Linear Algebra I

Matrices

Year 1 · Semester IILast updated

May 21 Lecture. Links to the board and the recording.

Notation:

A single capital letter denotes a matrix. (A,B,C,S,T…A, B, C, S, T\dots)

Am×n=[a11a12a13…a1na21…⋮⋱am1]A_{m\times n}=\begin{bmatrix} a_{11}&a_{12}&a_{13}&\dots&a_{1n}\\ a_{21}&\dots\\ \vdots&\ddots\\ a_{m1} \end{bmatrix}

mm rows by nn columns
aij=a_{ij}= entry in ith row and jth column (1≤i≤m,1≤j≤n1≤i≤m, 1≤j≤n)

Transpose operation

if AA is an m×nm \times n matrix, then ATA^T is the n×mn\times m matrix defined by AT=[bij]A^T=[b_{ij}] with bij=ajib_{ij}=a_{ji} for 1≤i≤m,1≤j≤n1≤ i ≤ m, 1≤ j≤n
Example:

A=[13−20−14]AT=[103−1−24]A=\begin{bmatrix}1&3&-2\\0&-1&4\end{bmatrix}\quad A^T=\begin{bmatrix}1&0\\3&-1\\-2&4\end{bmatrix}

A symmetric matrix AA is one with the property A=ATA=A^T.

Symmetric Matrices - In English

This means that the matrix is perfectly reflected across the main diagonal.
This can only ever happen with a square matrix

Matrix multiplication

General Notation in this section

Matrix AA refers to the left operand in multiplication.
Matrix BB refers to the right operand in multiplication.
Matrix CC refers to the product of A×BA\times B

The matrices A and B can be multiplied whenever the number of columns in A is the same number of row in B
Example:

A=[13−20−14]B=[2−1030121−310−2]A=\begin{bmatrix}1&3&-2\\0&-1&4\end{bmatrix}\qquad B=\begin{bmatrix}2&-1&0&3\\0&1&2&1\\-3&1&0&-2\end{bmatrix}

The CijC_{ij} is dot product of row ii of AA with col jj of BB

C=[82410−12−12−9]C=\begin{bmatrix}8&2&4&10\\-12&-1&2&-9\end{bmatrix}

General definition: If A is a m×nm\times n matrix and B is a n×pn\times p matrix, then C=A⋅BC=A\cdot B is a m×pm\times p matrix defined by Cij=(ith row of A)⋅(jth row of B)C_{ij}=(i^{th}\text{ row of }A)\cdot (j^{th}\text{ row of }B)

Cij=(ith row of A)⋅(jth column of B)=∑k=1naikbkjC_{ij}=(\text{i}^{th}\text{ row of }A)\cdot(\text{j}^{th}\text{ column of }B)=\sum_{k=1}^na_{ik}b_{kj}
Matrix Multiplication - In English (and TypeScript)

To multiply a matrix, you must ensure that the dimensions work out. The columns of AA must equal the rows of BB. This is required so we can take the dot product, as the operands of a dot product operation must be of equal dimension.
The dimension of the product will inversely be the rows of AA by the columns of BB.
For the resulting matrix CC, the value in CijC_{ij} is the dot product of Arow iA_{\text{row }i} and Bcol jB_{\text{col }j}

The code for matrix multiplication in TypeScript is as follows:

function matrixMultiplication(A: Matrix, B: Matrix){
	let result = new Matrix({
		rowCount: A.columns.length,
		columnCount: B.rows.length
	})
	for (let i = 0; i < A.rows.length; i++) {
		for (let j = 0; j < B.columns.length; j++) {
			result[i][i] = dotProduct(A.rows[i], B.columns[j]);
		}
	}
	return result;
}
Properties of Multiplication:
  • Not commutative (A⋅B≠B⋅AA\cdot B\not=B\cdot A)
  • Is Associative ((A⋅B)⋅C=A⋅(B⋅C)(A\cdot B)\cdot C=A\cdot (B\cdot C))
  • For any matrix A,BA,B and any scalar c∈Rc\in\R:
    • (c⋅A)⋅B=A⋅(c⋅B)=c⋅(A⋅B)(c\cdot A)\cdot B=A\cdot (c\cdot B)=c\cdot (A\cdot B)
  • If row ii is all 0 in A, then the ithi^{th} row of (A⋅BA\cdot B) is all zeros
  • If col jj is all 0 in BB, then the jthj^{th} col of (A⋅BA\cdot B) is all zeros
  • (A⋅B)T=BT⋅AT{(A\cdot B)}^T=B^T\cdot A^T
  • For any matrix AA, A⋅ATA\cdot A^T is a symmetric matrix
  • (A⋅AT)T=(AT)T⋅AT=(A⋅AT){(A\cdot A^T)}^T={(A^T)}^T\cdot A^T=(A\cdot A^T)
Tips and Tricks

