Math V2020 - Honors Linear Algebra
Professor A.J. de Jong,
Columbia university,
Department of Mathematics.
Much of this page was shamelessly copied from Robert Lipschitz's
page
from when he taught the course in the Fall of 2008.
Basic information:
- Time: TTh 2:40 -- 3:55 PM
- Place: Math 407
- Textbook: Linear Algebra by Klaus Jänich
- Office hours: Monday, 10:00-11:00 AM, 1:00-2:00 PM in Math, Rm 523
- Teaching assistant: Yifei Zhao. His help room hours
are Mondays, 2-4pm, in Room 406, Mathematics
-
Help room hours
- Final exam: TBA
Description and goals.
At heart, linear algebra is about linear equations and
linear transformations. Linear algebra's importance to both
mathematics and its applications rivals -- and perhaps exceeds --
that of calculus. Unlike calculus, however, linear algebra becomes
clearer in a somewhat more abstract setting -- that of vector spaces,
linear transformations and inner products. This course will discuss
abstract linear algebra. Two things will keep us anchored: a
multitude of concrete examples, computations and applications;
and proofs, to track what is true in general -- and why.
The main goals of the course are:
- To provide the essential tools from linear algebra needed
in mathematics, science and engineering.
- To provide an introduction to writing and reading mathematics,
suitable for taking more advanced courses or studying further
mathematics on one's own.
- To begin to explore a deep and beautiful subject, and its
relationships with other parts of mathematics and science.
How is V2020 different from V2010?
While there will be a lot of overlap between "Linear Algebra"
and "Honors Linear Algebra," there will also be several
differences. In particular, V2020 will:
- Present somewhat more material than V2010. In particular, we
will talk about Jordan Normal Form and
several versions of the Spectral Theorem.
- Work in a more abstract setting -- that of vector spaces and
linear maps -- than V2010. This requires a change in mind set at
the beginning, which can be difficult. Ultimately, however, the ideas
are clearer from a more abstract perspective.
- Involve writing proofs. By the end of the semester students should
be able to write clear mathematical arguments. (Students are not
expected to have had experience writing proofs before taking the
course.)
Consequently, the material in V2020 will be presented at a somewhat
higher pace than in V2010. Although students in V2020 will learn to
compute everything that students in V2010 do, there will be somewhat
less practice computing. We will also talk about some applications
of linear algebra but may not
spend as much time on applications as V2010.
You should strongly consider V2020 instead of V2010 if:
- You are planning to major in mathematics.
- You enjoy thinking about (abstract?) mathematical questions.
- You would like a sense of what more advanced mathematics courses are like.
- You are planning to take more advanced mathematics courses later.
- You want a deeper background in mathematics for your major or other studies.
You should not take V2020 if:
- You already took V2010 or V1207-1208.
- You do not like thinking about abstract mathematics
(or strongly prefer numbers to variables, say).
- You are only taking linear algebra because of a degree requirement.
Policies
Grading
The grade will be determined by a weighted average of the
homework scores, the midterm, and the final exam. Roughly the
percentages will be 30%, 30%, 40%.
Homework
There will be weekly problem sets due each Tuesday one week after they
appear on the webpage here (below). To hand in homework, please find the
dropbox marked V2020 on the fourth floor of the math building. You can hand
in late, but every day late will cost you 10% of the score on that set.
Late means after 6PM.
Missed exams
If you have a conflict with any of the exam dates, you must contact me
ahead of time so we can make arrangements. If you are unable to take the
exam because of a medical problem, you must go to the health center and
get a note from them -- and contact me as soon as you can.
Syllabus and schedule
This will change as we go along. Namely, the pace set by the
schedule below is too optimistic and we will move things later.
