📘 Lecture 40 — Matrices and Systems of Linear First-Order Equations
📖 Overview: This lecture introduces the matrix representation of systems of linear first-order differential equations, a powerful tool for simplifying complex systems into compact matrix equations. It covers solution vectors, initial-value problems, existence theorems, superposition principles, and concepts of linear dependence and independence, which are foundational for solving such systems efficiently.
🗂️ Topics Covered
The lecture begins by showing how to write any system of linear first-order ODEs in matrix form using coefficient matrix A and vector F. It then defines solution vectors and verifies solutions through examples. The initial-value problem is formally introduced with an existence and uniqueness theorem. The superposition principle is explained and demonstrated for homogeneous systems. Finally, the concepts of linear dependence and independence of solution vectors are defined with illustrative examples.
📝 Lecture Summary
40.1 Matrix form of a system
A system of n linear first-order differential equations can be written compactly using matrices. Let X be the column vector of unknown functions, A(t) be the n×n coefficient matrix, and F(t) be the column vector of non-homogeneous terms. Then the system becomes:
$$\frac{d\mathbf{X}}{dt} = \mathbf{A}(t)\mathbf{X} + \mathbf{F}(t)$$
For a homogeneous system where F(t) = 0, this simplifies to:
$$\frac{d\mathbf{X}}{dt} = \mathbf{A}(t)\mathbf{X}$$
Alternatively, we can write these as $\mathbf{X}' = \mathbf{A}\mathbf{X} + \mathbf{F}$ and $\mathbf{X}' = \mathbf{A}\mathbf{X}$ respectively.
🔑 Definition — Solution vector: On an interval I, a solution vector of a homogeneous system $\frac{d\mathbf{X}}{dt} = \mathbf{A}\mathbf{X}$ is any column matrix $\mathbf{X} = \begin{pmatrix} x_1(t) \ x_2(t) \ \vdots \ x_n(t) \end{pmatrix}$ whose entries are differentiable functions satisfying each equation of the system on I.
📌 Example 1: Write the non-homogeneous system $\frac{dx}{dt} = -2x + 5y + e^t - 2t$, $\frac{dy}{dt} = 4x - 3y + 10t$ in matrix form.
Let $\mathbf{X} = \begin{pmatrix} x \ y \end{pmatrix}$. Then:
$$\frac{d\mathbf{X}}{dt} = \begin{pmatrix} -2 & 5 \ 4 & -3 \end{pmatrix}\mathbf{X} + \begin{pmatrix} e^t - 2t \ 10t \end{pmatrix}$$
or $\mathbf{X}' = \begin{pmatrix} -2 & 5 \ 4 & -3 \end{pmatrix}\mathbf{X} + \begin{pmatrix} 1 \ 0 \end{pmatrix}e^t + \begin{pmatrix} -2 \ 10 \end{pmatrix}t$
📌 Example 2: Verify that $\mathbf{X}_1 = \begin{pmatrix} 1 \ -1 \end{pmatrix}e^{-2t} = \begin{pmatrix} e^{-2t} \ -e^{-2t} \end{pmatrix}$ and $\mathbf{X}_2 = \begin{pmatrix} 3 \ 5 \end{pmatrix}e^{6t} = \begin{pmatrix} 3e^{6t} \ 5e^{6t} \end{pmatrix}$ are solutions of $\mathbf{X}' = \begin{pmatrix} 1 & 3 \ 5 & 3 \end{pmatrix}\mathbf{X}$ on $(-\infty, \infty)$.
For $\mathbf{X}_1$: $\mathbf{X}_1' = \begin{pmatrix} -2e^{-2t} \ 2e^{-2t} \end{pmatrix}$ and $\mathbf{A}\mathbf{X}_1 = \begin{pmatrix} 1 & 3 \ 5 & 3 \end{pmatrix}\begin{pmatrix} e^{-2t} \ -e^{-2t} \end{pmatrix} = \begin{pmatrix} e^{-2t} - 3e^{-2t} \ 5e^{-2t} - 3e^{-2t} \end{pmatrix} = \begin{pmatrix} -2e^{-2t} \ 2e^{-2t} \end{pmatrix} = \mathbf{X}_1'$ ✓
For $\mathbf{X}_2$: $\mathbf{X}_2' = \begin{pmatrix} 18e^{6t} \ 30e^{6t} \end{pmatrix}$ and $\mathbf{A}\mathbf{X}_2 = \begin{pmatrix} 1 & 3 \ 5 & 3 \end{pmatrix}\begin{pmatrix} 3e^{6t} \ 5e^{6t} \end{pmatrix} = \begin{pmatrix} 3e^{6t} + 15e^{6t} \ 15e^{6t} + 15e^{6t} \end{pmatrix} = \begin{pmatrix} 18e^{6t} \ 30e^{6t} \end{pmatrix} = \mathbf{X}_2'$ ✓
40.2 Initial-Value Problem
Let $t_0$ be a point in interval I, and let $\mathbf{X}(t_0) = \begin{pmatrix} x_1(t_0) \ x_2(t_0) \ \vdots \ x_n(t_0) \end{pmatrix}$ and $\mathbf{X}_0 = \begin{pmatrix} \gamma_1 \ \gamma_2 \ \vdots \ \gamma_n \end{pmatrix}$ where $\gamma_i$ are given constants.