If row ii of AA is all zeros, the corresponding row in CC will be zeros too.
If column jj of BB is all zeros, the corresponding column in CC will be zeros too.
Note, col’s of 0s in AA or rows of 0s in BB does not tell us anything.

Systems of Equations

A system of equations 3x+4x−2z=1−1x+1y+1z=32x−1y+3z=−5\begin{aligned}3x+4x-2z&=1\\-1x+1y+1z&=3\\2x-1y+3z&=-5\end{aligned} can be written as follows:

[342−1112−13][xyz]=[13−5]\begin{bmatrix}3&4&2\\-1&1&1\\2&-1&3\end{bmatrix}\begin{bmatrix}x\\y\\z\end{bmatrix}=\begin{bmatrix}1\\3\\-5\end{bmatrix}

General notation for system: A⋅x⃗=b⃗A\cdot\vec x=\vec b
where AA is the restricted coefficient matrix,
x⃗\vec x is the vector of unknowns (so spooky 👻),
and b⃗\vec b is the vector of free coefficients.


May 28 Lecture. Links to the board and the recording.

Square Matrices (n×nn\times n)

A square matrix is the only type which can be multiplied by itself (A⋅A=A2A\cdot A=A^2)

A=[1−132]A2=[1−132]×[1−132]=[−2−391]A=\begin{bmatrix} 1&-1\\ 3&2 \end{bmatrix} \qquad A^2= \begin{bmatrix} 1&-1\\ 3&2 \end{bmatrix} \times \begin{bmatrix} 1&-1\\ 3&2 \end{bmatrix} = \begin{bmatrix} -2&-3\\ 9&1 \end{bmatrix}
Some Important Square Matrices:
Upper Triangular

All nonzero entries are on or above the main diagonal.
Formal Definition:

If i>ji>j then Aij=0A_{ij}=0

Lower Triangular

All nonzero entries are on or below the main diagonal.
Formal Definition:

If i<ji<j then Aij=0A_{ij}=0

Diagonal Matrix

All nonzero entries are on the main diagonal.
Formal Definition:

If i≠ji\not=j then Aij=0A_{ij}=0

Neat facts

The product of Upper Triangles is an Upper Triangle.
The product of Lower Triangles is a Lower Triangle.
The product of Diagonal Matrices is a Diagonal Matrix

Identity Matrix

Definition:
InI_n is the n×nn\times n diagonal matrix in which every entry on the main diagonal is 11.
I3I_3=[100011001]\begin{bmatrix}1&0&0\\0&1&1\\0&0&1\end{bmatrix}
If AA is any m×nm\times n matrix, then Im⋅A=A⋅In=AI_m\cdot A=A\cdot I_n=A
Specifically for square matrix AA, where II is the same order as AA:
I⋅A=A⋅I=AI\cdot A=A\cdot I=A

Scalar matrix:

Definition:
A scalar matrix is the n×nn\times n diagonal matrix in which every entry on the main diagonal is c∣∀c∈Rc\mid\forall c\in \R
Example:
c⋅I3=[c000c000c]c\cdot I_3=\begin{bmatrix}c&0&0\\0&c&0\\0&0&c\end{bmatrix}
Claim:
If AA is any m×nm\times n matrix, then:
cIm⋅A=cAcI_m\cdot A=cA
A⋅cIn=cAA\cdot cI_n=cA

Example
[200020002][−1301−12314]=[−2602−24628]\begin{bmatrix} 2&0&0\\ 0&2&0\\ 0&0&2\\ \end{bmatrix} \begin{bmatrix} -1&3&0\\ 1&-1&2\\ 3&1&4\\ \end{bmatrix}= \begin{bmatrix} -2&6&0\\ 2&-2&4\\ 6&2&8\\ \end{bmatrix}

Inverse Matrices

Definition:
If AA is an n×nn\times n matrix, and AB=IAB=I, then AA is called invertable, and BB is the inverse matrix of AA, denoted B=A−1B=A^{-1}.