Date |
Material |
Textbook |
Announcements |
09/02 |
sets, maps |
1.1, 1.2 |
|
09/04 |
vector spaces, complex numbers, subspaces |
2.1, 2.2, 2.3 |
|
09/09 |
fields, independence, span |
2.5, 3.1 |
Problem set 1 due |
09/11 |
bases, dimension |
3.1, 3.2 |
|
09/16 |
dimension (proofs), linear maps |
3.4, 4.1 |
Problem set 2 due |
09/18 |
linear maps, matrices |
4.1, 4.2 |
|
09/23 |
matrices, rotations, multiplication |
4.2, 4.5, 5.1 |
Problem set 3 due |
09/25 |
multiplication, rank |
5.1, 5.2 |
|
09/30 |
elementary, inverting, Gaussian |
5.3, 5.5, 7.3 |
Problem set 4 due |
10/02 |
systems, Gaussian |
7.1, 7.3 |
|
10/07 |
Cramer's rule and determinants |
6.1, 6.2, 7.1 |
Problem set 5 due |
10/09 |
determinants |
6.3, 6.4, 6.5, 6.7 |
|
10/14 |
Review |
1.1 - 7.5 |
Problem set 6 due |
10/16 |
Midterm on material above |
1.1 - 7.5 |
|
10/21 |
Inner products, orthogonal vectors |
8.1, 8.2 |
Problem set 7 due |
10/23 |
Orthogonal maps, groups |
8.3, 8.4 |
|
10/28 |
Symmetric Bilinear forms, completing the square, subgroups, Orthogonal groups |
Ch 8 |
Problem set 8 due |
10/30 |
Eigenvalues, Characteristic Polynomial |
9.1, 9.2 |
|
11/4 |
Election day |
--- |
|
11/6 |
Polynomials |
9.4 |
Problem set 9 due |
11/11 |
Slef-adjoint, Symmetric |
10.1, 10.2 |
Problem set 10 due |
11/13 |
Principal axes |
10.3 |
|
11/18 |
Rank Thm, Jordan Normal Form, Nilpotent Endos |
11.2, 11.3, Extra |
Problem set 11 due |
11/20 |
Jordan Normal Form |
11.3, Extra |
|
11/25 |
Examples of JNF |
|
Problem set 12 due |
11/27 |
Thanksgiving |
--- |
|
12/2 |
Sylvester Inertia |
11.5 |
Problem set 13 due |
12/4 |
Review |
|
|
12/18 |
EXAM 1:10-4:00 Rm 407 |
All sections of the book listed above + extra material listed below |
Extra theory
Some of the material we discussed in the lectures is not in
the book. Here is a list of these things:
- Row echelon form. See for example
Wikipedia here
- Elementary matrices and their relationship with
row operations. See for example
Wikipedia here
- Symmetric bilinear forms and symmetric matrices. See the
two sections on Definition and Matrix representation in
wikipedia.
- Subgroups. See the first part of
wikipedia).
- If V is a Euclidean vector space and f : V ---> V is an endomorphism
of the vector space V, then f is self-adjoint if and only if the map
Bf : V x V ---> R, (v, w) |---> <f(v), w>
is a symmetric bilinear form.
- If V is a finite dimensional Euclidean vector space and
B : V x V ---> R
is a symmetric bilinear form, then there is a unique self-adjoint
endomorphism f : V ---> V of V such that B = Bf.
- Nilpotent endomorphisms. See Lecture 21.
There is a typo on page 149: the statement is missing
"(v_1, ..., v_n) is in Jordan Normal Form". There is a mistake
on page 153: the upper left corner of the displayed matrix is wrong
unless we use v_m, ..., v_1, w_1, ..., w_{dim(W)} as our basis.
- Proof of Jordan Normal Form. See Lecture 22.
Observe how one of the pages is a page from my own research; it is the unique
page without a page number, please skip it.
Problem sets
The book has this interesting feature where in each chapter there is
a test. When I ask below that you do these, it means that you go through
the test, recording your answers to the multiple choice questions, and
then you check in the back of the book as to whether you got them right.
If not, please look at the suggestions given below the answers as to
how to improve. Do not hand in the tests.
- Problem set 1
- Do test 1.3 from the book.
- From Section 1.5 do exercises 1.1,
1.2 (prove the displayed statement), and 1.3.
- Do test 2.4 from the book.
