1ZC3 Linear Algebra – Winter 2021 – Page 1

Lecture 1 – Systems of Linear Equations (Section 1.1)

A system of linear equations can have a unique solution, no solution (inconsistent) or infinite solutions.
Augmented Matrix: A coefficient matrix with the right side tacked on, to represent a system.

Elementary Row Operations of a matrix:

a) Multiply a row by a nonzero constant, 𝑟𝑖 → 𝑘𝑟𝑖 .
b) Add to a given row some multiple of another row, 𝑟𝑖 → 𝑟𝑖 + 𝑘𝑟𝑗 .
c) Exchange two rows, 𝑟𝑖 ↔ 𝑟𝑗 .
Combining 1 and 2 (𝑟𝑖 = 𝑎𝑟𝑖 + 𝑏𝑟𝑗 ) is NOT an elementary row operation.

Lecture 2 – Gaussian Elimination (Section 1.2)

Row-Echelon Form of a matrix:
a) The first non-zero entry in each row is a leading 1.
b) Leading 1s occur further to the right for every lower row.
c) Zero rows are at the bottom.
Reduced Row-Echelon Form:
d) In addition, each leading 1 is the only nonzero entry in its column.

Gaussian Elimination: Use elementary row operations to find the row-echelon form. Eliminate entries one
column at a time from the left, below the diagonal entries. To solve the system, recreate the equations and
use back substitution.
Gauss-Jordan Elimination: Continue from Gaussian Elimination to find the reduced row-echelon form.
Eliminate entries one column at a time from the right, above the diagonal entries. Recreate the equations
and solve for leading variables. No back substitution is required.

Homogeneous Linear System: 𝐴𝑥⃗ = 0⃗⃗, always has a trivial solution, where (𝑥⃗ = 0
⃗⃗)
Theorem 1.2.2: A homogeneous systems with more unknowns than equations must have ∞ solutions.

Lecture 3 – Matrix Operations (Section 1.3)

Matrix: A rectangle array of numbers (𝑚 rows by 𝑛 columns, or 𝑛 × 𝑛 if square)
Addition: If 𝐴 and 𝐵 are the same size, then 𝐴 + 𝐵 adds the matrices together element-wise.
Scaling: If 𝑐 is a scalar, then 𝑐𝐴 multiplies each entry in 𝐴 by 𝐶.

Matrix Multiplication: If 𝐴 is an 𝑚 × 𝑟 matrix and 𝐵 is an 𝑟 × 𝑛 matrix, then 𝐶 = 𝐴𝐵 is the 𝑚 × 𝑛 matrix

whose (𝑖, 𝑗)𝑡ℎ entry is the dot product 𝑐𝑖𝑗 = (𝑟𝑜𝑤 𝑖 𝑜𝑓 𝐴) ∙ (𝑐𝑜𝑙𝑢𝑚𝑛 𝑗 𝑜𝑓 𝐵).

The product 𝐶 = 𝐴𝐵 might not exist, when the inner size 𝑟 does not match.
𝐴𝐵 ≠ 𝐵𝐴 in general. When they do, they are said to commute.
A linear system of equations can be written in matrix notation as 𝐴x⃗⃗ = ⃗⃗
b.

Transpose: denoted by 𝐴𝑇 , is the 𝑛 × 𝑚 matrix by interchanging the rows and columns of 𝐴.

Associative Rule: (𝐴𝐵)𝐶 = 𝐴(𝐵𝐶) is true of matrices when the product exists.
Trace: denoted by 𝑡𝑟(𝐴), is the sum of the diagonal entries of an 𝑛 × 𝑛 matrix. Note that 𝑡𝑟(𝐴𝐵) = 𝑡𝑟(𝐵𝐴) and
𝑡𝑟(𝐴 + 𝐵) = 𝑡𝑟(𝐴) + 𝑡𝑟(𝐵)

Lecture 4, 5 – Inverses and Properties (Section 1.4)

Matrix Properties: All properties of numbers except 𝐴𝐵 ≠ 𝐵𝐴, except the following
a) 𝐴𝐵 = 𝐴𝐶 does not imply 𝐵 = 𝐶
b) 𝐴𝐵 = ⃗0⃗ does not imply 𝐴 = ⃗0⃗ or 𝐵 = ⃗0⃗.
 1ZC3 Linear Algebra – Winter 2021 – Page 2

Identity Matrix: An 𝑛 × 𝑛 matrix with 1 on the diagonal and 0 elsewhere. It has the property that 𝐴𝐼 = 𝐴 and
𝐼𝐴 = 𝐴 whenever the products are defined.
Inverse Matrix: If 𝐴 is a square matrix, then 𝐵 = 𝐴−1 is the inverse matrix of 𝐴 if 𝐴𝐵 = 𝐵𝐴 = 𝐼. If such 𝐵
exists, then the matrix 𝐴 is called invertible or non-singular.
Theorem 1.4.4: If 𝐵 and 𝐶 are both inverses of 𝐴, then 𝐵 = 𝐶
𝑎 𝑏 1 𝑑 −𝑏
Theorem 1.4.5: Let 𝐴 = [ ]. If and only if 𝑎𝑑 − 𝑏𝑐 ≠ 0, then 𝐴 is invertible and 𝐴−1 = 𝑎𝑑−𝑏𝑐 [ ].
𝑐 𝑑 −𝑐 𝑎

Solving a system: If 𝐴𝑥⃗ = 𝑏⃗⃗ and 𝐴 is invertible, then 𝑥⃗ = 𝐴−1 𝑏⃗⃗.

Theorem 1.4.6: If 𝐴 and 𝐵 are 𝑛 × 𝑛 and invertible, then the 𝐴𝐵 is also invertible and (𝐴𝐵)−1 = 𝐵−1 𝐴−1 .

Theorem 1.4.7: If 𝐴 is invertible, then,

a) 𝐴−1 is invertible and (𝐴−1 )−1 = 𝐴
b) If 𝑛 is a non-negative integer, then 𝐴𝑛 is invertible and (𝐴𝑛 )−1 = (𝐴−1 )𝑛
1
c) If 𝑘 is a nonzero scalar, then (𝑘𝐴)−1 = 𝐴−1 .
𝑘
Theorem 1.4.8: Properties of transposes
a) (𝐴𝑇 )𝑇 = 𝐴
b) (𝐴 + 𝐵)𝑇 = 𝐴𝑇 + B T
c) (𝐴 − 𝐵)𝑇 = 𝐴𝑇 − B T
d) (𝑘𝐴)𝑇 = 𝑘AT
e) (AB)T = B T AT
−1
Theorem 1.4.9: If A is invertible, then (AT ) = (A−1 )T

Lecture 5, 6 – Elementary Matrices (Section 1.5)

Elementary Matrix: denoted by 𝐸, is a matrix obtained from 𝐼 using an elementary row operation.
Theorem 1.5.1: 𝐸𝐴 does the same operation on 𝐴 as on 𝐼.
Theorem 1.5.2: Every elementary matrix is invertible, and E −1 is also elementary.

