各位西交利物浦的同学们大家好啊!期末在即,为了不挂科,我费了好大力气,尽自己所能的整理了这些提纲。内容粗浅,如果大家能看到,并且为大家的学习提供了一点点的帮助,那真的再好不过了。哦还有,我是按 MTH007 的课程进行整理的,由于课程不同,可能知识点有所欠缺,欢迎各位大佬在评论区补充,本人才疏学浅,若内容有误,欢迎大家指正。^_^
本博客仅供学习交流使用,侵删。
文章目录
- Chapter 1 : Linear Equations in Linear Algebra
- Chapter 2 : Matrix Algebra
- Chapter 3 : Determinants
- Chpater 4 : Vector Spaces
- Chapter 5 : Eigenvalues and Eigenvectors
Chapter 1 : Linear Equations in Linear Algebra
1.1 Systems of Linear Equation
1.1.1 Linear Equation
- A Linear equation in the variables x 1 , x 2 , ⋅ ⋅ ⋅ , x n x_1,x_2,···,x_n x1,x2,⋅⋅⋅,xn is an equation that can be written in the form a 1 x 1 + a 2 x 2 + ⋅ ⋅ ⋅ + a n x n = b a_1x_1+a_2x_2+···+a_nx_n=b a1x1+a2x2+⋅⋅⋅+anxn=b, where b b b and the coefficients a 1 , a 2 , ⋅ ⋅ ⋅ , a n a_1,a_2,···,a_n a1,a2,⋅⋅⋅,an are real or complex numbers.
- A system of linear equations is a collection of one or more linear equations involving the same variables.
- A solution of the system is a list ( s 1 , s 2 , ⋅ ⋅ ⋅ , s n ) (s_1,s_2,···,s_n) (s1,s2,⋅⋅⋅,sn) of numbers that makes each equation a true statement when the values s 1 , s 2 , ⋅ ⋅ ⋅ , s n s_1,s_2,···, s_n s1,s2,⋅⋅⋅,sn are substituted for x 1 , x 2 , ⋅ ⋅ ⋅ , x n x_1,x_2,···,x_n x1,x2,⋅⋅⋅,xn respectively.
- The set of all possible solutions is called the solution set of the linear system.
- Two linear systems are called equivalent if they have the same solution set.
A system of linear equations has :
- no solution
- exactly one solution
- infinitely many solutions
A system of linear equations is said to be consistent (相容的) if it has either one solution
or infinitely many solutions.
A system is inconsistent (不相容) if it has no solution.
1.1.2 Matrix
with the coefficients of each variable aligned in columns, the matrix is called the coefficient matrix (系数矩阵) of the system
An augmented matrix (增广矩阵) of a system consists of the coefficient matrix with an added column containing the constants from the right sides of the equations.
1.1.3 Elementary Row Operations
- Replacement (倍加变换) : Replace one row by the sum of itself and a multiple of another row.
- Interchange (对换变换) : Interchange two rows.
- Scaling (倍乘变换) : Multiply all entries in a row by a nonzero constant.
Two matrices are called row equivalent if there is a sequence of elementary row operations that transforms one matrix into the other.
1.2 Row Reduction and Echelon Forms
1.2.1 Echelon Form
A nonzero row or column in a matrix means a row or column that contains at least one nonzero entry, and a leading entry (先导元素) of a row refers to the leftmost nonzero entry in a nonzero row.
A rectangular matrix is in echelon form if it has the following three properties:
- All nonzero rows are above any rows of all zeros.
- Each leading entry of a row is in a column to the right of the leading entry of the
row above it.- All entries in a column below a leading entry are zeros.
If a matrix in echelon form satisfies the following additional conditions, then it is in reduced echelon form :
- The leading entry in each nonzero row is 1.
- Each leading 1 is the only nonzero entry in its column.
The reduced echelon form one obtains from a matrix is unique.
Theorem 1 : Uniqueness of the Reduced Echelon Form
Each matrix is row equivalent to one and only one reduced echelon matrix.
1.2.2 Pivot Position
A pivot position in a matrix A A A is a location in A A A that corresponds to a leading 1 in the reduced echelon form of A A A. A A A pivot column is a column of A A A that contains a pivot position.
1.2.3 Solutions of Linear Systems
The variables x 1 x_1 x1 and x 2 x_2 x2 corresponding to pivot columns in the matrix are called basic variables (基本变量). The other variable, x 3 x_3 x3, is called a free variable (自由变量).
Theorem 2 : Existence and Uniqueness Theorem
A linear system is consistent if and only if the rightmost column of the augmented matrix is not a pivot column. If a linear system is consistent, then the solution set contains either
(i) a unique solution, when there are no free variables, or
(ii) infinitely many solutions, when there is at least one free variable.
1.3 Vector Equations
A matrix with only one column is called a column vector, or simply a vector.
The set of all vectors with n entries is denoted by R 2 ℝ^2 R2
Definition
If v ⃗ 1 , ⋅ ⋅ ⋅ , v ⃗ p \vec{v}_1,···,\vec{v}_p v1,⋅⋅⋅,vp are in R n ℝ^n Rn , then the set of all linear combinations of v ⃗ 1 , ⋅ ⋅ ⋅ , v ⃗ p \vec{v}_1,···,\vec{v}_p v1,⋅⋅⋅,vp is denoted by Span { v ⃗ 1 , ⋅ ⋅ ⋅ , v ⃗ p } \{\vec{v}_1,···,\vec{v}_p\} {v1,⋅⋅⋅,vp} and is called the subset of R n ℝ^n Rn spanned by v ⃗ 1 , ⋅ ⋅ ⋅ , v ⃗ p \vec{v}_1,···,\vec{v}_p v1,⋅⋅⋅,vp.