- From Section 2.9 do exercises 2.1 (list axioms you are using in
each step as explained in the book), 2.2, and 2.3.
- Problem set 2
- Prove, using the axioms, that in a field the product of two
nonzero elements is always nonzero.
- From Section 3.7 do exercises 3.1, 3.2, and 3.3.
- We will use R to denote the field of real numbers
and we will denote Rn to denote the usual
vector space of n-tuples of real numbers.
- Find a basis for V = {(x,y) in R2 | 3x + 5y = 0}.
What is the dimension of V?
- Find a basis for W = {(x,y,z) in R3 | x + 2y + 4z = 0}.
What is the dimension of W?
- The vector (10, -1, -2) is an element of W. Write it as a linear
combination of the basis vectors you found above.
- Let P be the vector space of all polynomials in x over the real
numbers. Prove that P is infinite dimensional.
- Problem set 3
- Do test 3.3 from the book.
- From Section 4.7 do exercise 4.1.
- Let F : R2 ---> R2 be the
linear map given by F(x, y) = (3x + 2y, -6x -4y).
- Find a basis for the kernel of F.
- Find a basis for the image of F.
- Draw the kernel of F and the set {(x, y) such that F(x, y) = (3, -6)}
in the same picture.
- How many solutions are there to the equation F(x, y) = (1, 1)?
- Let F : R3 ---> R3 be the
linear map given by F(x, y, z) = (2x + y + 8z, y + 2z, x + y + 5z).
- Find a basis for the kernel of F.
- Find a basis for the image of F.
- Find a vector v in R3 which is not in the image of F.
- Find all the solutions to the equation F(x, y, z) = (2, 0, 1).
- Give an example of a surjective linear map
R3 ---> R3.
(Yes, this is silly.)
- Let P be the real vector space of all real polynomials
which you have previously shown to be infinite dimensional.
This exercise shows you how infinite dimensional vector spaces
behave differently from finite dimensional ones.
- Give an example of an injective but not surjective linear map
f : P ---> P.
- Give an example of a surjective but not injective linear map
g : P ---> P.
- Problem set 4
- Do test 4.3 from the book.
- From Section 4.7 do exercise 4.2.
- Compute the following matrix multiplications
(sorry about the type setting, I do not know how
to make round braces using html)
- The following product
- The following product
- Compute the 8th power of the matrix
A =
in other words, compute the 8-fold self-product A A A A A A A A.
(Hint: to do this
you only need to do 3 matrix multiplications.)
- Let
A =
. Find the inverse B of A, in other words, the matrix whose
associated linear map is the inverse of the linear map associated to A.
Here I want you to use the following approach: write
B =
The condition that B is the inverse of A boils down to a system of linear
equations which you can solve.
- Multiplication with elementary matrices. No explanations necessary.
- Let E =
and let A be a 3x3 matrix. Describe in your own words:
What is the effect of left multiplying A by E?
What is the effect of right multiplying A by E?
- Let E =
and let A be a 3x3 matrix. Describe in your own words:
What is the effect of left multiplying A by E?
What is the effect of right multiplying A by E?
- In this problem we let
f : R2 ---> R2
be the linear map given by f(x, y) = (3x + 4y, -x).
Let v_1 = (1, 1), v_2 = (1, 0) and let w_1 = (2, 1) and w_2 = (0, 2).
What is the matrix A of f with respect to the bases (v_1, v_2) and
(w_1, w_2)? [In other words, your matrix A = (aij) should
have the property that f(v_j) = a1j w_1 + a2j w_2
for j = 1, 2.]
- Problem set 5
- Do the test 5.4 from the book.
- Read about row echelon form, for example on
Wikipedia here
- Read about elementary matrices and their relationship with
row operations, for example on
Wikipedia here
- Using elementary row operations (one per step) reduce each of the
following matrices to row echelon form:
and
1 |
2 |
3 |
4 |
1 |
-2 |
-3 |
-4 |
1 |
2 |
-3 |
-4 |
3 |
2 |
-3 |
-4 |
and
- Use the process described in class and in Section 5.5 of the book
to invert the matrix
A =
Please show all the steps.