Theorem 1.5.3: (TFAE) If 𝐴 is an 𝑛 × 𝑛 matrix, then the following are equivalent:

a) 𝐴 is invertible.
b) 𝐴𝑥⃗ = ⃗0⃗ has only the trivial solution (𝑥⃗ = ⃗0⃗).
c) The reduced row-echelon form of 𝐴 is 𝐼.
d) 𝐴 can be written as a product of elementary matrices.

Inversion Algorithm: the inverse can be found by row reducing [𝐴 𝐼 ] into [𝐼 𝐴−1 ].

Lecture 7 – More on Linear Systems (Section 1.6)

Theorem 1.6.2: If 𝐴 is 𝑛 × 𝑛 and invertible, then 𝑥⃗ = 𝐴−1 𝑏⃗⃗ is the only solution to 𝐴𝑥⃗ = 𝑏⃗⃗ for every 𝑛 × 1 𝑏⃗⃗.
Theorem 1.6.3: to do.

Theorem 1.6.4: Additions to TFAE

e) 𝐴𝑥⃗ = 𝑏⃗⃗ is consistent for every 𝑏⃗⃗.
f) 𝐴𝑥⃗ = 𝑏⃗⃗ has exactly one solution for every 𝑏⃗⃗.

Theorem 1.6.5: If 𝐴𝐵 is invertible, and 𝐴 and 𝐵 are square, then 𝐴 and 𝐵 are invertible.

Lecture 8 – Diagonal, Triangular, and Symmetric Matrices (Section 1.7)

Diagonal Matrix: Ab 𝑛 × 𝑛 matrix with zeroes on off-diagonal entries. If all diagonal entries in a diagonal matrix
1
𝐷 is nonzero, then 𝐷 is invertible, and (𝐷 −1 )𝑖𝑖 = . If 𝑘 is a positive integer, then (𝐷 𝑘 )𝑖𝑖 = (𝐷𝑖𝑖 )𝑘 .
𝑑𝑖𝑖
Upper Triangular Matrix: A matrix with all zeroes below the diagonal.
 1ZC3 Linear Algebra – Winter 2021 – Page 3

Lower Triangular Matrix: A matrix with all zeroes above the diagonal.
Triangular: A matrix that is either upper or lower triangular.

Theorem 1.7.1:
a) If 𝐴 is upper triangular then 𝐴𝑇 is lower triangular. If 𝐴 is lower triangular then 𝐴𝑇 is upper triangular.
b) If 𝐴 and 𝐵 are lower triangular than 𝐴𝐵 is lower triangular. If 𝐴 and 𝐵 are upper triangular than 𝐴𝐵 is
upper triangular.
c) If 𝐴 is triangular and all diagonal entries are nonzero, then 𝐴 is invertible.
d) Suppose 𝐴 is invertible. If 𝐴 is upper triangular, then 𝐴−1 is upper triangular. If 𝐴 is lower triangular, then
𝐴−1 is lower triangular.

Symmetric Matrix: A square matrix 𝐴 that satisfies 𝐴 = 𝐴𝑇 , where entries are symmetric across the diagonal.
Theorem 1.7.2: If 𝐴 and 𝐵 are 𝑛 × 𝑛 symmetric matrices, then:
a) 𝐴𝑇 and 𝐵𝑇 are symmetric.
b) 𝐴 + 𝐵 and 𝐴 − 𝐵 are symmetric.
c) If 𝑘 is a scalar, then 𝑘𝐴 and 𝑘𝐵 are symmetric.

Theorem 1.7.4: If 𝐴 is invertible and symmetric, then 𝐴−1 is also symmetric.

Skew Symmetric: A square matrix 𝐴 that satisfies 𝐴𝑇 = −𝐴.

Lecture 9 – Determinants (Section 2.1)

𝑎 𝑏 𝑎 𝑏
2 by 2 Determinant: If 𝐴 = [ ], then det(𝐴) = | | = 𝑎𝑑 − 𝑏𝑐.
𝑐 𝑑 𝑐 𝑑

Minor: If 𝐴 is a square matrix, then the minor of 𝑎𝑖𝑗 , denoted by 𝑀𝑖𝑗 , is the determinant of the matrix obtained
from 𝐴 by deleting row 𝑖 and column 𝑗.
Cofactor: denoted by 𝑐𝑖𝑗 , is equal to 𝑐𝑖𝑗 = (−1)𝑖+𝑗 𝑀𝑖𝑗 . It changes sign in every alternating row and column.

Determinant: If 𝐴 is a 𝑛 × 𝑛 matrix then det(𝐴) = 𝑎11 𝑐11 + 𝑎12 𝑐12 + ⋯ + 𝑎1𝑛 𝑐1𝑛 is called the cofactor
expansion of 𝐴 along row 1. The determinant can be obtained using cofactor expansion along any row or
column, with the row/column with the most 0s being the easiest.

Theorem 2.1.2: If 𝐴 is triangular then det(𝐴) is the product of the diagonal entries of 𝐴.

Lecture 10 – Determinants by Row Reduction (Section 2.2)

Theorem 2.2.1: If 𝐴 is a 𝑛 × 𝑛 matrix and has a row or column of zeroes, then det(𝐴) = 0.
Theorem 2.2.2: If 𝐴 is a 𝑛 × 𝑛 matrix then det(𝐴) = det(𝐴𝑇 ).

Theorem 2.2.3: Let 𝐴 be a 𝑛 × 𝑛 matrix:

a) A nonzero scalar can be factored out of any row or column of a determinant
b) If 𝐵 is obtained from 𝐴 by a row/column exchange, then det(𝐵) = − det(𝐴).
c) If 𝐵 is obtained from 𝐴 by adding to a row/column some multiple of another, then det(𝐵) = det(𝐴).

Theorem 2.2.4: If 𝐸 is an elementary matrix then det(𝐸) ≠ 0.

Lecture 11, 12 – Properties of Determinants (Section 2.3)

Lemma 2.3.2: If 𝐸 and 𝐵 are 𝑛 × 𝑛 matrices, and 𝐸 is elementary, then det(𝐸𝐵) = det(𝐸) det (𝐵).