1.4 Matrix Equation
If A A A is an m × n m\times n m×n matrix, with columns a 1 , ⋅ ⋅ ⋅ , a n a_1,···,a_n a1,⋅⋅⋅,an, and if x ⃗ \vec{x} x is in R n ℝ^n Rn, then the product of A A A and x ⃗ \vec{x} x, denoted by A x ⃗ A\vec{x} Ax, is the linear combination of the columns of A A A using the corresponding entries in x ⃗ \vec{x} x as weights.
A x ⃗ A\vec{x} Ax is defined only if the number of columns of A A A equals the number of entries in x ⃗ \vec{x} x.
Theorem 3
Let A A A be an m × n m\times n m×n matrix. Then the following statements are logically equivalent.
- For each b b b in R m ℝ^m Rm, the equation A x ⃗ = b A\vec{x}=b Ax=b has a solution.
- Each b b b in R m ℝ^m Rm is a linear combination of the columns of A.
- The columns of A span R m ℝ^m Rm
- A has a pivot position in every row.
The matrix with 1’s on the diagonal and 0’s elsewhere is called an identity matrix and is denoted by I I I
1.5 Homogeneous Linear Systems (齐次线性系统)
A system of linear equations is said to be homogeneous if it can be written in the form A x ⃗ = 0 A\vec{x}=0 Ax=0, where A A A is an matrix and 0 0 0 is the zero vector in R m ℝ^m Rm.
Such a system A x ⃗ = 0 A\vec{x}=0 Ax=0 always has at least one solution x ⃗ = 0 \vec{x}=0 x=0, this solution is usually called the trivial solution. The important question is whether there exists a nontrivial solution, that is, a nonzero vector x ⃗ \vec{x} x that satisfies A x ⃗ = 0 A\vec{x}=0 Ax=0.
The equation of the form x ⃗ = s u ⃗ + t v ⃗ \vec{x}=s\vec{u}+t\vec{v} x=su+tv is called a parametric vector equation (参数向量方程) of the plane. A solution set is described explicitly with vectors, we say that the solution is in parametric vector form.
Theorem 4
Suppose the equation A x ⃗ = b A\vec{x}=b Ax=b is consistent for some given b b b, and let p p p be a solution. Then the solution set of A x ⃗ = b A\vec{x}=b Ax=b is the set of all vectors of the form w ⃗ = p + v ⃗ h \vec{w}=p+\vec{v}_h w=p+vh, where v ⃗ h \vec{v}_h vh is any solution of the homogeneous equation.
1.7 Linear Independence
An indexed set of vectors { v ⃗ 1 , ⋅ ⋅ ⋅ , v ⃗ p } \{\vec{v}_1,···,\vec{v}_p\} {v1,⋅⋅⋅,vp} in R n ℝ^n Rn is said to be linearly independent if the vector equation has only the trivial solution. The set { v ⃗ 1 , ⋅ ⋅ ⋅ , v ⃗ p } \{\vec{v}_1,···,\vec{v}_p\} {v1,⋅⋅⋅,vp} is said to be linearly dependent if there exist the nontrivial solution.
Theorem 5 : Characterization of Linearly Dependent Sets
An indexed set S = { v ⃗ 1 , ⋯ , v ⃗ p } S=\{\vec{v}_1,⋯,\vec{v}_p\} S={v1,⋯,vp} of two or more vectors is linearly dependent if and only if at least one of the vectors in S is a linear combination of the others.
Theorem 6
Any set { v ⃗ 1 , ⋅ ⋅ ⋅ , v ⃗ p } \{\vec{v}_1,···,\vec{v}_p\} {v1,⋅⋅⋅,vp} in R n ℝ^n Rn is linearly dependent if p > n p>n p>n or contains the zero vector.
1.8 Introduction to Linear Transformations (线性变换)
- A transformation (or function or mapping) T T T from R n ℝ^n Rn to R m ℝ^m Rm is a rule that assigns to each vector x ⃗ \vec{x} x in R n ℝ^n Rn a vector T ( x ⃗ ) T(\vec{x}) T(x) in R m ℝ^m Rm .
- The set R n ℝ^n Rn is called domain of T T T, and R m ℝ^m Rm is called the codomain of T T T.
- For x ⃗ \vec{x} x in R n ℝ^n Rn, the vector T ( x ⃗ ) T(\vec{x}) T(x) in R m ℝ^m Rm is called the image of x ⃗ \vec{x} x and the set of all images T ( x ⃗ ) T(\vec{x}) T(x) is called the range of T T T.
- The transformation T : R 2 → R 2 T:ℝ^2\to ℝ^2 T:R2→R2 defined by T ( x ) = A x ⃗ T(x)=A\vec{x} T(x)=Ax is called a shear transformation (剪切变换)
Definition : Linear Transformation
A transformation T T T is linear if:
(i) T ( u ⃗ + v ⃗ ) T(\vec{u}+\vec{v}) T(u+v)= T ( u ⃗ ) T(\vec{u}) T(u)+ T ( v ⃗ ) T(\vec{v}) T(v)
(ii) T ( c u ⃗ ) = c T ( u ⃗ ) T(c\vec{u})=cT(\vec{u}) T(cu)=cT(u)
Given a scalar r r r, define T : R 2 → R 3 T:ℝ^2\toℝ^3 T:R2→R3 by T ( x ⃗ ) = r x ⃗ T(\vec{x})=r\vec{x} T(x)=rx. T T T is called a contraction when 0 ≤ r ≤ 1 0\le r\le1 0≤r≤1 and a dilation when r > 1 r>1 r>1 .
1.9 The Matrix of a Linear Transformation
Theorem 7
Let T : R n → R m T:ℝ^n\toℝ^m T:Rn→Rm be a linear transformation. Then there exists a unique matrix A A A such that T ( x ⃗ ) = A x ⃗ T(\vec{x})=A\vec{x} T(x)=Ax for all x ⃗ \vec{x} x in R n ℝ^n Rn. In fact, A A A is the m × n m\times n m×n matrix whose j t h j^{th} jth column is the vector T ( e j ) T(e_j) T(ej), where e j e_j ej is the j t h j^{th} jth column of the identity matrix in R n ℝ^n Rn , and is called the standard matrix for the linear transformation T T T.