Inverse Matrices — In English

Every scalar number has its inverse. The inverse of 22 is 121\over2. The inverse of 14\frac14 is 44. The inverse of 85\frac85 is 58\frac58.
The common ground is that multiplying a scalar by its inverse gives you the “Identity Scalar” - that is, 11.

With matrices, the same concept applies. The inverse of a matrix AA is the matrix you multiply AA with to get the Identity Matrix - that is, II.

Example

[2153][3−1−52]=[1001]=I\begin{bmatrix}2&1\\5&3\\\end{bmatrix}\begin{bmatrix}3&-1\\-5&2\\\end{bmatrix}=\begin{bmatrix}1&0\\0&1\\\end{bmatrix}=I

If A=[2153]A=\begin{bmatrix}2&1\\5&3\\\end{bmatrix}, then A−1=[3−1−52]A^{-1}=\begin{bmatrix}3&-1\\-5&2\\\end{bmatrix}.

Not every matrix is invertible.

Any matrix AA with a row or column of zeros is not invertible, because ABAB will always have a row of zeros, hence AB≠IAB\not=I.
This is not the only thing that will cause AA to be non-invertible.

Properties

Let AA be an invertible matrix.

  1. A−1A^{-1} is also invertible, and (A−1)−1=A{(A^{-1})}^{-1}=A.
  2. A⋅A−1=A−1⋅A=IA\cdot A^{-1}=A^{-1}\cdot A=I (They commute).
  3. A−1A^{-1} is unique. There are no two different matrices that can invert AA.
  4. If A,BA,B are invertible, then ABAB is invertible, and (AB)−1=B−1⋅A−1{(AB)}^{-1}=B^{-1}\cdot A^{-1}.
    • This applies to nn invertible matrices:
      (A1⋅A2⋅⋯⋅An)−1=An−1⋅⋯⋅A2−1⋅A1−1{(A_1\cdot A_2\cdot \dots\cdot A_n)}^{-1}=A_n^{-1}\cdot\dots\cdot A_2^{-1}\cdot A_1^{-1}
    • The inverse of products is the product of the inverses, in reverse order.
      zoom
Using Inverse Matrices to Solve Systems of Equations

Example:
{2x+y=15x+3y=−1  ⟹  [2153][xy]=[1−1](A⋅x⃗=b⃗)\begin{cases}2x+y&=1\\5x+3y&=-1\end{cases}\implies\begin{bmatrix}2&1\\5&3\end{bmatrix}\begin{bmatrix}x\\y\end{bmatrix}=\begin{bmatrix}1\\-1\end{bmatrix}\qquad(A\cdot\vec x=\vec b)
Note that AA is invertible, see above example, A−1=[3−1−52]A^{-1}=\begin{bmatrix}3&-1\\-5&2\\\end{bmatrix}
Since we want to isolate x⃗\vec x, let’s cancel out AA. We do this by multiplying both sides by A−1A^{-1} on the left like so:

A−1(Ax⃗)=A−1b⃗  ⟹  [3−1−52][2153][xy]=[3−1−52]⋅[1−1]\textcolor{darkorange}{A^{-1}}(A\vec x)=\textcolor{darkorange}{A^{-1}}\vec b\implies\textcolor{darkorange}{\begin{bmatrix}3&-1\\-5&2\\\end{bmatrix}}\begin{bmatrix}2&1\\5&3\end{bmatrix}\begin{bmatrix}x\\y\end{bmatrix}=\textcolor{darkorange}{\begin{bmatrix}3&-1\\-5&2\\\end{bmatrix}}\cdot\begin{bmatrix}1\\-1\end{bmatrix}