Theorem 2.3.3: (TFAE) 𝐴 is invertible if and only if det(𝐴) ≠ 0.

Theorem 2.3.4: If 𝐴 and 𝐵 are 𝑛 × 𝑛 matrices, then det(𝐴𝐵) = det(𝐴) det (𝐵). This also implies that
det(𝐴𝐵) = det(A)det(B) = det(𝐵𝐴), since order can be changed for scalar multiplication.
 1ZC3 Linear Algebra – Winter 2021 – Page 4
1
Theorem 2.3.5: If 𝐴 is invertible then det(𝐴−1 ) = .
det(𝐴)

Adjoint: denoted by adj(𝐴), is the transpose of the matrix of cofactors.

1
Theorem 2.3.6: If A is invertible then 𝐴−1 = det(𝐴) adj(𝐴)

Lecture 12, 13 –Eigenvalues and Eigenvectors (Section 5.1)

Eigenvectors and Eigenvalues: Let 𝐴 be an 𝑛 × 𝑛 matrix. 𝑥⃗ is called an eigenvector of 𝐴 with eigenvalue of 𝜆 if
𝑥⃗ ≠ 0⃗⃗, and 𝐴𝑥⃗ = 𝜆𝑥⃗. Notice that 𝐴𝑥⃗ = 𝜆𝑥⃗ = 𝜆𝐼𝑥⃗, which can be rearranged to (𝜆𝐼 − 𝐴)𝑥⃗ = 0
⃗⃗, where 𝑥⃗ ≠
⃗0⃗. By TFAE, 𝜆𝐼 − 𝐴 has to be singular and not invertible, and therefore det(𝜆𝐼 − 𝐴) = 0. To solve for
eigenvalues, solve det(𝜆𝐼 − 𝐴) = 0 for lambda. To find all the eigenvectors, find nonzero solutions to
(𝜆𝐼 − 𝐴)𝑥⃗ = ⃗0⃗ for each eigenvalue found in the previous step.

Linear Combination: Multiplying vectors by constants and adding the up: 𝑥⃗ = 𝑐1 𝑣 ⃗⃗⃗⃗⃗1 + 𝑐2 ⃗⃗⃗⃗⃗
𝑣2 + ⋯ + 𝑐𝑛 ⃗⃗⃗⃗⃗.
𝑣𝑛
Independence: Let 𝑆 = {𝑣1 , 𝑣2 , … , 𝑣𝑛 }. 𝑆 is independent if none of the vectors in 𝑆 can be written as a linear
combination of other vectors in 𝑆.
Eigenspace: 𝑆 is a basis for the eigenspace for an eigenvalue 𝜆 if 𝑆 is independent and every eigenvector 𝑥⃗ can be
written as a linear combination of vectors in 𝑆.

Theorem 5.1.2: If 𝐴 is triangular then the eigenvalues are the diagonal entries of 𝐴.
Theorem 5.1.4: (TFAE) 𝐴 is invertible if and only if 𝜆 = 0 is not an eigenvalue of 𝐴.

Lecture 14, 15 – Diagonalization (Section 5.2)

Similarity: If 𝐴 and 𝐵 are 𝑛 × 𝑛 matrices, then they are similar if there is an invertible 𝑛 × 𝑛 matrix 𝑃, such that
𝐵 = 𝑃−1 𝐴𝑃. If they are indeed similar, then the following are the same for both 𝐴 and 𝐵:
a) Determinant
b) Trace
c) Characteristic Polynomial det(𝜆𝐼 − 𝐴)
d) Eigenvalues
e) Invertibility

Diagonalization: An 𝑛 × 𝑛 matrix 𝐴 is diagonalizable if there exists an invertible matrix 𝑃 and a diagonal matrix
𝐷, such that 𝑃−1 𝐴𝑃 = 𝐷 (in other words, 𝐴 is similar to 𝐷). To find such a 𝑃 and 𝐷, let 𝑃 =
[𝑝1 𝑝2 𝑝3 … ], where 𝑝𝑞 is a basis vector for each eigenspace. Let 𝐷 be a diagonal matrix with all the
eigenvalues 𝜆1 , 𝜆2 , 𝜆3 , … (same order as 𝑝𝑞 ) as the diagonal entries. Then 𝑃−1 𝐴𝑃 = 𝐷 if there are
enough independent eigenvectors to fill the matrix 𝑃.

Algebraic Multiplicity: of eigenvalue 𝜆0 , is the power of (𝜆 − 𝜆0 ) in the characteristic polynomial.

Geometric Multiplicity: The number of vectors in a basis for the eigenspace corresponding to 𝜆0 .

Theorem 5.2.4: Let 𝐴 be a square matrix, then,

a) For every eigenvalue 𝜆 of 𝐴, the G.M. is less than or equal to the A.M.
b) 𝐴 is diagonalizable if and only if det(𝜆𝐼 − 𝐴) can be expressed as a product of linear factors, and for
every eigenvalue 𝜆 G.M. is equal to A.M.

Powers of a Matrix: If 𝐴 is diagonalizable and 𝑘 is a positive integer, then 𝐴𝑘 = 𝑃𝐷 𝑘 𝑃−1 (since all middle 𝑃𝑃−1
terms cancel out), which is much easier to compute because (𝐷 𝑘 )𝑖𝑖 = (𝐷𝑖𝑖 )𝑘 .

Lecture 16,17 – Differential Equations (Section 5.4)

Differential Equation: An equation involving a function and its derivatives (ex. 𝑦 ′′ + 3𝑦 ′ − 2𝑦 = 0). A solution
is a function that satisfies such an equation.
Natural Exponentials: The general solution to 𝑦 ′ = 𝑎𝑦 is 𝑦 = 𝑐𝑒 𝑎𝑥 , where 𝑎 and 𝑐 are constants.
 1ZC3 Linear Algebra – Winter 2021 – Page 5

First-Order Linear Systems: A constant first-order homogeneous linear system can be expressed as 𝑦⃗ ′ = 𝐴𝑦⃗,
where 𝐴 is a matrix of constants. If 𝐴 is diagonal, then each individual equation satisfies 𝑦 ′ = 𝑎𝑦 and can
be solved using natural exponentials.
Solution by Diagonalization: Suppose that 𝑦⃗ ′ = 𝐴𝑦⃗, where 𝐴 = 𝑃𝐷𝑃−1 is diagonalizable. Let 𝑦⃗ = 𝑃𝑢 ⃗⃗, 𝑦⃗ ′ =
′ ′
𝑃𝑢⃗⃗ , where 𝑢
⃗⃗ is an unknown vector. Substitute into first equation to get 𝑃𝑢 ⃗⃗ = 𝐴𝑃𝑢⃗⃗. Multiply both sides
by 𝑃−1 to get 𝑢 ⃗⃗′ = 𝑃 −1 𝐴𝑃𝑢
⃗⃗. But 𝑃−1 𝐴𝑃 = 𝐷, so 𝑢 ⃗⃗′ = 𝐷𝑢⃗⃗. This system is now solvable because the
coefficients are diagonal. When 𝑢 ⃗⃗ is solved, 𝑦⃗ can be obtained by the assumption 𝑦⃗ = 𝑃𝑢 ⃗⃗.

