113B-EEDP10016
線性代數 Linear Algebra
T56R2, CY202
助教:
褚亭妏 59209 <
[email protected] >、
郭羿萱 59209 <
[email protected]>
教科書及參考資料:
Gilbert Strang,
Introduction to Linear Algebra. International Edition. Wellesley-Cambridge
Google/OCW: MIT 18.06
Supplement:
1. Stephen Boyd and Lieven Vandenberghe, Introduction to Linear Algebra. Vectors, Matrices
and Least Squares, Cambridge University Press
2. 3Bule1Brown, Khan Academy, Steve Brunton in UW
3. 交大周志成教授，線代啟示錄
4. 資料科學與 Python 程式實務 Numerical Linear Algebra
評分標準:
Midterm exam (4/22 in 10th week): 25%
Final exam (6/5 in 16th week): 25%
Five Quizzes (best 5 out of 6): 50% (Quiz Time: R1)
Office Hour:
After class (Tues7 Thur34) or appointment.
1
113B-EEDP10016
Schedule:
The following schedule is an approximation. Please use it to plan your reading, and preview
the related material before coming the class.
week date class Key ideas/outline
W1 2/18 L0. Single variable Calculus,
Introduction to Linear Algebra Multivariable calculus,
Vector Analysis
Linear Algebra
2/20 L1.
The Geometry of Linear equations Linear equations (n unknowns; n equations)
Linear equations Ax = b (2.1) Row/Column picture;
Square matrix form in linear equation
Solving Linear equations Ax = b
Idea of elimination, back substitution
Matrix-vector operations
L2. Elimination using Matrices
The Idea of Elimination (2.2)
Elimination using Matrices (2.3)
W2 2/25 L3. 4 ways to see the matrix multiplication
Rules for Matrix Operations (2.4) Inverse
-1
Inverse Matrices (2.5) Gauss-Jordan Elimination A
2/27 L4. Production of elimination matrices
Elimination and Factorization Matrix factorization (LU or LDU)
A = LU (2.6) Why so bothering? (Algorithm)
Cost of Elimination
W3 3/4 L5. Transpose: column to row
Transposes and Permutations (2.7) Permutation (order exchange)
3/6 L6. Vectors in ℝ𝑛𝑛 space;
Vector spaces and Subspaces (3.1) Vector, linear combination, vector space,
2
113B-EEDP10016
subspace;
Column space = range
W4 3/11 L7. Ax = 0
Null Space N(A): Solving Ax = 0 Pivot number (rank: r);
and Rx = 0 (3.2) Free variables (n-r)
Rank and the RREF(A) (3.3) rref(A) = R;
Rxn = 0 special solutions
A = CR
3/13 L8. Ax = b; x= xp+xn;
Complete Solution to Ax= b (3.4) xp: particular solution
xn: special solution in N(A)
Ax = b through rank analysis
W5 3/18 L9. Linear independence
Independence, Basis and Spanning a space;
Dimension (3.5) Basis for a vector space;
Dimension of vector space
C(A ) + N(A) in ℝ𝑛𝑛
T
L10.
C(A) + N(A ) in ℝ𝑚𝑚
T
The four fundamental subspaces
(3.6)
3/20 L11. Orthogonal vectors
Orthogonality of the four Orthogonal space
C(A ) ⊥ N(A)
T
subspaces (4.1)
C(A) ⊥ N(A )
T
W6 3/25 L12. Projections (4.2) Why projection:
Ax = b has no solution
Projection matrix
T
A A is invertible when A has independent
column vectors.
3/27 L13. A is an overdetermined system
Least square approximation (4.3) Ax = b (m>n)
W7 4/1 L14. Orthogonal bases q1, q2, …
Orthogonal Bases and Orthogonal matrix Q
Gram-Schmidt (4.4) Gram-Schmidt A → Q
Orthogonal matrix Q Factorization A = QR
4/3 Spring Break
3
113B-EEDP10016
W8 4/8 L15.
The properties of determinants 10 properties of determinants
(5.1)
Permutations, cofactors (5.2) Formula for det(A)
Big formula by permutations
Cofactors formula
Special Tridiagonal matrices
4/10 L16. Solving Ax = b by det(A) and Cramer’s rule
-1
Cramer’s rule, Inverses (5.3) Solving A by det(A) and cofactos
W9 4/15 L17. Ax = λx, find λ’s first, then soling the
Introduction to Eigenvalues (6.1) corresponding x’s
det(A-λI) = 0
Trace(A) = λ1+λ2+… ; Det(A) = λ1⋅λ2…
X AX = Λ
-1
4/17 L18.
K
Diagonalizing a Matrix (6.2) Power of A: generation problem (A )
X: Eigenvector matrices Fibonacci numbers
W10
4/22 Midterm Exam, Chap. 1-4
4/24
Review of midterm
W11 4.29 L19. T
S = QΛQ (Q AQ =Λ)
T
Real Symmetric matrices S (6.4) Real eigenvalues λ’s
Orthogonal eigenvectors Q
5/1 L20. Complex inner product
Complex vectors of Matrices (10) Real symmetric (S) vs Complex Hermitian
Hermitian matrices H (H)
Unitary matrices U Orthogonal (Q) vs Unitary (U)
W12 5/6 L21. Markov matrices;
Applications Discrete Fourier Transform
5/8 L22-1. Tests of positive matrix
T
Positive Definite Matrices (6.5) -I Minimum x Sx > 0
2 2
Elliptical equation ax +2bxy+cy = 1
T
S = A A symmetric and positive def
4
113B-EEDP10016
W13 5/13 L22-2 Why positive definite? Optimization!
Positive Definite Matrices (6.5)- II
-1
5/15 L23 Similar Matrices (6.6) B = M AM, for some invertible matrix M
Geometric Multiplicity (GM);
Algebraic Multiplicity (AM)
Jordan form
W14 5/20 L24-1. A = UΣV
T
Singular Value Decomposition U, V: orthogonal matrix;
(SVD) (6.7) Σ: diagonal matrix
5/22 L25-2. Why SVD: Dual Basis Analysis
Singular Value Decomposition
(SVD) (6.7)
W15 5/27 L24. Matrix of linear transformation
The matrix of linear Change of Basis
transformations (7.1, 7.2)
5/29 L25. Four fundamental spaces, SVD, basis change,
Diagonalization and Pseudoinverse to linear transformation
W16 6/3
Review of all semester
6/5
Final Exam, Chap. 5, 6, 7, 10