(A−1A)x⃗=A−1b⃗  ⟹  I⋅[xy]=[3−1−52]⋅[1−1](A^{-1}A)\vec x=A^{-1}\vec b\implies I\cdot\begin{bmatrix}x\\y\end{bmatrix}=\begin{bmatrix}3&-1\\-5&2\\\end{bmatrix}\cdot\begin{bmatrix}1\\-1\end{bmatrix}
x⃗=A−1b⃗  ⟹  [xy]=[4−7]  ⟹  {x=4y=−7\vec x=A^{-1}\vec b\implies\begin{bmatrix}x\\y\end{bmatrix}=\begin{bmatrix}4\\-7\end{bmatrix}\implies\begin{cases}x&=4\\y&=-7\end{cases}

Big Questions:
  1. How can we tell if AA is invertible?
  2. If it is, how do we find A−1A^{-1}?
    Stay tuned for the answers…

Elementary Matrices

Definition:
An elementary matrix is a matrix obtained from II by a single elementary row operation.
Notation:
EijE_{ij} - The matrix obtained from II by Ri↔RjR_i\leftrightarrow R_j
Ei(c)E_i(c) - The matrix obtained from II by Ri→cRiR_i\to cR_i
Eij(c)E_{ij}(c) - The matrix obtained from II by Ri→Ri+cRjR_i\to R_i+cR_j
Examples:
E12=[010100001]E_{12}=\begin{bmatrix}0&1&0\\1&0&0\\0&0&1\end{bmatrix} (The matrix obtained from II by R1↔R2R_1\leftrightarrow R_2)
E2(3)=[100030001]E_2(3)=\begin{bmatrix}1&0&0\\0&3&0\\0&0&1\end{bmatrix} (The matrix obtained from II by R2→3R2R_2\to 3R_2)
E21(1)=[100110001]E_{21}(1)=\begin{bmatrix}1&0&0\\1&1&0\\0&0&1\end{bmatrix} (The matrix obtained from II by R2→R2+1R1R_2\to R_2+1R_1)


June 4 Lecture. Links to the board and the recording.

The Big Claims™:
  1. Let AA be any n×nn\times n matrix, and let EE be any n×nn\times n elementary matrix. The matrix EAEA is exactly the matrix obtained from AA by performing the single row operation through which II becomes EE.
  2. Every elementary matrix is invertible.
    • Eij−1=Eij{E_{ij}}^{-1}=E_{ij} (self-inverse)
    • Ei(c)−1=Ei(1c){E_i(c)}^{-1}=E_i(\frac1c)
    • Eij(c)−1=Eij(−c){E_{ij}(c)}^{-1}=E_{ij}(-c)

Determining if AA is invertible

  1. Use the standard row operations to reduce AA as much as possible, keeping note of the operations and the order.
    1. If you are able to reach the Identity Matrix, it’s invertible, congrats!
      Read on to find the inverse. Otherwise, AA is not invertable.
  2. Rewrite this reduction as a sequence of multiplications on the left by elementary matrices.
    • Since you have [series of elementary matrices]⋅A=I[\text{series of elementary matrices}]\cdot A=I, AA is invertible.
    • A−1=[series of elementary matrices]A^{-1}=[\text{series of elementary matrices}]

Example:

A=[01110211 0  ]→E21R2↔R1[102011110]→E31(−1)R3→R3−R1[10201101−2]→E32(−1)R3→R3−R2[10201100−3]→E3(−13)R2↔R1[102011001]→E23(−1)R2→R2−R3[102010001]→E13(−2)R1→R1−2R3[10001000  1 ]\begin{aligned}A=&\begin{bmatrix}0 & 1 & 1 \\1 & 0 & 2 \\1 & 1 &\ 0 \ \ \\ \end{bmatrix} \xrightarrow[E_{21}]{R_2\leftrightarrow R_1} \begin{bmatrix}1 & 0 & 2 \\0 & 1 & 1 \\1 & 1 & 0 \\ \end{bmatrix} \xrightarrow[E_{31}(-1)]{R_3\rightarrow R_3-R_1} \begin{bmatrix}1 & 0 & 2 \\0 & 1 & 1 \\0 & 1 &-2 \\ \end{bmatrix} \xrightarrow[E_{32}(-1)]{R_3\rightarrow R_3-R_2}\\ &\begin{bmatrix}1 & 0 & 2 \\0 & 1 & 1 \\0 & 0 &-3 \\ \end{bmatrix} \xrightarrow[E_{3}(-\frac13)]{R_2\leftrightarrow R_1} \begin{bmatrix}1 & 0 & 2 \\0 & 1 & 1 \\0 & 0 & 1 \\ \end{bmatrix} \xrightarrow[E_{23}(-1)]{R_2\rightarrow R_2-R_3} \begin{bmatrix}1 & 0 & 2 \\0 & 1 & 0 \\0 & 0 & 1 \\ \end{bmatrix} \xrightarrow[E_{13}(-2)]{R_1\rightarrow R_1-2R_3}\\ &\begin{bmatrix}1 & 0 & 0 \\0 & 1 & 0 \\0 & 0 &\ \ 1 \ \\\end{bmatrix}\end{aligned} [10−2010001]E13(−2)⋅([10001−1001]E23(−1)⋅([10001000−13]E3(−13)⋅([1000100−11]E32(−1)⋅([100010−101]E31(−1)⋅([010100001]E12⋅A)))))=I\underset{E_{13}(-2)}{\begin{bmatrix}1 & 0 & -2 \\0 & 1 & 0 \\0 & 0 & 1 \\\end{bmatrix}}\cdot\left(\underset{E_{23}(-1)}{\begin{bmatrix}1 & 0 & 0 \\0 & 1 & -1 \\0 & 0 & 1 \\\end{bmatrix}}\cdot\left(\underset{E_{3}(-\frac13)}{\begin{bmatrix}1 & 0 & 0 \\0 & 1 & 0 \\0 & 0 & -\frac13 \\\end{bmatrix}}\cdot\left(\underset{E_{32}(-1)}{\begin{bmatrix}1 & 0 & 0 \\0 & 1 & 0 \\0 & -1 & 1 \\\end{bmatrix}}\cdot\left(\underset{E_{31}(-1)}{\begin{bmatrix}1 & 0 & 0 \\0 & 1 & 0 \\-1 & 0 & 1 \\\end{bmatrix}}\cdot\left(\underset{E_{12}}{\begin{bmatrix}0 & 1 & 0 \\1 & 0 & 0 \\0 & 0 & 1 \\\end{bmatrix}}\cdot A\right)\right)\right)\right)\right)=I

By associativity of matrix multiplication, [Product of EM’s]⋅A=I[\text{Product of EM's}]\cdot A=I

Therefore, AA is invertible, and A−1=[Product of EM’s]A^{-1}=[\text{Product of EM's}]

A−1A^{-1} as a product of Elementary Matrices

A−1A^{-1} is the product of the Elementary Matrices obtained by reducing AA to II, in the reverse of the order in which you obtained them.

This is frequently an exam question. Be prepared.

Finding A−1A^{-1}

To find the inverse of AA, we will make a double window matrix, like this: [AI]\left[\begin{array}{c|c}A&I\end{array}\right].
Perform each row operation on both windows.
At the end of the reduction, the left window will be II, and the right window will be A−1A^{-1}.

Example:

[011100102010110001]→E21R2↔R1[102010011100110001]→E31(−1)R3→R3−R1[10201001110001−20−11]→E32(−1)R3→R3−R2[10201001110000−3−1−11]→E3(−13)R3→−13R3[1020100111000011313−13]→E23(−1) & E13(−2)R2→R2−R3 & R1→R1−2R3[100−23132301023−13130011313−13][I∣A−1]  ⟹  A−1=13[−2122−1111−1]\begin{aligned} &\left[\begin{array}{ccc|ccc} 0&1&1&1&0&0\\ 1&0&2&0&1&0\\ 1&1&0&0&0&1 \end{array}\right] \xrightarrow[E_{21}]{R_2\leftrightarrow R_1} \left[\begin{array}{ccc|ccc} 1&0&2&0&1&0\\ 0&1&1&1&0&0\\ 1&1&0&0&0&1 \end{array}\right] \xrightarrow[E_{31}(-1)]{R_3\rightarrow R_3 - R_1} \left[\begin{array}{ccc|ccc} 1&0&2&0&1&0\\ 0&1&1&1&0&0\\ 0&1&-2&0&-1&1 \end{array}\right] \xrightarrow[E_{32}(-1)]{R_3\rightarrow R_3 - R_2}\\\\ &\left[\begin{array}{ccc|ccc} 1&0&2&0&1&0\\ 0&1&1&1&0&0\\ 0&0&-3&-1&-1&1 \end{array}\right] \xrightarrow[E_{3}(-\frac13)]{R_3\rightarrow -\frac13 R_3} \left[\begin{array}{ccc|ccc} 1&0&2&0&1&0\\ 0&1&1&1&0&0\\ 0&0&1&\frac13&\frac13&-\frac13 \end{array}\right] \xrightarrow[E_{23}(-1)\ \&\ E_{13}(-2)]{R_2\rightarrow R_2-R_3\ \&\ R_1\rightarrow R_1-2R_3}\\\\ &\underset{[I|A^{-1}]}{\left[\begin{array}{ccc|ccc} 1&0&0&-\frac23&\frac13&\frac23\\ 0&1&0&\frac23&-\frac13&\frac13\\ 0&0&1&\frac13&\frac13&-\frac13 \end{array}\right]} \implies A^{-1}=\frac13\begin{bmatrix} -2&1&2\\ 2&-1&1\\ 1&1&-1 \end{bmatrix} \end{aligned}
In summary: Matrix Inversion

