线性方程组 Linear Equations

简介 Introduction

在线性代数中,两个重要的方程组形式是 。这两个方程组涉及到矩阵 和向量 ,它们的解集有着不同的性质和几何意义。理解这些解集对于线性代数的学习和应用至关重要。

In linear algebra, two important forms of systems of equations are and . These systems involve a matrix and vectors and , and their solution sets have different properties and geometric meanings. Understanding these solution sets is crucial for learning and applying linear algebra.

齐次线性方程组 Homogeneous System

定义与性质 Definition and Properties

  1. 定义 Definition

这里,$A$ 是一个 $m \times n$ 的矩阵,$x$ 是一个 $n$ 维列向量,零向量 $0$ 是一个 $m$ 维列向量。 Here, $A$ is an $m \times n$ matrix, $x$ is an $n$-dimensional column vector, and the zero vector $0$ is an $m$-dimensional column vector. 2. **解集 Solution Set**: $Ax = 0$ 的解集被称为矩阵 $A$ 的 **零空间(null space)** 或 **核(kernel)**,记作 $N(A)$。 这个解集是一个向量空间。 The solution set of $Ax = 0$ is called the **null space** or **kernel** of the matrix $A$, denoted as $N(A)$. This solution set is a vector space. ### 求解方法 Solution Methods 1. **行简化 Row Reduction**: 将矩阵 $A$ 进行 **简化行阶梯形矩阵(reduced row echelon form, RREF)** 变换。 Reduce the matrix $A$ to its **reduced row echelon form (RREF)**. 例 Example:

A = \begin{pmatrix}

1 & 2 & 1 \

2 & 4 & 2 \

3 & 6 & 3

\end{pmatrix} \rightarrow \text{RREF} \rightarrow \begin{pmatrix}

1 & 2 & 1 \

0 & 0 & 0 \

0 & 0 & 0

\end{pmatrix}

通过 RREF,可以发现 $x_1 + 2x_2 + x_3 = 0$,即 $$x = t_1 \begin{pmatrix} -2 \\ 1 \\ 0 \end{pmatrix} + t_2 \begin{pmatrix} -1 \\ 0 \\ 1 \end{pmatrix}$,其中 $t_1, t_2 \in \mathbb{R}$$。 By RREF, we find that $x_1 + 2x_2 + x_3 = 0$, hence $$x = t_1 \begin{pmatrix} -2 \\ 1 \\ 0 \end{pmatrix} + t_2 \begin{pmatrix} -1 \\ 0 \\ 1 \end{pmatrix}$, where $t_1, t_2 \in \mathbb{R}$$. 2. **使用分块矩阵和转置矩阵 Constructing Solution with Block Matrix and Transpose Matrix**: 假设矩阵 $A$ 的 RREF 可以表示为分块矩阵:

\text{RREF}(A) = \begin{pmatrix}

I & B \

O & O

\end{pmatrix}

其中 $I$ 是单位矩阵,$B$ 是适当大小的矩阵,$O$ 是零矩阵。用 $B$ 的转置 $B^T$ 和单位矩阵来构造零空间的基础解系。 Suppose the RREF of matrix $A$ can be represented as a block matrix:

\text{RREF}(A) = \begin{pmatrix}

I & B \

O & O

\end{pmatrix}

where $I$ is the identity matrix, $B$ is an appropriately sized matrix, and $O`$ is the zero matrix. Use the transpose of $B$, $B^T$, and the identity matrix to construct the basis for the null space. 例 Example:

A = \begin{pmatrix}

1 & 2 & 1 & 3 \

2 & 4 & 2 & 6 \

3 & 6 & 3 & 9

\end{pmatrix} \rightarrow \text{RREF} \rightarrow \begin{pmatrix}

1 & 2 & 1 & 3 \

0 & 0 & 0 & 0 \

0 & 0 & 0 & 0

\end{pmatrix}

从 RREF 可以看到 $x_2, x_3, x_4$ 是自由变量。构造 $B^T$:

B = \begin{pmatrix}

2 & 1 & 3

\end{pmatrix}, \quad B^T = \begin{pmatrix}

2 \

1 \

3

\end{pmatrix}

x = t_1 \begin{pmatrix}

-2 \

1 \

0 \

0

\end{pmatrix} + t_2 \begin{pmatrix}

-1 \

0 \

1 \

0

\end{pmatrix} + t_3 \begin{pmatrix}

-3 \

0 \

0 \

1

\end{pmatrix}

其中 $t_1, t_2, t_3 \in \mathbb{R}$。 From the RREF, we see that $x_2, x_3, x_4$ are free variables. Construct $B^T$:

B = \begin{pmatrix}

2 & 1 & 3

\end{pmatrix}, \quad B^T = \begin{pmatrix}

2 \

1 \

3

\end{pmatrix}

x = t_1 \begin{pmatrix}

-2 \

1 \

0 \

0

\end{pmatrix} + t_2 \begin{pmatrix}

-1 \

0 \

1 \

0

\end{pmatrix} + t_3 \begin{pmatrix}

-3 \

0 \

0 \

1

\end{pmatrix}

where $t_1, t_2, t_3 \in \mathbb{R}$. ## 非齐次线性方程组 $Ax = b$ Non-Homogeneous System $Ax = b$ ### 定义与性质 Definition and Properties 1. **定义 Definition**:

Ax = b

这里,$b$ 是一个 $m$ 维列向量。 Here, $b$ is an $m$-dimensional column vector. 2. **解集 Solution Set**: 如果 $Ax = b$ 有解,那么解集可以表示为一个特解 $x_p$ 加上齐次方程组 $Ax = 0$ 的解空间,即

x = x_p + N(A)

If $Ax = b$ has a solution, the solution set can be expressed as a particular solution $x_p$ plus the solution space of the homogeneous system $Ax = 0$, i.e.,

x = x_p + N(A)

### 求解方法 Solution Methods 1. **行简化 Row Reduction**: 将增广矩阵 $[A | b]$ 进行行简化,直到得到简化行阶梯形矩阵。 Perform row reduction on the augmented matrix $[A | b]$ until you obtain the reduced row echelon form (RREF). 例 Example:

A = \begin{pmatrix}

1 & 2 & 1 \

2 & 4 & 2 \

3 & 6 & 3

\end{pmatrix}, \quad b = \begin{pmatrix}

1 \

2 \

3

\end{pmatrix} \rightarrow [A | b] = \begin{pmatrix}

1 & 2 & 1 & 1 \

2 & 4 & 2 & 2 \

3 & 6 & 3 & 3

\end{pmatrix} \rightarrow \text{RREF} \rightarrow \begin{pmatrix}

1 & 2 & 1 & 1 \

0 & 0 & 0 & 0 \

0 & 0 & 0 & 0

\end{pmatrix}

通过 RREF,可以发现 $x_1 + 2x_2 + x_3 = 1$,解集为 $$x = \begin{pmatrix} 1 \\ 0 \\ 0 \end{pmatrix} + t_1 \begin{pmatrix} -2 \\ 1 \\ 0 \end{pmatrix} + t_2 \begin{pmatrix} -1 \\ 0 \\ 1 \end{pmatrix}$$,其中 $t_1, t_2 \in \mathbb{R}$。 By RREF, we find that $x_1 + 2x_2 + x_3 = 1$, hence the solution set is $$x = \begin{pmatrix} 1 \\ 0 \\ 0 \end{pmatrix} + t_1 \begin{pmatrix} -2 \\ 1 \\ 0 \end{pmatrix} + t_2 \begin{pmatrix} -1 \\ 0 \\ 1 \end{pmatrix}$$, where $t_1, t_2 \in \mathbb{R}$。 ## Rouché-Capelli Theorem ### Definition / 定义 **Rouché-Capelli Theorem**: This theorem states that a system of linear equations $\mathbf{A}\mathbf{x} = \mathbf{b}$ has at least one solution if and only if the rank of the coefficient matrix $\mathbf{A}$ is equal to the rank of the augmented matrix $[\mathbf{A} | \mathbf{b}]$.

\text{rank}(\mathbf{A}) = \text{rank}([\mathbf{A} | \mathbf{b}])

**Rouché-Capelli 定理**:该定理指出,对于线性方程组 $\mathbf{A}\mathbf{x} = \mathbf{b}$,当且仅当系数矩阵 $\mathbf{A}$ 的秩等于增广矩阵 $[\mathbf{A} | \mathbf{b}]$ 的秩时,方程组至少有一个解。

\text{rank}(\mathbf{A}) = \text{rank}([\mathbf{A} | \mathbf{b}])

### Properties / 性质 1. **Consistent System**: If $\text{rank}(\mathbf{A}) = \text{rank}([\mathbf{A} | \mathbf{b}])$, the system is consistent (i.e., it has at least one solution). - **一致性系统**:如果 $\text{rank}(\mathbf{A}) = \text{rank}([\mathbf{A} | \mathbf{b}])$,则该系统是一致的(即,它至少有一个解)。 2. **Inconsistent System**: If $\text{rank}(\mathbf{A}) \neq \text{rank}([\mathbf{A} | \mathbf{b}])$, the system is inconsistent (i.e., it has no solutions). - **不一致性系统**:如果 $\text{rank}(\mathbf{A}) \neq \text{rank}([\mathbf{A} | \mathbf{b}])$,则该系统是不一致的(即,它没有解)。 3. **Unique Solution**: If $\text{rank}(\mathbf{A}) = \text{rank}([\mathbf{A} | \mathbf{b}]) = n$ (where $n$ is the number of unknowns), the system has a unique solution. - **唯一解**:如果 $\text{rank}(\mathbf{A}) = \text{rank}([\mathbf{A} | \mathbf{b}]) = n$(其中 $n$ 是未知数的数量),则该系统有唯一解。 4. **Infinite Solutions**: If $\text{rank}(\mathbf{A}) = \text{rank}([\mathbf{A} | \mathbf{b}]) < n$, the system has infinitely many solutions. - **无限多解**:如果 $\text{rank}(\mathbf{A}) = \text{rank}([\mathbf{A} | \mathbf{b}]) < n$,则该系统有无限多个解。