- Use the process described in class and in Section 5.5 of the book
to invert the matrix
Please show all the steps.
- Use Gaussian elimination (row reduction on extended matrix)
to find all solutions in R3 to the following system
of equations
2x + 3y + z = 0, x + y + z = 0, 3x + 4y + 2z = 0, y + z = 0.
- Use Gaussian elimination (row reduction on extended matrix)
to find all solutions in R3 to the following system
of equations
3x + y - 3z = 14, 2x + y - 3z = 9, -2x - y + 4z = -8.
- Use Gaussian elimination (row reduction on extended matrix)
to find all solutions in R3 to the following system
of equations
x + y + 3z = 5, -2x - 2y - 6z = -20.
- Problem set 6
- Do test 6.6 from the book.
- Compute the determinant of the matrix
and please show explicitly how you did the computation.
- Compute the determinant of the matrix
0 |
5 |
0 |
0 |
0 |
12 |
11 |
2 |
0 |
10 |
3 |
13 |
0 |
0 |
14 |
9 |
8 |
6 |
1 |
7 |
0 |
15 |
0 |
0 |
4 |
and please show explicitly how you did the computation.
- Compute the determinant of the matrix
a |
b |
0 |
0 |
0 |
0 |
c |
d |
0 |
0 |
0 |
0 |
0 |
0 |
e |
f |
0 |
0 |
0 |
0 |
g |
h |
0 |
0 |
0 |
0 |
0 |
0 |
i |
j |
0 |
0 |
0 |
0 |
k |
l |
and please explain how you did the computation.
- Let A be an n-by-n matrix. Let λ be a scalar. Prove that
det(λ A) = λn det(A).
- Let A be the n-by-n matrix which has a 1 in the upper right hand
corner and 1's just below the diagonal and 0's everywhere else. For
example, if n = 4 then you get
0 |
0 |
0 |
1 |
1 |
0 |
0 |
0 |
0 |
1 |
0 |
0 |
0 |
0 |
1 |
0 |
.
Find a formula for the determinant of A in terms of n and
prove it carefully.
- Suppose that A is a n-by-n matrix all of whose entries are
even integers. Prove that det(A) is an even integer.
- Problem set 7
- Do test 7.4 from the book (I should have asked you to do this last week).
- Start reading section 8.1 about inner products.
- Problem set 8
- Do test 8.5 from the book.
- With the standard inner product on R3
compute the lengths of (1, 1, 1) and (1, 2, 3) and the
angle between them. The answer is not "nice" -- it is fine if you
explain exactly what the final calculation step is without
actually putting it into a calculator.
- From Section 8.7 do exercises 8.1, 8.2, 8.3.
- Use the Gram-Schmidt process to find an orthnormal bases for the
image of the matrix
with respect to the standard inner product.
- Let V = R2. Show that the rule defined by
= 2x_1y_1 + 3x_1y_2 + 3x_2y_1 + 5x_2y_2 defines an inner
product on V.
- Problem set 9
- Read up about symmetric bilinear forms and symmetric matrices
(use your notes, or the two sections on Definition and Matrix representation in wikipedia).
- For which values of t in R does the symmetric matrix
define a positive definite symmetric bilinear form. Be precise and
explain your argument.
- Read up about subgroups (use your notes, or the first part of
wikipedia).
- Consider the group R*=R - {0} of nonzero
real numbers with group law given by multiplication. Find a subgroup with
two elements and prove there is no subgroup with 3 elements.
- Show that if A is an nxn matrix with integer entries
which is invertible, then the inverse has integer entries
also if and only if the determinant of A is 1 or -1.
(Hint: In one direction use the determinental formula for the inverse.
For the other direction use that the determinant is multiplicative.)
- Use the result of the previous exercise to prove
GL(n, Z) = {A in M(n x n, Z) with det(A) = 1 or -1}
is a subgroup of GL(n, R) but that on the other hand
{A in M(n x n, Z) with det(A) not zero} is not a subgroup
of GL(n, R).