To check if AA is invertible, set up a “double window” [AI]\left[\begin{array}{c|c}A&I\end{array}\right] and reduce AA, performing the same operations on both sides at every step.
If AA can’t be reduced to II, then AA is not invertible.
If AA can be reduced to II, then the right window will contain A−1A^{-1} at the end of the reduction.

Summary Theorem
Summary Theorem - This WILL be on the exam in some form

A n×nn\times n matrix AA which satisfies one of the below satisfies all of the below:

  1. AA is invertible
  2. AA can be written as a product of Elementary Matrices
  3. AA can be reduced to II
  4. The system Ax⃗=b⃗  ⟹  x⃗=A−1b⃗A\vec x=\vec b\implies\vec x=A^{-1}\vec b has a unique solution
  5. rank(A)=n\text{rank}(A)=n (See Rank)
  6. ∣A∣≠0|A|\not=0 (See Determinants)

June 8 Tirgul. Links to the board and the recording.


June 11 Lecture. Links to the board and the recording.

Row Space, Column Space, and Rank of a matrix

Let AA be a m×nm\times n matrix, in these examples, A=[10−23−114−11105]A=\begin{bmatrix}1&0&-2&3\\-1&1&4&-1\\1&1&0&5\end{bmatrix}

Row Space

Definition:
The row space of AA is Span(rows of A)≤Rn\text{Span}(\text{rows of } A)\leq\R^n
(note that ≤\leq denotes subspace)
Example:
The row space of AA is Span{(1,0,−2,3),(−1,1,4,−1),(1,1,0,5)}≤Rn\text{Span}\{(1,0,-2,3),(-1,1,4,-1),(1,1,0,5)\}\leq\R^n

Column Space

Definition:
The column space of AA is Span(columns of A)≤Rm\text{Span}(\text{columns of } A)\leq\R^m
Example:
The column space of AA is Span{(1,−1,1),(0,1,1),(−2,4,0),(3,−1,5)}≤Rm\text{Span}\{(1,-1,1),(0,1,1),(-2,4,0),(3,-1,5)\}\leq\R^m

Rank

Definition:
The row rank of AA is the dimension of the row space of AA.
The column rank of AA is the dimension of the column space of AA.
The row rank will always equal the column rank.
We therefore define rank as the row rank and the column rank, which are all equivalent.
Finding the rank of a matrix:

  1. Reduce AA
  2. rank AA = number of rows which don’t turn to all zeros after AA is reduced.
    Example:
    Given our example matrix above, the reduced form is [10−2301220000]\begin{bmatrix}1&0&-2&3\\0&1&2&2\\0&0&0&0\end{bmatrix}
    rank A=2A=2
    Claim:
    An n×nn\times n matrix AA is invertible iff rank A=nA=n.
    [Proof omitted, see page 5 of linked board.]