Theorem 5.4.1 (restatement of above): Suppose that 𝐴 is diagonalizable with eigenvectors 𝑥⃗1 , 𝑥⃗2 , … , 𝑥⃗𝑛 and
corresponding eigenvalues 𝜆1 , 𝜆2 , … , 𝜆𝑛 , then the general solution to the system of differential equations
𝑦⃗ ′ = 𝐴𝑦⃗ is 𝑦⃗ = 𝑐1 𝑥⃗1 𝑒 𝜆1 𝑥 + 𝑐2 𝑥⃗2 𝑒 𝜆2 𝑥 + ⋯ + 𝑐𝑛 𝑥⃗𝑛 𝑒 𝜆𝑛 𝑥 .

Higher-Order ODE: Introduce new functions and create a linear system.

Lecture 18 – Complex Numbers (Section 10.1, 10.2)

Complex Number: A number in the form 𝑧 = 𝑎 + 𝑏𝑖, where 𝑎, 𝑏 are real numbers and 𝑖 = √−1 (and 𝑖 2 = −1).
𝑧 can be represented as a vector in the complex plane, where 𝑦 is the imaginary and 𝑥 is the real axis.
Modulus: If 𝑧 = 𝑎 + 𝑏𝑖, then the modulus of 𝑧 is |𝑧| = √𝑎2 + 𝑏 2 , |𝑧| ≥ 0.
Conjugate: If 𝑧 = 𝑎 + 𝑏𝑖, then the conjugate of 𝑧 is 𝑧̅ = 𝑎 − 𝑏𝑖.

Complex Addition: Like vectors, (𝑎 + 𝑏𝑖) + (𝑐 + 𝑑𝑖) = (𝑎 + 𝑐) + (𝑏 + 𝑑)𝑖.

Complex Multiplication: Expanding, (𝑎 + 𝑏𝑖)(𝑐 + 𝑑𝑖) = 𝑎𝑐 + 𝑎𝑑𝑖 + 𝑏𝑐𝑖 + 𝑏𝑑𝑖 2 = (𝑎𝑐 − 𝑏𝑑) + (𝑎𝑑 + 𝑏𝑐)𝑖.
𝑎+𝑏𝑖 𝑎+𝑏𝑖 𝑎−𝑏𝑖
Complex Division: 𝑐+𝑑𝑖 = 𝑐+𝑑𝑖 𝑐−𝑑𝑖 (multiply top and bottom by the conjugate of the denominator).

Lecture 19 – Polar Form of a Complex Number (Section 10.3)

Polar Form: 𝑧 = 𝑎 + 𝑏𝑖 can instead be specified by 𝑟 and 𝜃, in the form of 𝑧 = 𝑟(cos 𝜃 + 𝑖 sin 𝜃). 𝜃 represents
the angle from the positive real axis and repeats every 2𝜋 radians. If −𝜋 < 𝜃 ≤ 𝜋, then 𝜃 is called the
principal argument of 𝑧.

Exponential Notation: 𝑒 𝑖𝜃 means the same thing as cos 𝜃 + 𝑖 sin 𝜃. It has the property that if 𝑧1 = 𝑟1 𝑒 𝑖𝜃1 and
𝑧2 = 𝑟2 𝑒 𝑖𝜃2 , then 𝑧1 𝑧2 = 𝑟1 𝑟2 𝑒 𝑖(𝜃1 +𝜃2 ) , an alternate way to multiply complex numbers.
DeMoivre’s Theorem: If 𝑧 = 𝑟𝑒 𝑖𝜃 , then 𝑧 𝑛 = 𝑟 𝑛 𝑒 𝑖𝑛𝜃 .

Lecture 21 – Vectors (Section 3.1, 3.2)

Vector Addition: 𝑤 ⃗⃗⃗ is a linear combination of the vectors 𝑣 ⃗⃗⃗⃗⃗,
1 ⃗⃗⃗⃗⃗,
𝑣2 … , 𝑣
⃗⃗⃗⃗⃗
𝑘 if there exists scalars 𝑐1 , 𝑐2 , … , 𝑐𝑘 such
that 𝑤
⃗⃗⃗ = 𝑐1 ⃗⃗⃗⃗⃗
𝑣1 + 𝑐2 ⃗⃗⃗⃗⃗
𝑣2 + ⋯ + 𝑐𝑘 ⃗⃗⃗⃗⃗.
𝑣𝑘
Norm: If 𝑣⃗ = (𝑣1 , 𝑣2 , … , 𝑣𝑛 ) is in ℝ𝑛 then the norm/length/magnitude of 𝑣⃗ is ‖𝑣⃗‖ = √𝑣12 + 𝑣22 + ⋯ + 𝑣𝑛2 . It is
always positive and is nonzero unless 𝑣⃗ = 0 ⃗⃗. It also has the properties that ‖𝑘𝑣⃗‖ = 𝑘‖𝑣⃗‖, and 1 𝑣⃗
‖𝑣
⃗⃗‖
produces a unit vector in the direction of 𝑣⃗.
Distance: The distance between two vectors 𝑢 ⃗⃗ and 𝑣⃗ is 𝑑(𝑢 ⃗⃗ − 𝑣⃗‖. It is √(𝑥2 − 𝑥1 )2 + (𝑦2 − 𝑦1 )2 in ℝ2 .
⃗⃗, 𝑣⃗) = ‖𝑢

Dot Product: The dot product of two vectors 𝑢 ⃗⃗ and 𝑣⃗ is 𝑢

⃗⃗ ∙ 𝑣⃗ = 𝑢1 𝑣1 + 𝑢2 𝑣2 + ⋯ + 𝑢𝑛 𝑣𝑛 . It has the properties
that 𝑣⃗ ∙ 𝑣⃗ = ‖𝑣⃗‖2, 𝑢
⃗⃗ ∙ 𝑣⃗ = 𝑣⃗ ∙ 𝑢 ⃗⃗ ∙ (𝑣⃗ + 𝑤
⃗⃗, and 𝑢 ⃗⃗⃗) = 𝑢 ⃗⃗ ∙ 𝑣⃗ + 𝑢
⃗⃗ ∙ 𝑤
⃗⃗⃗.

