Eigenvalues and Eigenvectors: Core Concepts Explained | 矩阵特征值与特征向量：核心概念解析

📚 Eigenvalues and Eigenvectors: Core Concepts Explained | 矩阵特征值与特征向量：核心概念解析

Eigenvalues and eigenvectors are among the most powerful ideas in linear algebra. They reveal the intrinsic structure of a matrix and appear everywhere from physics to data science. In this article, we will clarify the definitions, computation methods, key properties, and applications at an IB Mathematics level.

特征值与特征向量是线性代数中最强大的概念之一。它们揭示了矩阵的内在结构，从物理学到数据科学无处不在。本文将以 IB 数学的难度，澄清定义、计算方法、核心性质与应用。

---

1. Definition | 定义

Let A be a square matrix. A nonzero vector v is called an eigenvector of A if multiplying A by v only scales v. That is, there exists a scalar λ such that:

设 A 是一个方阵。若非零向量 v 满足 A 乘以 v 只是对 v 进行缩放，则称 v 为 A 的特征向量。也就是说，存在一个标量 λ，使得：

Av = λv

Here λ is called the eigenvalue associated with v. The zero vector is excluded because it always satisfies the equation with any λ, which would make the definition trivial.

这里的 λ 称为与 v 对应的特征值。零向量通常被排除在外，因为零向量配上任意 λ 都满足方程，那样定义就没有意义了。

---

2. Geometric Meaning | 几何意义

A matrix can be viewed as a linear transformation that stretches, rotates, or reflects space. An eigenvector is a special direction that does not change direction under that transformation; it only gets stretched or compressed by the factor λ. If λ is negative, the vector flips to the opposite direction.

矩阵可以看作是对空间的线性变换，例如拉伸、旋转或反射。特征向量是在该变换下方向不变的特定方向，它只是被因子 λ 拉伸或压缩。如果 λ 为负，向量会翻转到相反方向。

For example, the identity matrix I scales every vector by 1. Thus every nonzero vector is an eigenvector with eigenvalue 1.

例如，单位矩阵 I 将所有向量按 1 缩放，因此每个非零向量都是特征向量，特征值为 1。

• λ > 1: the eigenvector is stretched.
• 0 < λ < 1: the eigenvector is compressed.
• λ < 0: the eigenvector is reversed and stretched/compressed.

• λ > 1：特征向量被拉伸。
• 0 < λ < 1：特征向量被压缩。
• λ < 0：特征向量被反转并拉伸/压缩。

---

3. The Characteristic Equation | 特征方程

To find eigenvalues, we rewrite Av = λv as:

为了求特征值，我们把 Av = λv 改写为：

(A − λI)v = 0

For a nonzero v to exist, the matrix A − λI must be singular, so its determinant must be zero:

要使非零向量 v 存在，矩阵 A − λI 必须是奇异的，因此其行列式必须为零：

det(A − λI) = 0

This polynomial in λ is called the characteristic polynomial. Its roots are the eigenvalues of A.

这个关于 λ 的多项式称为特征多项式，它的根就是 A 的特征值。

---

4. Computing Eigenvalues: 2×2 Example | 计算特征值：2×2 示例

Let A = [[2, 1], [1, 2]]. We form A − λI:

设 A = [[2, 1], [1, 2]]。我们构造 A − λI：

det[[2−λ, 1], [1, 2−λ]] = (2−λ)² − 1 = 0

Expanding gives λ² − 4λ + 3 = 0, so λ = 1 or λ = 3. These are the eigenvalues.

展开得 λ² − 4λ + 3 = 0，因此 λ = 1 或 λ = 3。这就是特征值。

---

5. Computing Eigenvectors | 计算特征向量

For each eigenvalue λ, substitute λ into (A − λI)v = 0 and solve the linear system.

对每个特征值 λ，将 λ 代入 (A − λI)v = 0，并解线性方程组。

For λ = 3 in the example above:

在上面的例子中，对 λ = 3：

[[−1, 1], [1, −1]][x; y] = [0; 0]

This gives −x + y = 0, so x = y. A basis eigenvector is [1; 1]. Similarly for λ = 1, we get x = −y, with eigenvector [1; −1].

得到 −x + y = 0，所以 x = y。一个基础特征向量是 [1; 1]。类似地，对 λ = 1，得到 x = −y，特征向量为 [1; −1]。

---

6. Properties of Eigenvalues | 特征值的性质

Several useful facts connect eigenvalues to matrix invariants:

以下重要事实将特征值与矩阵不变量联系起来：

• The sum of all eigenvalues (counting multiplicity) equals the trace of the matrix: tr(A).
• The product of all eigenvalues equals the determinant: det(A).
• If A is triangular, its eigenvalues are exactly the diagonal entries.
• A and its transpose Aᵀ have the same eigenvalues.

• 所有特征值之和（按重数计）等于矩阵的迹 tr(A)。
• 所有特征值之积等于行列式 det(A)。
• 若 A 是三角矩阵，其特征值恰好是对角线元素。
• A 与其转置 Aᵀ 有相同的特征值。

These properties often allow quick checks of computed eigenvalues.