Definition : onto (满射)
A mapping T : R n → R m T:ℝ^n\toℝ^m T:Rn→Rm is said to be onto R m ℝ^m Rm if each b ⃗ \vec{b} b in R m ℝ^m Rm is the image of at least one x ⃗ \vec{x} x in R n ℝ^n Rn. That is, T T T maps R n ℝ^n Rn onto R m R^m Rm if for each b ⃗ \vec{b} b in the codomain R m ℝ^m Rm, there exists at least one solution of T ( x ⃗ ) = b ⃗ T(\vec{x})=\vec{b} T(x)=b, or the columns of A A A span R m ℝ^m Rm.
Definition : one-to-one
A mapping T : R n → R m T:ℝ^n\toℝ^m T:Rn→Rm is said to be one-to-one if each b ⃗ \vec{b} b in R m ℝ^m Rm is the image of at most one x ⃗ \vec{x} x in R n ℝ^n Rn. That is, T T T is one-to-one if the equation T ( x ⃗ ) = 0 T(\vec{x})=0 T(x)=0 has only the trivial solution, or the columns of A A A are linearly independent.
Chapter 2 : Matrix Algebra
2.1 Matrix Operations
Definition : Diagonal Matrix (对角矩阵)
A diagonal matrix is a square n × n n\times n n×n matrix whose nondiagonal entries are zero.
The two matrices are equal if they have the same size and if their corresponding columns are equal.
2.1.1 Sums
(1) A + B = B + A A+B=B+A A+B=B+A
(2) ( A + B ) + C = A + ( B + C ) (A+B)+C=A+(B+C) (A+B)+C=A+(B+C)
(3) A + 0 = A A+0=A A+0=A
(4) r ( A + B ) = r A + r B r(A+B)=rA+rB r(A+B)=rA+rB
(5) ( r + s ) A = r A + s A (r+s)A=rA+sA (r+s)A=rA+sA
(6) r ( s A ) = ( r s ) A r(sA)=(rs)A r(sA)=(rs)A
2.1.2 Scalar Multiples
If A A A is an m × n m\times n m×n matrix and B B B is an n × p n\times p n×p matrix, then the product A B AB AB is the m × p m\times p m×p matrix whose columns are A b 1 , ⋅ ⋅ ⋅ , A b p Ab_1,···,Ab_p Ab1,⋅⋅⋅,Abp
(1) A ( B C ) = ( A B ) C A(BC)=(AB)C A(BC)=(AB)C
(2) A ( B + C ) = A B + A C A(B+C)=AB+AC A(B+C)=AB+AC
(3) ( B + C ) A = B A + C A (B+C)A=BA+CA (B+C)A=BA+CA
(4) r ( A B ) = ( r A ) B = A ( r B ) r(AB)=(rA)B=A(rB) r(AB)=(rA)B=A(rB)
(5) I m A = A = A I n I_mA=A=AI_n ImA=A=AIn
(6) A B ≠ B A AB\ne BA AB=BA
(7) A B = A C ⇏ B = C AB=AC\ \nRightarrow\ B=C AB=AC ⇏ B=C
(8) A B = 0 ⇏ B = 0 AB=0\ \nRightarrow\ B=0 AB=0 ⇏ B=0 or A = 0 A=0 A=0
2.1.3 The Transpose of a Matrix
(1) ( A T ) T = A (A^T)^T=A (AT)T=A
(2) ( A + B ) T = A T + B T (A+B)^T=A^T+B^T (A+B)T=AT+BT
(3) ( r A ) T = r A T (rA)^T=rA^T (rA)T=rAT
(4) ( A B ) T = B T A T (AB)^T=B^TA^T (AB)T=BTAT
2.2 The Inverse of a Matrix
Definition : Invertible
An n × n n\times n n×n matrix A A A is said to be invertible if there is an n × n n\times n n×n matrix C C C such that :
C A = I CA=I CA=I and A C = I AC=I AC=I
We say that C C C is an inverse of A A A, denoted by A − 1 A^{-1} A−1.
Theorem 1
Let A = [ a c b d ] A=\begin{bmatrix}a&c\\b&d\end{bmatrix} A=[abcd]. If a d − b c ≠ 0 ad-bc\ne0 ad−bc=0, then A A A is invertible and A − 1 = 1 a d − b c [ d − c − b a ] A^{-1}=\frac{1}{ad-bc}\begin{bmatrix}d&-c\\-b&a\end{bmatrix} A−1=ad−bc1[d−b−ca]
a d − b c ad-bc ad−bc is called the determinant (行列式) of A A A, that is det A = a d − b c \det A=ad-bc detA=ad−bc.
Theorem 2
If A A A is an invertible n × n n\times n n×n matrix, then for each b ⃗ \vec{b} b in R n ℝ^n Rn, the equation A x ⃗ = b ⃗ A\vec{x}=\vec{b} Ax=b has the unique solution x ⃗ = A − 1 b ⃗ \vec{x}=A^{-1}\vec{b} x=A−1b
Theorem 3
(1) ( A − 1 ) − 1 = A (A^{-1})^{-1}=A (A−1)−1=A
(2) ( A B ) − 1 = B − 1 A − 1 (AB)^{-1}=B^{-1}A^{-1} (AB)−1=B−1A−1
(3) ( A T ) − 1 = ( A − 1 ) T (A^T)^{-1}=(A^{-1})^T (AT)−1=(A−1)T
Definition : Elementary Matrix
An elementary matrix is one that is obtained by performing a single elementary row operation on an identity matrix.
Theorem 4
An n × n n\times n n×n matrix A A A is invertible if and only if A A A is row equivalent to I n I_n In, and in this case, any sequence of elementary row operations that reduces A A A to I n I_n In also transforms I n I_n In into A − 1 A^{-1} A−1.
2.3 Characterizations of Invertible Matrices
Let A A A be a square n × n n\times n n×n matrix. Then the following statements are equivalent.
