n维常微分方程组下阿贝尔定理的严谨推导与正式表述问询

Great question—Abel's theorem for linear homogeneous ODE systems is a key result that connects the Wronskian of solutions to the trace of the coefficient matrix, and Boyce & DiPrima's informal mention in their text hints at its importance. Let's break down the formal statement and walk through a rigorous derivation using your notation.

Background & Setup #

First, let's formalize the system we're working with:
We have an n-dimensional linear homogeneous system of ordinary differential equations:
$$\dot{\mathbf{x}}(t) = P(t)\mathbf{x}(t)$$
Where:

• $\mathbf{x}(t) = (x_1(t), x_2(t), \dots, x_n(t))^T$ is the state vector in $\mathbb{R}^n$ (or $\mathbb{C}^n$),
• $P(t) = [p_{ij}(t)]$ is an $n \times n$ matrix with entries that are continuous functions on some interval $I \subseteq \mathbb{R}$,
• $\dot{\mathbf{x}}(t)$ denotes the component-wise derivative of $\mathbf{x}(t)$.

A fundamental matrix $\Phi(t)$ for this system is an $n \times n$ matrix whose columns are linearly independent solutions to the system. In other words:
$$\Phi(t) = \left[ \mathbf{x}_1(t), \mathbf{x}_2(t), \dots, \mathbf{x}_n(t) \right]$$
where each $\mathbf{x}_i(t)$ satisfies $\dot{\mathbf{x}}_i(t) = P(t)\mathbf{x}_i(t)$, and the columns are linearly independent on $I$. The determinant of $\Phi(t)$, denoted $\det(\Phi(t))$, is called the Wronskian of the system.

Formal Statement of Abel's Theorem #

For the linear homogeneous system $\dot{\mathbf{x}} = P(t)\mathbf{x}$, the Wronskian of any fundamental matrix satisfies:
$$\det(\Phi(t)) = \det(\Phi(t_0)) \cdot \exp\left( \int_{t_0}^t \text{tr}(P(s)) ds \right)$$
for all $t, t_0 \in I$. Here, $\text{tr}(P(s)) = p_{11}(s) + p_{22}(s) + \dots + p_{nn}(s)$ is the trace of the matrix $P(s)$ (the sum of its diagonal entries).

Key Takeaways #

• The Wronskian is either identically zero on $I$ (if it's zero at any point $t_0 \in I$) or never zero on $I$ (if it's non-zero at any $t_0$). This means linear independence of solutions is preserved across the entire interval—you only need to check it at one point!
• The evolution of the Wronskian depends solely on the trace of $P(t)$, not the full matrix. This is a surprising simplification.

Rigorous Derivation #

To prove the theorem, we'll use a key property of determinants: the derivative of a determinant of a differentiable matrix. Let's walk through this step by step.

Step 1: Derivative of the Determinant #

For any $n \times n$ matrix $A(t)$ with differentiable entries, the derivative of $\det(A(t))$ is:
$$\frac{d}{dt} \det(A(t)) = \sum_{i=1}^n \det(A_i(t))$$
where $A_i(t)$ is the matrix formed by replacing the $i$-th column of $A(t)$ with its derivative $\dot{A}_i(t)$.

Step 2: Apply to the Fundamental Matrix #

For our fundamental matrix $\Phi(t)$, each column $\mathbf{x}_i(t)$ is a solution to the system, so $\dot{\mathbf{x}}_i(t) = P(t)\mathbf{x}_i(t)$. This means the derivative of $\Phi(t)$ is:
$$\dot{\Phi}(t) = P(t)\Phi(t)$$
(because the derivative of the matrix is the matrix of column derivatives, and each column derivative is $P(t)\mathbf{x}_i$).

Step 3: Compute the Wronskian's Derivative #

Let $W(t) = \det(\Phi(t))$. Using the determinant derivative rule:
$$\dot{W}(t) = \sum_{i=1}^n \det(\Phi_i(t))$$
where $\Phi_i(t)$ has its $i$-th column replaced by $\dot{\mathbf{x}}_i(t) = P(t)\mathbf{x}_i(t)$.

Now, $P(t)\mathbf{x}i(t)$ is a linear combination of the columns of $\Phi(t)$:
$$P(t)\mathbf{x}i(t) = \sum{j=1}^n p{ji}(t)\mathbf{x}j(t)$$
So the $i$-th column of $\Phi_i(t)$ is this linear combination. Using linearity of determinants in columns, we can split $\det(\Phi_i(t))$ into:
$$\det(\Phi_i(t)) = \sum{j=1}^n p_{ji}(t) \det\left( [\mathbf{x}_1, \dots, \mathbf{x}_j, \dots, \mathbf{x}_n] \right)$$

But for all $j \neq i$, the matrix in the determinant has two identical columns (the $j$-th and $i$-th), so those determinants are zero. Only the term where $j = i$ remains:
$$\det(\Phi_i(t)) = p_{ii}(t) \cdot \det(\Phi(t)) = p_{ii}(t) W(t)$$

Step 4: Sum to Get the ODE for $W(t)$ #

Summing over all $i$:
$$\dot{W}(t) = \sum_{i=1}^n p_{ii}(t) W(t) = \text{tr}(P(t)) W(t)$$

This is a first-order linear ODE for $W(t)$. The solution to this equation is found using an integrating factor or separation of variables:
$$W(t) = W(t_0) \exp\left( \int_{t_0}^t \text{tr}(P(s)) ds \right)$$

Which is exactly the statement of Abel's theorem.

Connection to Boyce & DiPrima's Informal Mention #

Boyce & DiPrima often frame this theorem in the context of verifying linear independence of solutions—since the Wronskian either stays non-zero everywhere or is zero everywhere, you can just check it at a single point. Their informal version likely skips some of the determinant calculus details to focus on this practical takeaway, but the full derivation above fills in those gaps.

内容的提问来源于stack exchange，提问作者Eduardo Montesuma