- Let V = {(x_1, x_2, x_3, ...) with x_n in C} the
complex vector space of infinite sequences of complex numbers. Addition
and scalar multiplication is componentwise, as in the case of
the standard vector space Cn. Let f : V ---> V be the
endomorphism which sends (x_1, x_2, x_3, ...) to (0, x_1, x_2, x_3, ...).
Show that f has no eigenvector.
- Compute the eigenvalues and eigenvectors of the matrix
Hint: First guess the eigenvalues and then compute the eigenvectors.
You may use theory from the book about how many eigenvalues there
can be even if we haven't covered it yet in the lectures.
- Problem set 10
- Do test 9.3 from the book.
- Do exercise 9.1 from Section 9.5 of the book.
- Let n, m be positive integers.
Let A, B be square matrices of size n and m.
Let C be an n x m matrix.
Consider the square matrix of size n + m which
in block form looks like this:
. What this means is that the entries
in the left-lower m x n block are all zero.
Show that the determinant of this matrix
is equal to the product of det(A) and det(B).
- Find the roots and the algebraic multiplicities of the polynomial
x^7 - x^6 - 6*x^5 + 10*x^4 + 5*x^3 - 21*x^2 + 16*x - 4 over the complex
numbers. Totally fine
if you use some computer algebra to do this. Just the answer is OK here.
- Find the roots and the algebraic multiplicities of the polynomial
x^4 + 3*x^2 + 1 over the complex numbers. Same remarks as previous
question.
- Problem set 11
- Do test 10.4 from the book.
- Find a real 3 x 3 matrix which has eigenvalue 11 with
algebraic and geometric multiplicity 1 and no other
eigenvalues. Please explain why the example works.
- Find a real 6 x 6 matrix A which has eigenvalue 13 with
algebraic multiplicity 4 and geometric multiplicity 2 and
eigenvalue 17 with algebraic and geometric multiplicty 2.
- From Section 10.5 of the book do exercise 10.1.
- Give an example of a 2 x 2 complex symmetrix matrix A which
is not diagonalizable over the complex numbers.
- Call a complex n x n matrix A Hermitian if its transpose
is equal to its complex conjugate: in other words the entry aij
is equal to the complex conjugate of aji. Recall that
to get the complex conjugate of a complex number, you change the sign
of the imaginary part. In particular, a real symmetrix matrix is
Hermitian when viewed as a complex matrix. Show that a nonzero
2 x 2 Hermitian matrix A has two distinct real eigenvalues (in particular
it is diagonalizable over the complex numbers).
[Edit 11/24/2014: This is not quite correct, it should say that the complex
eigenvalues of A are real and that A is diagonalizable over the
complex numbers.]
- Problem set 12.
- Let f : V ---> V be an endomorphism of a vector space.
Let v be a vector and m > 0 an integer such that fm(v) = 0 but
fm - 1(v) is not zero. Prove that
v, f(v), f2(v), ..., fm - 1(v)
are linearly independent.
- Let U, W be subspaces of a vector space V such that U ∩ W = 0.
Let v be a vector in V which is not in U + W. Let W' be the span of
W and v, in other words
W' = {w + λ v | w ∈ W, λ is a scalar}.
Show that U ∩ W' = 0.
- Do exercise 11.2 from Section 11.7 of the book.
- For every triple a, b, c of complex numbers
determine the Jordan normal form for the matrix
0 |
0 |
abc |
1 |
0 |
-ab - ac - bc |
0 |
1 |
a + b + c |
Hint: Compute the characteristic polynomial and find its roots in terms
of a, b, c. Then carefully distinguish the cases where some of the roots
coincide.
- Suppose a linear self map f : V ---> V of a finite dimensional
complex vector space has eigenvalues λi where
i = 1, ..., r. What are the eigenvalues of f^n?
- Is the result you stated (and argued) in the previous execise
also true for an endomorphism of a finite dimensional real vector space?
- Problem set 13: Empty problem set.