(1) A A A is an invertible matrix.
(2) A A A is row equvivalent to the n × n n\times n n×n identity matrix.
(3) A A A has n n n pivot positions.
(4) The equation A x ⃗ = 0 A\vec{x}=0 Ax=0 has only the trivial solution.
(5) The columns of A A A span R n ℝ^n Rn.
(6) The columns of A A A form a linearly independent set.
(7) The linear transformation x ⃗ ↦ A x ⃗ \vec x\mapsto A\vec x x↦Ax is onr-to-one.
(8) The equation A x ⃗ = b ⃗ A\vec x=\vec b Ax=b has at least one solution for each b ⃗ \vec b b in R n ℝ^n Rn.
(9) The linear transformation x ⃗ ↦ A x ⃗ \vec x\mapsto A\vec x x↦Ax maps R n ℝ^n Rn onto R n R^n Rn.
(10) A T A^T AT is an invertible matrix.
Theorem 5
Let T : R n → R n T:ℝ^n\to ℝ^n T:Rn→Rn be a linear fransformation and let A A A be the standard matrix for T T T. Then T T T is invertible if and only if A A A is an invertible matrix.
2.4 Partitioned Matrices
- If matrices A A A and B B B are the same size and are partitioned in exactly the same way, then it is natural to make the same partition of the ordinary matrix sum A + B A+B A+B. In this case, each block of A + B A+B A+B is the sum of the corresponding blocks of A A A and B B B.
- Multiplication of a partitioned matrix by a scalar is also computed block by block.
- Partitioned matrices can be multiplied by the usual row—column rule as if the block entries were scalars, provided that for a product AB, the column partition of A matches the row partition of B. (We say that the partitions of A and B are conformable for block multiplication. )
Theorem : Column—Row Expansion of A B AB AB
if A A A is m × n m\times n m×n and B B B is n × p n\times p n×p, then
A B = [ c o l 1 ( A ) c o l 2 ( A ) ⋯ c o l n ( A ) ] [ r o w 1 ( B ) r o w 2 ( B ) ⋮ r o w n ( B ) ] = c o l 1 ( A ) r o w 1 ( B ) + ⋯ + c o l n ( A ) r o w n ( B ) AB=\begin{bmatrix}\mathrm{col}_1(A)&\mathrm{col}_2(A)&\cdots&\mathrm{col}_n(A)\end{bmatrix}\begin{bmatrix}\mathrm{row}_1(B)\\\mathrm{row}_2(B)\\\vdots\\\mathrm{row}_n(B)\end{bmatrix}\\=\mathrm{col}_1(A)\mathrm{row}_1(B)+\cdots+\mathrm{col}_n(A)\mathrm{row}_n(B) AB=[col1(A)col2(A)⋯coln(A)]⎣⎢⎢⎢⎡row1(B)row2(B)⋮rown(B)⎦⎥⎥⎥⎤=col1(A)row1(B)+⋯+coln(A)rown(B)
Chapter 3 : Determinants
3.1 Introduction To Determinants
For n ≥ 2 n\ge 2 n≥2, the determinant of an n × n n\times n n×n matrix A = [ a i j ] A=[a_{ij}] A=[aij] is the sum of n n n terms of the form ± a 1 j det A 1 j \pm a_{1j}\det A_{1j} ±a1jdetA1j.
det A = ∑ j = 1 n ( − 1 ) j + 1 a 1 j det A 1 j \det A=\sum\limits^{n}_{j=1}(-1)^{j+1}a_{1j}\det A_{1j} detA=j=1∑n(−1)j+1a1jdetA1j
Theorem 1
Let C i j = ( − 1 ) i + j det A i j C_{ij}=(-1)^{i+j}\det A_{ij} Cij=(−1)i+jdetAij , then det A = ∑ j = 1 n a 1 j C 1 j \det A=\sum\limits^{n}_{j=1}a_{1j}C_{1j} detA=j=1∑na1jC1j
This fomula is called a cofactor (余子式) expansion across the first row of A A A. The determinant of an n × n n\times n n×n matrix A A A can be computed by a cofactor across any row or down any column, that is :
det A = ∑ j = 1 n a i j C i j \det A=\sum\limits^{n}_{j=1}a_{ij}C_{ij} detA=j=1∑naijCij or det A = ∑ i = 1 n a i j C i j \det A=\sum\limits^{n}_{i=1}a_{ij}C_{ij} detA=i=1∑naijCij
Theorem 2
If A A A is a triangular matrix, then det A \det A detA is the product of the entries on the main diagonal of A A A.
3.2 Properties of determinants
Theorem 3
Let A A A be a square matrix
(1) If a multiple of one row of A A A is added to another row to produce a matrix B B B, then det B = det A \det B=\det A detB=detA .
(2) If two rows of A A A are interchanged to produce B B B, then det B = − det A \det B=-\det A detB=−detA .
(3) If one row of A A A is multiplied by k k k to produce B B B, then det B = k ⋅ det A \det B=k\cdot\det A detB=k⋅detA .
Theorem 4
det A T = det A \det A^T=\det A detAT=detA
det A B = ( det A ) ( det B ) \det AB=(\det A)(\det B) detAB=(detA)(detB)
3.3 Cramer’s Rule, Volume and Linear Transformations
Theorem 1 : Cramer’s Rule
Let A A A be an invertible n × n n\times n n×n matrix. For any b ⃗ \vec b b in R n ℝ^n Rn, the unique solution x ⃗ \vec x x of A x ⃗ = b ⃗ A\vec x=\vec b Ax=b has entries given by
x i = det A i ( b ⃗ ) det A , i = 1 , 2 , ⋯ , n x_i=\frac{\det A_i(\vec b)}{\det A},\ i=1,2,\cdots,n xi=detAdetAi(b), i=1,2,⋯,n
where A i ( b ⃗ ) A_i(\vec b) Ai(b) denotes the i t h i\ th i th column replaced by the vector b ⃗ \vec b b .