⃗⃗ and 𝑣⃗ are in ℝ𝑛 then |𝑢

Theorem 3.2.4: (The Cauchy-Shwartz Inequality) If 𝑢 ⃗⃗ ∙ 𝑣⃗| ≤ ‖𝑢
⃗⃗‖‖𝑣⃗‖.

⃗⃗∙𝑣
𝑢 ⃗⃗
⃗⃗ ∙ 𝑣⃗ = ‖𝑢
Angle Formula: cos 𝜃 = ‖𝑢⃗⃗‖‖𝑣⃗⃗‖, or 𝑢 ⃗⃗‖‖𝑣⃗‖ cos 𝜃, where 𝜃 is the angle between 𝑢
⃗⃗ and 𝑣⃗.
⃗⃗ and 𝑣⃗ are in ℝ𝑛 then ‖𝑢
Theorem 3.2.5: (The Triangle Inequality) If 𝑢 ⃗⃗ + 𝑣⃗‖ ≤ ‖𝑢
⃗⃗‖ + ‖𝑣⃗‖.
 1ZC3 Linear Algebra – Winter 2021 – Page 6

Lecture 22 – Orthogonality, Geometry (Section 3.3, 3.4)

Orthogonality: The vectors 𝑢 ⃗⃗ and 𝑣⃗ are orthogonal/perpendicular if 𝑢
⃗⃗ ∙ 𝑣⃗ = 0
Normal Vectors: If 𝑛⃗⃗ = (𝑎, 𝑏) is perpendicular to a line then the equation of the line is 𝑎𝑥 + 𝑏𝑦 + 𝑐 = 0, with
some constant 𝑐. If 𝑛⃗⃗ = (𝑎, 𝑏, 𝑐) is perpendicular to a two-dimensional plane then the equation of the
plane is 𝑎𝑥 + 𝑏𝑦 + 𝑐𝑧 + 𝑑 = 0, with some constant 𝑑.

Projections: Given two vectors 𝑎⃗ and 𝑢 ⃗⃗, we want to find a new vector 𝑢 ⃗⃗⃗⃗⃗1 = Proj𝑎⃗⃗ 𝑢
⃗⃗, called the the projection of
𝑢
⃗⃗ along 𝑎⃗, such that 𝑢
⃗⃗⃗⃗⃗1 is parallel to 𝑎⃗ and 𝑢
⃗⃗ − 𝑢
⃗⃗⃗⃗⃗1 is perpendicular to 𝑎⃗. In other words, we want to find a
⃗⃗∙𝑎⃗⃗
𝑢 ⃗⃗∙𝑎⃗⃗
𝑢
⃗⃗⃗⃗⃗1 = 𝑡𝑎⃗ and (𝑢
constant 𝑡, such that 𝑢 ⃗⃗ − 𝑡𝑎⃗) ∙ 𝑎⃗ = 0. Rearranging, we get 𝑡 = ‖𝑎⃗⃗‖2 , and 𝑢
⃗⃗⃗⃗⃗1 = ‖𝑎⃗⃗‖2 𝑎⃗. Note
that 𝑢
⃗⃗ − ⃗⃗⃗⃗⃗
𝑢1 is called the component of 𝑢
⃗⃗ orthogonal to 𝑎⃗. Also, if only the magnitude of the projection is
𝑢⃗⃗∙𝑎⃗⃗
needed, the expression can be simplified to ‖Proj𝑎⃗⃗ 𝑢
⃗⃗‖ = ‖𝑎⃗⃗‖

Theorem 3.4.1: Let 𝐿 be the line in ℝ2 or ℝ3 that contains the point ⃗⃗⃗⃗⃗
𝑥0 and is parallel to the nonzero vector 𝑣⃗.
Then, the equation of the line 𝐿 is 𝑥⃗ = ⃗⃗⃗⃗⃗
𝑥0 + 𝑡𝑣⃗, where the components are parametric equations.
Theorem 3.4.2: Similarly, for ℝ3 , a plane can be expressed as 𝑥⃗ = ⃗⃗⃗⃗⃗
𝑥0 + 𝑡1 ⃗⃗⃗⃗⃗
𝑣1 + 𝑡2 ⃗⃗⃗⃗⃗.
𝑣2
Theorem 3.4.3: If 𝐴 is an 𝑚 × 𝑛 matrix then the solution set of the homogeneous linear system 𝐴𝑥⃗ = ⃗0⃗ consist of
all vectors in ℝ𝑛 that are orthogonal to every row vector of 𝐴.

Lecture 23 – Cross Product (Section 3.5)

Cross Product: If 𝑢 ⃗⃗ = (𝑢1 , 𝑢2 , 𝑢3 ) and 𝑣⃗ = (𝑣1 , 𝑣2 , 𝑣3 ) are in ℝ3 then the cross product 𝑢
⃗⃗ × 𝑣⃗ is defined as
𝑖̂ 𝑗̂ ̂
𝑘 𝑖̂ = (1, 0, 0)
⃗⃗ × 𝑣⃗ = |𝑢1 𝑢2 𝑢3 |, where 𝑗̂ = (0, 1, 0) . The result is a vector that is orthogonal to both 𝑢
𝑢 ⃗⃗ and 𝑣⃗.
𝑣1 𝑣2 𝑣3 ̂
𝑘 = (0, 0, 1)
Properties of the Cross Product:
a) 𝑢⃗⃗ × 𝑣⃗ = −(𝑣⃗ × 𝑢 ⃗⃗)
b) 𝑢⃗⃗ × (𝑣⃗ + 𝑤⃗⃗⃗) = 𝑢 ⃗⃗ × 𝑣⃗ + 𝑢⃗⃗ × 𝑤
⃗⃗⃗
c) 𝑢⃗⃗ × 𝑢⃗⃗ = 𝑢 ⃗⃗
⃗⃗ × 0 = 0 ⃗⃗
d) 𝑘(𝑢 ⃗⃗ × 𝑣⃗) = 𝑘𝑢 ⃗⃗ × 𝑣⃗ = 𝑢⃗⃗ × 𝑘𝑣⃗