Determinants

Definition:
The determinant of a square matrix AA is a number computed from the entries of AA which has geometric and algebraic significance.
It is denoted: det A\text{det }A, or ∣A∣|A|, or ∣a11…a1n⋮an1…ann∣\left|\begin{matrix}a_{11}&\dots&a_{1n}\\\vdots\\a_{n1}&\dots&a_{nn}\end{matrix}\right|
The determinant of a 2×22\times 2 matrix ∣abcd∣\left|\begin{matrix}a&b\\c&d\end{matrix}\right| is ad−bcad-bc.
Examples:

  • ∣3714∣=5\left|\begin{matrix}3&7\\1&4\end{matrix}\right|=5
  • ∣1−3−412∣=0\left|\begin{matrix}1&-3\\-4&12\end{matrix}\right|=0
    • Note, since the determinant is zero, the matrix is not invertible.
    • Also note, R2=−3R1R_2=-3R_1
Computing ∣An∣|A_n| for n>2n > 2:

Ex. A=[112−1010−11]A=\begin{bmatrix}1&1&2\\-1&0&1\\0&-1&1\end{bmatrix}
Definition:

Definition of Matrix Minors

Let AA be an n×nn\times n matrix. The (i,j)(i, j)-minor of AA, denoted as MijM_{ij}, is the (n−1)×(n−1)(n-1)\times(n-1) matrix obtained by “deleting” the ithi^{th} row and the jthj^{th} column of AA.

Example of matrix minors

Using the above matrix AA:
M13=[−100−1](112−1010−11)M_{13}=\begin{bmatrix}-1&0\\0&-1\end{bmatrix} \qquad\begin{pmatrix}\color{red}1&\color{red}1&\color{red}\bf{2}\\-1&0&\color{red}1\\0&-1&\color{red}1\end{pmatrix}
M21=[12−11](112−1010−11)M_{21}=\begin{bmatrix}1&2\\-1&1\end{bmatrix}\qquad\begin{pmatrix}\color{red}1&1&2\\\color{red}\bf{-1}&\color{red}0&\color{red}1\\\color{red}0&-1&1\end{pmatrix}

Formula for ∣An∣|A_n|

For any i∈[1,n]i\in [1, n]

∑j=1n−1j+i⋅aij⋅∣Mij∣\sum_{j=1}^n-1^{j+i}\cdot a_{ij}\cdot|M_{ij}|
Example of ∣A3∣|A_3|

Using the above matrix AA:

A=[112−1010−11]∣A∣=+(+1)⏟a11⋅∣01−11∣⏟M11−(+1)⏟a12⋅∣01−11∣⏟M12+(+2)⏟a13⋅∣01−11∣⏟M13=1(1)−1(−1)+2(1)=4\begin{aligned}A&=\begin{bmatrix}1&1&2\\-1&0&1\\0&-1&1\end{bmatrix}\\|A|&=+\underbrace{(+1)}_{a_{11}}\cdot\underbrace{\begin{vmatrix}0&1\\-1&1\end{vmatrix}}_{M_{11}}-\underbrace{(+1)}_{a_{12}}\cdot \underbrace{\begin{vmatrix}0&1\\-1&1\end{vmatrix}}_{M_{12}}+\underbrace{(+2)}_{a_{13}}\cdot\underbrace{\begin{vmatrix}0&1\\-1&1\end{vmatrix}}_{M_{13}}\\&=1(1)-1(-1)+2(1)\\&=4\end{aligned}
Theorem: There is nothing special about the first row.

∣A∣|A| can be computed along any row or any column, using the formula:
Along row ii:

∣A∣=∑j=1n−1j+i⋅aij⋅∣Mij∣|A|=\sum_{j=1}^n-1^{j+i}\cdot a_{ij}\cdot|M_{ij}|

Along column jj:

∣A∣=∑i=1n−1j+i⋅aij⋅∣Mij∣|A|=\sum_{i=1}^n-1^{j+i}\cdot a_{ij}\cdot|M_{ij}|

It is therefore advantageous to take the determinant along a row or column with lots of zeros, to reduce the number of minor determinants you have to calculate.


June 15 Tirgul. Links to the board and the recording.

Properties of Determinants
  1. If AA has a row or column of zeros, then ∣A∣=0|A|=0
  2. ∣AT∣=∣A∣|A^T|=|A|
  3. If AA is a triangular matrix (upper, lower, or diagonal), then
    ∣A∣=|A|= the product of the entries on the main diagonal.
Important Proofs Omitted!

The proofs for these properties are likely important to know for the exam. Make sure you are familiar with these.


June 18 Lecture. Links to the board and the recording.