Theorem 2
Let A A A be an invertible n × n n\times n n×n matrix. Then A − 1 = a d j A det A A^{-1}=\frac{\mathrm{adj}\ A}{\det A} A−1=detAadj A , where a d j A \mathrm{adj} A adjA is called the adjugate (伴随矩阵) of A .
a d j A = [ C 11 C 21 ⋯ C n 1 C 12 C 22 ⋯ C n 2 ⋮ ⋮ C 1 n C 2 n ⋯ C n n ] \mathrm{adj}A=\begin{bmatrix}C_{11}&C_{21}&\cdots& C_{n1}\\C_{12}&C_{22}&\cdots&C_{n2}\\\vdots&&&\vdots\\ C_{1n}&C_{2n}&\cdots&C_{nn}\end{bmatrix} adjA=⎣⎢⎢⎢⎡C11C12⋮C1nC21C22C2n⋯⋯⋯Cn1Cn2⋮Cnn⎦⎥⎥⎥⎤
Theorem 3 : Determinants as Area or Volume
If A A A is a 2 × 2 2\times2 2×2 matrix, the area of the parallelogram determinant by the columns of A A A is ∣ det A ∣ |\det A| ∣detA∣ If A A A is a 3 × 3 3\times 3 3×3 matrix, the volume of the parallelepiped determined by the columns of A A A is ∣ det A ∣ |\det A| ∣detA∣ .
Theorem 4 : Linear Transformations
Let T : R 2 → R 2 T:ℝ^2\toℝ^2 T:R2→R2 be the linear transformation determined by a 2 × 2 2\times2 2×2 matrix A A A . If S S S is a parallelogram in R 2 ℝ^2 R2, and p ⃗ \vec p p is a vector, then
{ a r e a o f T ( p ⃗ + S ) } = ∣ det A ∣ ⋅ { a r e a o f S } \{area\ of\ T(\vec p+S)\}=|\det A|\cdot\{area\ of\ S\} {area of T(p+S)}=∣detA∣⋅{area of S}
Chpater 4 : Vector Spaces
4.1 Vector Spaces and Subspaces
Definition : Vector Spaces
A set V ≠ ∅ V\ne \varnothing V=∅ admits V × V = V V\times V=V V×V=V (加法) and R × V = V ℝ\times V=V R×V=V (纯量乘法) that obey the rules u ⃗ + v ⃗ = v ⃗ + u ⃗ \vec u+\vec v=\vec v+\vec u u+v=v+u , ( u ⃗ + v ⃗ ) + w ⃗ = u ⃗ + ( v ⃗ + w ⃗ ) (\vec u+\vec v)+\vec w=\vec u+(\vec v+\vec w) (u+v)+w=u+(v+w) , c ( u ⃗ + v ⃗ ) = c u ⃗ + c v ⃗ c(\vec u+\vec v)=c\vec u+c\vec v c(u+v)=cu+cv , 1 u ⃗ = u ⃗ 1\vec u=\vec u 1u=u , ( c + d ) u ⃗ = c u ⃗ + d u ⃗ (c+d)\vec u=c\vec u+d\vec u (c+d)u=cu+du , ( c d ) u ⃗ = c ( d u ⃗ ) , ∀ c , d ∈ R , u , v , w ∈ V (cd)\vec u=c(d\vec u), \forall c,d\inℝ, u,v,w\in V (cd)u=c(du),∀c,d∈R,u,v,w∈V . Denote by 0 ⃗ = 0 u ⃗ \vec 0=0\vec u 0=0u , − u ⃗ = ( − 1 ) u ⃗ -\vec u=(-1)\vec u −u=(−1)u . We call V V V a vector space and the objects in V V V vectors.
Definition : Subspaces
A subspace of a vector space V V V is a subset H H H of V V V that has three properties:
(1) 0 ⃗ \vec 0 0 of V V V is also in H H H .
(2) H H H is closed under vector addition.
(3) H H H is closed under scalar multiple.
Definition : linear variety
Let H H H be a subspace of the vector space V V V. Translate H H H by a vector u ⃗ ∈ V \vec u\in V u∈V, we get u ⃗ + H : = { u ⃗ + v ⃗ ∣ v ⃗ ∈ H } \vec u+H:=\{\vec u+\vec v\ |\ \vec v\in H\} u+H:={u+v ∣ v∈H} called a linear variety (线性簇) .
- In R 3 ℝ^3 R3 , linear varieties can be any planes, lines and points that may be through or not through the origin.
- The solution set of a consistent linear system A x ⃗ = b ⃗ A\vec x=\vec b Ax=b is a linear variety u ⃗ + H \vec u+H u+H with u ⃗ \vec u u a particular solution of A x ⃗ = b ⃗ A\vec x=\vec b Ax=b and H : = { x h ∣ A x h = 0 } H:=\{x_h\ |\ Ax_h=0\} H:={xh ∣ Axh=0} a subspace.
H , K H,K H,K subspaces of V V V, then H + K : = { h ⃗ + k ⃗ ∣ h ⃗ ∈ H , k ⃗ ∈ K } H+K:=\{\vec h+\vec k\ |\ \vec h\in H,\vec k\in K\} H+K:={h+k ∣ h∈H,k∈K} is also a subspace of V V V. Then, H ∩ K H\cap K H∩K is again a subspace of V V V. However, H ∪ K H\cup K H∪K is not a subspace of V V V.
Theorem 1
If V V V is a vector space, v ⃗ 1 , ⋯ , v ⃗ p ∈ V \vec v_1,\cdots,\vec v_p\in V v1,⋯,vp∈V, then H = S p a n { v ⃗ 1 , ⋯ , v ⃗ p } H=\mathrm{Span}\{\vec v_1,\cdots,\vec v_p\} H=Span{v1,⋯,vp} is a subspace of V V V. And { v ⃗ 1 , ⋯ , v ⃗ p } \{\vec v_1,\cdots,\vec v_p\} {v1,⋯,vp} is a spanning set (生成集) of H H H.