⃗⃗ × 𝑣⃗‖2 = ‖𝑢
Lagrange’s Identity: ‖𝑢 ⃗⃗ ∙ 𝑣⃗)2
⃗⃗‖‖𝑣⃗‖ − (𝑢

⃗⃗ and 𝑣⃗ are in ℝ3 then ‖𝑢

Theorem 3.5.3: If 𝑢 ⃗⃗ × 𝑣⃗‖ is the area of the parallelogram determined by 𝑢
⃗⃗ and 𝑣⃗
𝑢1 𝑢2 𝑢3
Scalar Triple Product: 𝑢 ⃗⃗⃗) = | 𝑣1 𝑣2 𝑣3 |, is the volume of the parallelopiped formed by 𝑢
⃗⃗ ∙ (𝑣⃗ × 𝑤 ⃗⃗, 𝑣⃗, and 𝑤
⃗⃗⃗.
𝑤1 𝑤2 𝑤3

Lecture 24 – Real Vector Spaces (Section 4.1)

Vector Space: Let 𝑉 be a nonempty set of objects together with two operations called “addition” and “scalar
multiplication”. 𝑉 is called a vector space, and its elements called vectors in 𝑉, if and only if the
following ten axioms are satisfied:
1) Closure under addition: If 𝑢 ⃗⃗ and 𝑣⃗ are in 𝑉, then the sum 𝑢 ⃗⃗ + 𝑣⃗ is also in 𝑉.
2) 𝑢⃗⃗ + 𝑣⃗ = 𝑣⃗ + 𝑢 ⃗⃗ for all 𝑢 ⃗⃗, 𝑣⃗ in 𝑉.
3) 𝑢⃗⃗ + (𝑣⃗ + 𝑤 ⃗⃗⃗) = (𝑢 ⃗⃗ + 𝑣⃗) + 𝑤 ⃗⃗⃗ for all 𝑢
⃗⃗, 𝑣⃗, 𝑤
⃗⃗⃗ in 𝑉.
4) There is an object in 𝑉 called the zero vector, denoted by ⃗0⃗, such that 𝑢 ⃗⃗ + ⃗0⃗ = 𝑢⃗⃗ for all 𝑢
⃗⃗ in 𝑉.
5) For each 𝑢 ⃗⃗ in 𝑉 there is an object in 𝑉 called the negative of 𝑢 ⃗⃗, denoted by −𝑢 ⃗⃗, such that 𝑢 ⃗⃗ + (−𝑢 ⃗⃗.
⃗⃗) = 0
6) Closure under scalar multiplication: If 𝑢 ⃗⃗ is in 𝑉 and 𝑘 is a scalar, then the product 𝑘𝑢 ⃗⃗ is also in 𝑉.
7) 𝑘(𝑢 ⃗⃗ + 𝑣⃗) = 𝑘𝑢 ⃗⃗ + 𝑘𝑣⃗.
8) (𝑘 + 𝑚)𝑢 ⃗⃗ = 𝑘𝑢 ⃗⃗ + 𝑚𝑢 ⃗⃗.
9) 𝑘(𝑚𝑢 ⃗⃗) = (𝑘𝑚)𝑢 ⃗⃗
10) 1𝑢⃗⃗ = 𝑢 ⃗⃗
 1ZC3 Linear Algebra – Winter 2021 – Page 7

Theorem 4.1.1: If 𝑉 is a vector space then it also has the following properties:
a) −𝑢 ⃗⃗ = (−1)𝑢 ⃗⃗
b) 0𝑢⃗⃗ = 0 ⃗⃗
c) 𝑘0⃗ = ⃗0⃗
⃗
d) If 𝑘𝑢 ⃗⃗ = 0⃗⃗ then either 𝑘 = 0 or 𝑢 ⃗⃗
⃗⃗ = 0

Lecture 25 – Subspaces (Section 4.2)

Subspace: A subset 𝑊 of a vector space 𝑉 is a subspace of 𝑉 if 𝑊 itself is a vector space using the same addition
and scalar multiplication operations defined on 𝑉.
Theorem 4.2.1: If 𝑊 is a nonempty subset of a vector space 𝑉, then 𝑊 is a subspace if and only if the following
conditions hold for 𝑊 (no other axioms need to be checked):
a) Closure under addition
b) Closure under scalar multiplication

Lecture 26 – Span (Section 4.3)

Span: Let 𝑆 = {𝑤
⃗⃗⃗⃗⃗,
1 𝑤⃗⃗⃗⃗⃗⃗,
2 … , ⃗⃗⃗⃗⃗}
𝑤𝑟 be a nonempty set of vectors in a vector space 𝑉. The set of all possible linear
combinations of vectors in 𝑆 is called the span of 𝑆, denoted by span(𝑆). In set notation, the span is
written as {𝑐1 ⃗⃗⃗⃗⃗
𝑤1 + 𝑐2 ⃗⃗⃗⃗⃗⃗ 𝑤𝑟 | 𝑐1 , 𝑐2 , … , 𝑐𝑟 ∈ ℝ}. If 𝑉 = span(𝑆), then the vectors in 𝑆 are said
𝑤2 + ⋯ + 𝑐𝑟 ⃗⃗⃗⃗⃗
to span 𝑉, and every vector in 𝑉 can be written as a linear combination of the vectors in 𝑆.

Theorem 4.3.1: Let 𝑆 = {𝑤

⃗⃗⃗⃗⃗,
1 𝑤⃗⃗⃗⃗⃗⃗,
2 … , ⃗⃗⃗⃗⃗}
𝑤𝑟 be a nonempty set of vectors in a vector space 𝑉. Then 𝑊 = span(𝑆) is
a subspace of 𝑉.

Lecture 27 – Linear Independence (Section 4.4)

Theorem 4.4.1: Let 𝑆 = {𝑣 ⃗⃗⃗⃗⃗,
1 ⃗⃗⃗⃗⃗,
𝑣2 … , ⃗⃗⃗⃗}
𝑣𝑟 be a set of vectors in the vector space 𝑉. 𝑆 is independent if and only if
the equation 𝑘1 ⃗⃗⃗⃗⃗
𝑣1 + 𝑘2 ⃗⃗⃗⃗⃗ 𝑣𝑟 = ⃗0⃗ has only the trivial solution (ie. 𝑘1 = 𝑘2 = ⋯ = 𝑘𝑟 = 0).
𝑣2 + ⋯ + 𝑘𝑟 ⃗⃗⃗⃗

Theorem 4.4.3: Let 𝑆 = {𝑣

⃗⃗⃗⃗⃗,
1 ⃗⃗⃗⃗⃗, 𝑣𝑟 be a set of vectors in ℝ𝑛 . If 𝑟 > 𝑛 then 𝑆 is linearly dependent.
𝑣2 … , ⃗⃗⃗⃗}

Lecture 28,29 – Coordinates and Basis (Section 4.5)

Basis: Let 𝑆 = {𝑣
⃗⃗⃗⃗⃗,
1 ⃗⃗⃗⃗⃗,
𝑣2 … , ⃗⃗⃗⃗}
𝑣𝑟 be a set of vectors in the vector space 𝑉. 𝑆 is called a basis for 𝑉 if and only if 𝑆 is
independent and 𝑆 spans 𝑉.

