j x i j det(X [i, j]),
j=1

where X [i, j] ∈ R(n−1)×(n−1) is the (i, j)-minor of X , which is itself an (n − 1) × (n − 1) matrix.

Use this formula to explain why

∂ det(X )
= (−1)i+k det(X [i, k]) for any i, k ∈ {1, . . . , n}.
∂ x ik

(4d) Define the set of matrices

S = {X ∈ Rn×n : det(X ) = 1}.
2
Use the previous part to show that S is a (n2 − 1)-dimensional smooth manifold of Rn .

MAT237 Problem Set 4 - Page 7 of 7 November 27, 2025