{ v ⃗ 1 , ⋯ , v ⃗ p } \{\vec v_1,\cdots,\vec v_p\} {v1,⋯,vp} is a spanning set (生成集) of H H H if and only if
∀ b ⃗ ∈ H , ∃ x ⃗ = [ x 1 , ⋯ , x p ] T ∈ R p \forall\vec b\in H,\ \exists\vec x=[x_1,\cdots,x_p]^T\in ℝ^p ∀b∈H, ∃x=[x1,⋯,xp]T∈Rp such that x 1 v ⃗ 1 + ⋯ + x p v ⃗ p = b ⃗ x_1\vec v_1+\cdots+x_p\vec v_p=\vec b x1v1+⋯+xpvp=b
4.2 Null Space and Column Space
4.2.1 Null Space (零空间)
Let A ∈ M m × n A\in M_{m\times n} A∈Mm×n. We call N u l A : = { x ⃗ ∣ x ⃗ ∈ R n a n d A x ⃗ = 0 } \mathrm{Nul}A:=\{\vec x\ |\ \vec x\inℝ^n\ \mathrm{and}\ A\vec x=0\} NulA:={x ∣ x∈Rn and Ax=0} the Null Space of A A A.
Theorem 2
Let A ∈ M m × n A\in M_{m\times n} A∈Mm×n, then N u l A \mathrm{Nul}A NulA is a subspace of R n ℝ^n Rn
Properties of Null Space
Let A ∈ M m × n A\in M_{m\times n} A∈Mm×n
- N u l A = { 0 } \mathrm{Nul}A=\{0\} NulA={0} is equivalent to that A x ⃗ = 0 A\vec x=0 Ax=0 has only the trivial solution x ⃗ = 0 \vec x=0 x=0 .
- N u l A = { 0 } \mathrm{Nul}A=\{0\} NulA={0} is equivalent to that x ⃗ ↦ A x ⃗ \vec x\mapsto A\vec x x↦Ax is a ont-to-one map from R n ℝ^n Rn to R m ℝ^m Rm.
- N u l A = { 0 } \mathrm{Nul}A=\{0\} NulA={0} implies m ≥ n m\ge n m≥n
- N u l A = R n \mathrm{Nul}A=ℝ^n NulA=Rn is equivalent to A = 0 A=0 A=0
Definition : nullity (零化度)
零空间的维度,即 : dim N u l A \dim\mathrm{Nul}A dimNulA
4.2.2 Column Space (列空间)
Let A = [ c ⃗ 1 , ⋯ , c ⃗ n ] ∈ M m × n A=[\vec c_1,\cdots,\vec c_n]\in M_{m\times n} A=[c1,⋯,cn]∈Mm×n . We call C o l A : = S p a n { c ⃗ 1 , ⋯ , c ⃗ n } \mathrm{Col}A:=\mathrm{Span}\{\vec c_1,\cdots,\vec c_n\} ColA:=Span{c1,⋯,cn} the Column Space of A A A.
Properties of Column Space
Let A ∈ M m × n A\in M_{m\times n} A∈Mm×n .
- C o l A = { b ⃗ ∈ R m ∣ ∃ x ⃗ ∈ R n s . t . A x ⃗ = b ⃗ } \mathrm{Col}A=\{\vec b\inℝ^m\ |\ \exists\vec x\inℝ^n\ s.t.\ A\vec x=\vec b\} ColA={b∈Rm ∣ ∃x∈Rn s.t. Ax=b}
- C o l A = \mathrm{Col}A= ColA= range A A A with A : R n → R m , x ⃗ ↦ A x ⃗ A:ℝ^n\toℝ^m\ ,\ \vec x\mapsto A\vec x A:Rn→Rm , x↦Ax
- C o l A \mathrm{Col}A ColA is a subspace of R m ℝ^m Rm
- C o l A = R m \mathrm{Col}A=ℝ^m ColA=Rm if and only if ∀ b ⃗ ∈ R m , ∃ x ⃗ ∈ R n \forall\vec b\inℝ^m,\ \exists\vec x\inℝ^n ∀b∈Rm, ∃x∈Rn such that A x ⃗ = b ⃗ A\vec x=\vec b Ax=b
- C o l A = R m \mathrm{Col}A=ℝ^m ColA=Rm if and only if A : R n → R m , x ⃗ ↦ A x ⃗ A:ℝ^n\toℝ^m\ ,\ \vec x\mapsto A\vec x A:Rn→Rm , x↦Ax is onto.
- C o l A = R m \mathrm{Col}A=ℝ^m ColA=Rm if and only if the number of pivot rows equals m m m .
- C o l A = R m \mathrm{Col}A=ℝ^m ColA=Rm implies that m ≤ n m\le n m≤n
- if m = n m=n m=n then C o l A = R n ⇔ N u l A = { 0 } \mathrm{Col}A=ℝ^n\Leftrightarrow\mathrm{Nul}A=\{0\} ColA=Rn⇔NulA={0}
4.3 Basis
Theorem 1
Let V V V be a vector space. { v ⃗ 1 , ⋯ v ⃗ p } ⊂ V ( p ≥ 2 ) \{\vec v_1,\cdots\,\vec v_p\}\subset V(p\ge2) {v1,⋯vp}⊂V(p≥2) and v ⃗ 1 ≠ 0 \vec v_1\ne0 v1=0, then { v ⃗ 1 , ⋯ v ⃗ p } \{\vec v_1,\cdots\,\vec v_p\} {v1,⋯vp} is linearly dependent if and only if ∃ j ∈ { 2 , ⋯ , p } \exists j\in\{2,\cdots,p\} ∃j∈{2,⋯,p} such that v j v_j vj is a linear combination of { v ⃗ 1 , ⋯ , v ⃗ j − 1 } \{\vec v_1,\cdots,\vec v_{j-1}\} {v1,⋯,vj−1}
Theorem 2
Let S = { v ⃗ 1 , ⋯ , v ⃗ p } ⊂ S=\{\vec v_1,\cdots,\vec v_p\}\subset S={v1,⋯,vp}⊂ vector space V V V, H = S p a n S H=\mathrm{Span}S H=SpanS .