Theorem 4.5.1: If 𝑆 = {𝑣 ⃗⃗⃗⃗⃗,

1 ⃗⃗⃗⃗⃗,
𝑣2 … , ⃗⃗⃗⃗}
𝑣𝑟 is a basis for a vector space 𝑉 then every 𝑣⃗ ∈ 𝑉 can be written as the
combination 𝑣⃗ = 𝑐1 ⃗⃗⃗⃗⃗ 𝑣1 + 𝑐2 ⃗⃗⃗⃗⃗
𝑣2 + ⋯ + 𝑐𝑟 ⃗⃗⃗⃗
𝑣𝑟 in exactly one way.
Coordinates: Given 𝑣⃗ = 𝑐1 ⃗⃗⃗⃗⃗ 𝑣1 + 𝑐2 ⃗⃗⃗⃗⃗ 𝑣𝑟 in a vector space 𝑉 with the basis 𝑆 = {𝑣
𝑣2 + ⋯ + 𝑐𝑟 ⃗⃗⃗⃗ ⃗⃗⃗⃗⃗,
1 ⃗⃗⃗⃗⃗,
𝑣2 … , ⃗⃗⃗⃗},
𝑣𝑟
𝑐1 , 𝑐2 , … , 𝑐𝑛 are called the coordinates of 𝑣⃗ relative to 𝑆, and the vector (𝑐1 , 𝑐2 , … , 𝑐𝑛 ) is denoted (𝑣)𝑆 .

Lecture 31 – Gram Shmidt (Section 6.3)

Orthogonal Set: A set of vectors 𝑆 = {𝑣 ⃗⃗⃗⃗⃗,
1 ⃗⃗⃗⃗⃗, 𝑣𝑟 in ℝ𝑛 is orthogonal if and only if 𝑣𝑖 ⋅ 𝑣𝑗 = 0 for all 𝑖 ≠ 𝑗
𝑣2 … , ⃗⃗⃗⃗}
less than 𝑟. If in addition, each vector in 𝑆 has length 1, then 𝑆 is called orthonormal.

Theorem 6.3.1: Every set 𝑆 of nonzero orthogonal vectors is linearly independent.

Theorem 6.3.2: Let 𝑆 = {𝑣
⃗⃗⃗⃗⃗,
1 ⃗⃗⃗⃗⃗, 𝑣𝑟 be a basis for a subspace 𝑊 of ℝ𝑛 and let 𝑢
𝑣2 … , ⃗⃗⃗⃗} ⃗⃗ be any vector in 𝑊.
⃗⃗∙𝑣
𝑢 ⃗⃗⃗⃗⃗ ⃗⃗∙𝑣
𝑢 ⃗⃗⃗⃗⃗ ⃗⃗∙𝑣
𝑢 ⃗⃗⃗⃗⃗
⃗⃗ = ‖𝑣⃗⃗⃗⃗⃗‖12 𝑣
a) If 𝑆 is orthogonal then 𝑢 ⃗⃗⃗⃗⃗1 + ‖𝑣⃗⃗⃗⃗⃗‖22 ⃗⃗⃗⃗⃗
𝑣2 + ⋯ + ‖𝑣⃗⃗⃗⃗⃗‖𝑟2 ⃗⃗⃗⃗.
𝑣𝑟
1 2 𝑟
⃗⃗ = (𝑢
b) If 𝑆 is orthonormal then 𝑢 ⃗⃗ ⋅ 𝑣
⃗⃗⃗⃗⃗)𝑣 1 + (𝑢
1 ⃗⃗⃗⃗⃗ 𝑣2 ⃗⃗⃗⃗⃗2 + ⋯ + (𝑢
⃗⃗ ⋅ ⃗⃗⃗⃗⃗)𝑣 ⃗⃗ ⋅ ⃗⃗⃗⃗)𝑣
𝑣𝑟 ⃗⃗⃗⃗𝑟

Theorem 6.3.3: If a set 𝑆 of 𝑛 vectors in ℝ𝑛 spans ℝ𝑛 or is independent, then 𝑆 is a basis for ℝ𝑛 .

Orthogonal Projections: Let 𝑊 be a subspace of ℝ𝑛 and let 𝑆 = {𝑣

⃗⃗⃗⃗⃗,
1 ⃗⃗⃗⃗⃗,
𝑣2 … , ⃗⃗⃗⃗}
𝑣𝑟 be an orthogonal basis for for 𝑊.
⃗⃗∙𝑣
𝑢 ⃗⃗⃗⃗⃗ ⃗⃗∙𝑣
𝑢 ⃗⃗⃗⃗⃗ ⃗⃗∙𝑣
𝑢 ⃗⃗⃗⃗⃗
⃗⃗ ∈ ℝ𝑛 then the projection of 𝑢
If 𝑢 ⃗⃗ onto 𝑊 is ‖𝑣⃗⃗⃗⃗⃗‖12 𝑣
⃗⃗⃗⃗⃗1 + ‖𝑣⃗⃗⃗⃗⃗‖22 ⃗⃗⃗⃗⃗
𝑣2 + ⋯ + ‖𝑣⃗⃗⃗⃗⃗‖𝑟2 ⃗⃗⃗⃗,
𝑣𝑟 with these properties:
1 2 𝑟
 1ZC3 Linear Algebra – Winter 2021 – Page 8