这些性质常用于快速检验算出的特征值是否正确。

---

7. Diagonalization | 对角化

If an n×n matrix A has n linearly independent eigenvectors, it can be diagonalized. Let P be the matrix whose columns are these eigenvectors, and let D be the diagonal matrix with corresponding eigenvalues on its diagonal. Then:

如果 n×n 矩阵 A 有 n 个线性无关的特征向量，则它可以对角化。设 P 是以这些特征向量为列的矩阵，D 是对角线元素为对应特征值的对角矩阵，则：

A = PDP⁻¹

Diagonalization simplifies computing powers: Aᵏ = PDᵏP⁻¹, where Dᵏ is just raising each diagonal entry to the k-th power.

对角化简化了矩阵幂的计算：Aᵏ = PDᵏP⁻¹，其中 Dᵏ 只需将对角线每个元素取 k 次幂。

---

8. Symmetric Matrices | 对称矩阵

Real symmetric matrices (where Aᵀ = A) have very special eigenvalue properties:

实对称矩阵（满足 Aᵀ = A）具有非常特殊的特征值性质：

• All eigenvalues are real.
• Eigenvectors corresponding to distinct eigenvalues are orthogonal.
• Every symmetric matrix can be orthogonally diagonalized.

• 所有特征值都是实数。
• 不同特征值对应的特征向量相互正交。
• 每个对称矩阵都可以正交对角化。

This is why symmetric matrices are central in multivariable calculus and quadratic forms.

这就是为什么对称矩阵在多变量微积分和二次型中处于核心地位。

---

9. Matrix Powers and Markov Chains | 矩阵幂与马尔可夫链

Eigenvalues are indispensable for studying long-term behaviour of iterated systems. If A is a transition matrix of a Markov chain, its eigenvalues determine whether the chain converges and how fast.

在研究迭代系统的长期行为时，特征值不可或缺。若 A 是马尔可夫链的转移矩阵，其特征值决定链是否收敛以及收敛速度。

For a 2×2 matrix, if we write v in the eigenvector basis, then applying A repeatedly scales each component by the corresponding eigenvalue. If |λ| < 1, those components decay; if |λ| > 1, they grow.

对 2×2 矩阵，若把 v 写在特征向量基下，反复应用 A 相当于将各分量按对应特征值缩放。若 |λ| < 1，对应分量衰减；若 |λ| > 1，则增长。

---

10. Worked Problem | 完整例题

Find the eigenvalues and eigenvectors of B = [[1, 2], [2, 1]].

求矩阵 B = [[1, 2], [2, 1]] 的特征值和特征向量。

First solve det(B − λI) = 0:

首先解 det(B − λI) = 0：

(1−λ)² − 4 = 0

So (1−λ)² = 4, giving 1−λ = ±2. Hence λ = 3 or λ = −1.

因此 (1−λ)² = 4，得 1−λ = ±2。所以 λ = 3 或 λ = −1。

For λ = 3, solve [[−2, 2], [2, −2]][x; y] = [0; 0], giving x = y. Eigenvector: [1; 1].

对 λ = 3，解 [[−2, 2], [2, −2]][x; y] = [0; 0]，得 x = y。特征向量：[1; 1]。

For λ = −1, solve [[2, 2], [2, 2]][x; y] = [0; 0], giving x = −y. Eigenvector: [1; −1].

对 λ = −1，解 [[2, 2], [2, 2]][x; y] = [0; 0]，得 x = −y。特征向量：[1; −1]。

---

11. Common Pitfalls | 常见易错点

Students often make the following mistakes:

同学们常犯以下错误：

Error	Correction
Forgetting that eigenvectors are nonzero.	Always exclude the zero vector from eigenvector sets.
Writing det(A − λI) incorrectly.	Subtract λ only from the diagonal entries.
Assuming every 2×2 matrix has two distinct eigenvalues.	Eigenvalues may repeat (e.g. λ = 2, 2).
Using an eigenvector to compute another eigenvector incorrectly.	Check each eigenvector separately in the original equation.

错误	更正
忘记特征向量是非零向量。	特征向量集合中必须排除零向量。
det(A − λI) 写错。	只在对角线元素上减去 λ。
假设每个 2×2 矩阵都有两个不同特征值。	特征值可能重复（如 λ = 2, 2）。
用某个特征向量推导另一个特征向量时出错。	将每个特征向量分别代回原方程检验。

---

12. Summary | 总结

Eigenvalues and eigenvectors give deep insight into a matrix’s structure. To find them, solve the characteristic equation, then solve a linear system for each eigenvalue. Remember the key properties: trace, determinant, and diagonalization. Mastering this topic prepares you for advanced studies in engineering, economics, and natural sciences.

特征值与特征向量让我们深刻理解矩阵的结构。求特征值时先解特征方程，再对每个特征值解线性方程组。记住关键性质：迹、行列式和对角化。掌握这一主题，将为你学习工程、经济以及自然科学中的进阶内容打下坚实基础。

---

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）