(1) If S ∋ v ⃗ k ∈ S p a n ( S \ { v ⃗ k } ) S\ni\vec v_k\in\mathrm{Span}(S\backslash\{\vec v_k\}) S∋vk∈Span(S\{vk}), then S p a n ( S \ { v ⃗ k } ) = H \mathrm{Span}(S\backslash\{\vec v_k\})=H Span(S\{vk})=H.
(2) If S S S is linearly independent, then every proper subset (真子集) of S S S generate a proper subspace (真子空间) of H H H.
Definition : Basis
Let H H H be a subspace of the vector space V V V. B = { b ⃗ 1 , ⋯ , b ⃗ p } ⊂ H B=\{\vec b_1,\cdots,\vec b_p\}\subset H B={b1,⋯,bp}⊂H is said to be a basis for H H H if :
(i) B B B is linearly independent and (ii) S p a n B = H \mathrm{Span}B=H SpanB=H
Theorem 3
Let S = { v ⃗ 1 , ⋯ , v ⃗ p } ⊂ S=\{\vec v_1,\cdots,\vec v_p\}\subset S={v1,⋯,vp}⊂ vector space V V V, H = S p a n S H=\mathrm{Span}S H=SpanS. If H ≠ { 0 } H\ne\{0\} H={0}, then ∃ B ⊂ S \exists B\subset S ∃B⊂S is a basis of H H H.
Theorem 4
The pivot columns of A A A is a basis of H = C o l A H=\mathrm{Col}A H=ColA .
- Basis for a vector space is a spanning set as small as possible.
- Basis for a vector space is a linearly independent set as large as possible.
4.4 Coordinate Systems
Suppose B = { b ⃗ 1 , ⋯ , b ⃗ n } B=\{\vec b_1,\cdots,\vec b_n\} B={b1,⋯,bn} is a basis for a vector space V V V. Let x ⃗ ∈ V \vec x\in V x∈V. The coordinates of x ⃗ \vec x x relative to the basis B B B are the weights c 1 , ⋯ c n c_1,\cdots c_n c1,⋯cn. We denoted by [ x ⃗ ] B = [ c 1 ⋮ c n ] [\vec x]_B=\begin{bmatrix}c_1\\\vdots\\c_n\end{bmatrix} [x]B=⎣⎢⎡c1⋮cn⎦⎥⎤ the B-coordinate vector of x ⃗ \vec x x. The mapping x ⃗ ↦ [ x ⃗ ] B \vec x\mapsto[\vec x]_B x↦[x]B is the coordinate mapping.
Theorem 1
Let B = { b ⃗ 1 , ⋯ , b ⃗ n } B=\{\vec b_1,\cdots,\vec b_n\} B={b1,⋯,bn} is a basis for a vector space V V V. Then the coordinate mapping x ⃗ ↦ [ x ⃗ ] B \vec x\mapsto [\vec x]_B x↦[x]B is a one to one linear transformation from V V V onto R n ℝ^n Rn.
4.4.1 Isomorphic Vector Spaces (同构的向量空间)
If there exists a one-to-one linear map T T T from a vector space V V V onto a vector space W W W, then we say V V V and W W W are isomorphic and T T T is an isomorphism from V V V onto W W W.
- Coordinate mapping is an isomorphism.
- Every vector space with a basis of n n n vectors is isomorphic to R n ℝ^n Rn.
If the vector spaces V V V and W W W are isomorphic with an isomorphism T T T, then
- Linearly (in)dependent set in V V V is mapped to linearly (in)dependent set in W W W by T T T.
- Linearly dependence relations in V V V are carried over by T T T to W W W.
4.4.2 Coordinate Mapping in Subspaces of R n ℝ^n Rn
- Let H H H be a subspace of R n ℝ^n Rn and B = { b ⃗ 1 , ⋯ , b ⃗ p } B=\{\vec b_1,\cdots,\vec b_p\} B={b1,⋯,bp} be a basis for H H H. Then, we must have 1 ≤ p ≤ n 1\le p\le n 1≤p≤n.
- Let
P
B
=
[
b
⃗
1
,
⋯
,
b
⃗
p
]
P_B=[\vec b_1,\cdots,\vec b_p]
PB=[b1,⋯,bp], then the coordinate mapping is ‘inverse’ in some sense.
4.5 Dimension of a Vector Space
Theorem 1
If a vector space V V V has a basis B = { b ⃗ 1 , ⋯ , b ⃗ n } B=\{\vec b_1,\cdots,\vec b_n\} B={b1,⋯,bn}, then any set in V V V containing more than n n n vectors must be linearly dependent.
Corollary
A linearly independent set must have ≤ n \le n ≤n vectors.
Theorem 2
If a vector space V V V has a basis of n n n vectors, then every basis of V V V must contain exactly n n n vectors.
Definition : Dimension of a Vector Space
If V V V is spanned by a finite set, then V V V is said to be finite-dimensional, and the dimension of V V V, written as dim V \dim V dimV, is the number of vectors in a basis for V V V. The dimension of the zero vector space { 0 } \{0\} {0} is defined to be zero. If V V V is not spanned by a finite set, then V V V is said to be infinite-dimensional.
- Isomorphic vector spaces have the same dimension.
- Dimension is the cardinality of a basis.
- Basis for N u l A \mathrm{Nul}A NulA and C o l A \mathrm{Col}A ColA can be found by the echelon form of A.
- Number of free variables is dim N u l A \dim\mathrm{Nul}A dimNulA
- Number of pivot columns is dim C o l A \dim\mathrm{Col}A dimColA
- dim N u l A \dim\mathrm{Nul}A dimNulA+ dim C o l A = \dim\mathrm{Col}A= dimColA= total number of columns in A A A .