a) Proj𝑊 𝑢 ⃗⃗ is in 𝑊.
b) 𝑢
⃗⃗⃗⃗⃗1 = 𝑢
⃗⃗ − Proj𝑊 𝑢 ⃗⃗ is orthogonal to every vector in 𝑆.
c) 𝑢
⃗⃗⃗⃗⃗1 = 𝑢
⃗⃗ − Proj𝑊 𝑢 ⃗⃗ is orthogonal to every vector in 𝑊.
d) Proj𝑊 𝑢 ⃗⃗ is independent of the choice of orthogonal basis.
Gram Shmidt Process: Let 𝑆 = {𝑢 ⃗⃗⃗⃗⃗,
1 𝑢⃗⃗⃗⃗⃗, 𝑢𝑟 be a basis for a subspace 𝑊 of ℝ𝑛 . Then to find an orthogonal
2 … , ⃗⃗⃗⃗⃗}
⃗⃗⃗⃗⃗⃗∙𝑣
𝑢 ⃗⃗⃗⃗⃗ ⃗⃗⃗⃗⃗⃗∙𝑣
𝑢 ⃗⃗⃗⃗⃗ ⃗⃗⃗⃗⃗⃗∙𝑣
𝑢 ⃗⃗⃗⃗⃗
basis, 𝑆⊥ = {𝑣 ⃗⃗⃗⃗⃗,
1 ⃗⃗⃗⃗⃗,
𝑣2 … , ⃗⃗⃗⃗},
𝑣𝑟 let 𝑣 ⃗⃗⃗⃗⃗1 = ⃗⃗⃗⃗⃗,
𝑢1 ⃗⃗⃗⃗⃗ ⃗⃗⃗⃗⃗2 − 2 12 𝑣
𝑣2 = 𝑢 ⃗⃗⃗⃗⃗,
1 ⃗⃗⃗⃗⃗ ⃗⃗⃗⃗⃗3 − 3 12 𝑣
𝑣2 = 𝑢
‖𝑣⃗⃗⃗⃗⃗‖
⃗⃗⃗⃗⃗1 − 3 22 ⃗⃗⃗⃗⃗
‖𝑣
⃗⃗⃗⃗⃗‖ ‖𝑣
𝑣2 through all 𝑟
⃗⃗⃗⃗⃗‖
1 1 2
⃗⃗⃗⃗⃗⃗∙𝑣
𝑢 𝑟 ⃗⃗⃗⃗⃗
1 ⃗⃗⃗⃗⃗⃗∙𝑣
𝑢 𝑟 ⃗⃗⃗⃗⃗
2 ⃗⃗⃗⃗⃗⃗∙𝑣
𝑢𝑟 ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗
𝑟−1
until 𝑣𝑟 = 𝑢𝑟 − ‖𝑣 ⃗⃗⃗⃗⃗‖ 2𝑣⃗⃗⃗⃗⃗1 − ‖𝑣 ⃗⃗⃗⃗⃗‖ 2 ⃗⃗⃗⃗⃗
𝑣2 − ⋯ − ‖𝑣⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗‖2 𝑣
⃗⃗⃗⃗⃗⃗⃗⃗⃗,
𝑟−1 which forms that basis.
1 2 𝑟−1

Lecture 32 – Dimension (Section 4.6)

Dimension: The dimension of a vector space 𝑉 is the number of vectors in a basis for 𝑉.
Theorem 4.6.4: Let 𝑉 be a vector space with dimension 𝑛 and let 𝑆 be a set of vectors in 𝑉 with 𝑛 vectors. Then 𝑆
is a basis for 𝑉 if 𝑆 spans 𝑉 or 𝑆 is independent.

Lecture 33, 34 – Row Space, Column Space, and Null Space (Section 4.8)
Matrix Subspaces: Let 𝐴 be an 𝑚 × 𝑛 matrix.
a) The subspace of ℝ𝑛 spanned by the rows of 𝐴 is the row space of 𝐴.
b) The subspace of ℝ𝑛 spanned by the columns of 𝐴 is the column space of 𝐴.
c) The subspace of ℝ𝑛 consisting of all solutions to the equation 𝐴𝑥⃗ = ⃗0⃗ is the null space of 𝐴.

Theorem 4.8.5: 𝐴𝑥⃗ = 𝑏⃗⃗ is consistent if and only if 𝑏⃗⃗ is in the column space of 𝐴.
Theorem 4.8.3: Elementary row operations do not change the row space and null space of 𝐴. The do however
change the column space of 𝐴.
Theorem 4.8.4: Let 𝑅 be a row-echelon form of 𝐴.
a) The nonzero vectors of 𝑅 form a basis for the row space of 𝐴.
b) The columns of 𝐴 corresponding to those of 𝑅 with leading 1s form a basis for the column space of 𝐴.

Lecture 35,36 – Cryptography (Section 10.13)

Modular Arithmatic: If 𝑚 is a positive integer and 𝑎 and 𝑏 are any integers then 𝑎 is equivalent to 𝑏 modulo 𝑚,
denoted by 𝑎 ≡ 𝑏 (mod 𝑚), if 𝑎 − 𝑏 is an integer multiple of 𝑚.
Result 13.1: Every integer 𝑎 is equivalent modulo 𝑚 to exactly one of ℤ𝑚 = {0,1,2, … , 𝑚 − 1}.

Residue: If 𝑎 ≡ 𝑏 (mod 𝑚) and 𝑏 ∈ ℤ𝑚 then 𝑏 is the residue of 𝑎 modulo 𝑚.

Hill-2 Cipher:
𝑎 𝑏
1) Choose a 2 × 2 enciphering matrix 𝐴 = [ ].
𝑐 𝑑
2) Group letters into pairs and replace each by its numerical value (from 0 to 25)
𝑝1
3) For each pair 𝑝⃗ = [𝑝 ], let 𝑐⃗ = 𝐴𝑝⃗ be the ciphertext vector.
2
4) Replace each entry in 𝑐⃗ by its residue modulo 26.
5) Convert pairs back into plain text.

Reciprocals: If 𝑎 ∈ ℤ𝑚 then there is a number 𝑎−1 ∈ ℤ𝑚 such that 𝑎𝑎−1 ≡ 1 (mod 𝑚) called the reciprocal or
the multiplicative inverse of 𝑎 modulo 𝑚.
Result 13.2: If 𝑎 ∈ ℤ𝑚 then 𝑎 has a reciprocal modulo 𝑚 if and only if 𝑎 and 𝑚 have no common prime factors.
𝑎 𝑏
Result 13.3: Let 𝐴 = [ ] have entries in ℤ𝑚 . If the residue of 𝑎𝑑 − 𝑏𝑐 has a reciprocal modulo 𝑚 then 𝐴 has
𝑐 𝑑
𝑑 −𝑏
an inverse modulo 𝑚, and 𝐴−1 = (𝑎𝑑 − 𝑏𝑐)−1 [ ].
−𝑐 𝑎

Deciphering: Use the same algorithm as the cipher but replace 𝐴 with 𝐴−1.

And that’s the end of *these* notes.