Theorem 3 : Expand a Linearly Independent Set to a Basis
Let V V V be a vector space of finite dimension, H H H a nonzero subspace of V V V.
- If we expand a linearly independent set in H H H with more vectors from H H H until inserting any vector will make the set linearly dependent, then this maximal linearly independent set must be a basis for H H H.
- dim H ≤ dim V \dim H\le \dim V dimH≤dimV
Theorem 4
Let V V V be a p p p-dimensional vector space, p ≥ 1 p\ge1 p≥1 .
(1) Any linearly independent set of exactly p p p vectors in V V V is automatically a basis for V V V .
(2) Any spanning set of exactly p p p vectors is also a basis for V V V.
cardinality of any linearly independent set ≤ \ \le ≤ cardinality of any basis = dim V ≤ \ =\ \dim V\ \le = dimV ≤ cardinality of any spanning set
4.6 Row Space and Rank of a Matrix
4.6.1 Row Space
R o w A : = S p a n { r ⃗ 1 , ⋯ , r ⃗ m } \mathrm{Row}A:=\mathrm{Span}\{\vec r_1,\cdots,\vec r_m\} RowA:=Span{r1,⋯,rm} for A ∈ M m × n A\in M_{m\times n} A∈Mm×n with the rows { r ⃗ 1 , ⋯ , r ⃗ m } \{\vec r_1,\cdots,\vec r_m\} {r1,⋯,rm} is called the row space of A A A .
- R o w A \mathrm{Row}A RowA is a subspace of R n ℝ^n Rn
- R o w A \mathrm{Row}A RowA can be identified with C o l A T \mathrm{Col}A^T ColAT
Theorem 1
(1) If A A A and B B B are row equivalent, then R o w A = R o w B \mathrm{Row}A=\mathrm{Row}B RowA=RowB .
(2) If B B B is an echelon form of A A A, then the nonzero rows of B B B form a basis for R o w A \mathrm{Row}A RowA .
4.6.2 Rank
r a n k A : = dim C o l A \mathrm{rank}A:=\dim\mathrm{Col}A rankA:=dimColA
Rank Theorem
(1) r a n k A = dim R o w A \mathrm{rank}A=\dim\mathrm{Row}A rankA=dimRowA
(2) r a n k A + dim N u l A = n \mathrm{rank}A+\dim\mathrm{Nul}A=n rankA+dimNulA=n for A ∈ M m × n A\in M_{m\times n} A∈Mm×n
Chapter 5 : Eigenvalues and Eigenvectors
5.1 Eigenvalues and Eigenvectors
An eigenvector of an n × n n\times n n×n matrix A A A is a nonzero vector x ⃗ \vec x x such that A x ⃗ = λ x ⃗ A\vec x=\lambda \vec x Ax=λx for some scalar λ \lambda λ, and the scalar λ \lambda λ is called eigenvalue.
λ \lambda λ is an eigenvalue of an n × n n\times n n×n matrix A A A if and only if the equation ( A − λ I ) x ⃗ = 0 (A-\lambda I)\vec x=0 (A−λI)x=0 has a nontrivial solution. The set of all solutions is a subspace of R n ℝ^n Rn and is called the eigenspace of A A A corresponding to λ \lambda λ .
Theorem 1
The eigenvalues of a triangular matrix are the entries on its main diagonal.
Theorem 2
If v ⃗ 1 , ⋯ , v ⃗ r \vec v_1,\cdots,\vec v_r v1,⋯,vr are eigenvectors that correspond to distinct eigenvalues λ 1 , ⋯ , λ r \lambda_1,\cdots,\lambda_r λ1,⋯,λr of an n × n n\times n n×n matrix A A A, then the set { v ⃗ 1 , ⋯ , v ⃗ r } \{\vec v_1,\cdots,\vec v_r\} {v1,⋯,vr} is linearly independent.
λ k \lambda^k λk is the eigenvalue of matrix A k A^k Ak.
5.2 The Characteristic equation
Let A A A be an n × n n\times n n×n matrix, let U U U be any echelon form obtained from A A A by row replacements and row interchanges, and let r r r be the number of such row interchanges. The the determinant of A A A, is ( − 1 ) r (-1)^r (−1)r times the product of the diagonal entries.
Theorem 1
Let A A A be an n × n n\times n n×n matrix. Then A A A is invertible if and only if:
(1) The number 0 is not an eigenvalue of A A A .
(2) The determinant of A A A is not zero.
Definition : Characteristic Equation
The scalar equation det ( A − λ I ) = 0 \det(A-\lambda I)=0 det(A−λI)=0 is called the characteristic equation of A A A . A scalar λ \lambda λ is an eigenvalue of an n × n n\times n n×n matrix A A A if and only if λ \lambda λ satisfies the characteristic equation.
det ( A − λ I ) \det(A-\lambda I) det(A−λI) is a polynomial of degree n n n called the characteristic polynomial of A A A.
Definition : Similarity
If A A A and B B B are n × n n\times n n×n matrices, then A A A is similar to B B B if there is an invertible matrix P P P such that P − 1 A P = B P^{-1}AP=B P−1AP=B, or, equivalently, A = P B P − 1 A=PBP^{-1} A=PBP−1.
Change A A A into P − 1 A P P^{-1}AP P−1AP is called a similarity transformation.
Theorem 2
If n × n n\times n n×n matrices A A A and B B B are similar, then they have the same characteristic polynomial and hence the same eigenvalues.
Warnings :
- The matrices [ 2 1 0 2 ] \begin{bmatrix}2&1\\0&2\end{bmatrix} [2012] and [ 1 0 0 2 ] \begin{bmatrix}1&0\\0&2\end{bmatrix} [1002] are not similar even though they ha ve the same eigenvalues.
- Similarity is not the same as row equivalence. Row operations on a matrix usually change its eigenvalues.
洋洋洒洒 2w字,努力将所有知识点记了下来,也不知道会有多大用处,但如果对你起到作用了的话,请务必 关注+点赞+收藏 感激不尽。临表涕零,不知所言。