🔑 Definition — Initial-value problem: The problem of solving $\frac{d\mathbf{X}}{dt} = \mathbf{A}(t)\mathbf{X} + \mathbf{F}(t)$ subject to the initial condition $\mathbf{X}(t_0) = \mathbf{X}_0$ is called an initial-value problem on interval I.
40.3 Theorem: Existence of a Unique Solution
If the entries of matrices A(t) and F(t) in $\frac{d\mathbf{X}}{dt} = \mathbf{A}(t)\mathbf{X} + \mathbf{F}(t)$ are continuous functions on a common interval I that contains $t_0$, then there exists a unique solution of the initial-value problem on interval I.
💡 Why this matters: This theorem guarantees that if the coefficient functions are well-behaved (continuous), a unique solution exists for any given initial conditions, justifying our search for solutions.
40.4 Superposition Principle
Suppose $\mathbf{X}_1, \mathbf{X}_2, \ldots, \mathbf{X}_k$ are solution vectors of the homogeneous system $\frac{d\mathbf{X}}{dt} = \mathbf{A}(t)\mathbf{X}$ on interval I. Then the linear combination $\mathbf{X} = c_1\mathbf{X}_1 + c_2\mathbf{X}_2 + \cdots + c_k\mathbf{X}_k$ where $c_i$ are arbitrary constants is also a solution on I.
An immediate consequence: a constant multiple of any solution vector is also a solution.
📌 Example 3: For the system $\mathbf{X}' = \begin{pmatrix} 1 & 0 & 1 \ 1 & 1 & 0 \ -2 & 0 & -1 \end{pmatrix}\mathbf{X}$ with solution $\mathbf{X}_1 = \begin{pmatrix} \cos t \ -\frac{1}{2}\cos t + \frac{1}{2}\sin t \ -\cos t - \sin t \end{pmatrix}$, verify that $c_1\mathbf{X}_1$ is also a solution.
$\frac{d}{dt}(c_1\mathbf{X}_1) = c_1\begin{pmatrix} -\sin t \ \frac{1}{2}\cos t + \frac{1}{2}\sin t \ -\cos t + \sin t \end{pmatrix}$
$\mathbf{A}(c_1\mathbf{X}_1) = c_1\begin{pmatrix} 1 & 0 & 1 \ 1 & 1 & 0 \ -2 & 0 & -1 \end{pmatrix}\begin{pmatrix} \cos t \ -\frac{1}{2}\cos t + \frac{1}{2}\sin t \ -\cos t - \sin t \end{pmatrix} = c_1\begin{pmatrix} -\sin t \ \frac{1}{2}\cos t + \frac{1}{2}\sin t \ -\cos t + \sin t \end{pmatrix}$
Thus $\frac{d}{dt}(c_1\mathbf{X}_1) = \mathbf{A}(c_1\mathbf{X}_1)$, so $c_1\mathbf{X}_1$ is a solution. ✓
📌 Example 4: For the same system, $\mathbf{X}_2 = \begin{pmatrix} 0 \ e^t \ 0 \end{pmatrix}$ is also a solution since $\mathbf{X}_2' = \begin{pmatrix} 0 \ e^t \ 0 \end{pmatrix} = \begin{pmatrix} 1 & 0 & 1 \ 1 & 1 & 0 \ -2 & 0 & -1 \end{pmatrix}\begin{pmatrix} 0 \ e^t \ 0 \end{pmatrix} = \mathbf{A}\mathbf{X}_2$. By superposition, $\mathbf{X} = c_1\mathbf{X}_1 + c_2\mathbf{X}_2 = c_1\begin{pmatrix} \cos t \ -\frac{1}{2}\cos t + \frac{1}{2}\sin t \ -\cos t - \sin t \end{pmatrix} + c_2\begin{pmatrix} 0 \ e^t \ 0 \end{pmatrix}$ is also a solution.
40.5 Linear Dependence of Solution Vectors
Let $\mathbf{X}_1, \mathbf{X}_2, \ldots, \mathbf{X}_k$ be solution vectors of $\frac{d\mathbf{X}}{dt} = \mathbf{A}\mathbf{X}$ on interval I.
🔑 Definition — Linearly dependent: The set is linearly dependent on I if there exist constants $c_1, c_2, \ldots, c_k$, not all zero, such that $\mathbf{X}(t) = c_1\mathbf{X}_1(t) + c_2\mathbf{X}_2(t) + \cdots + c_k\mathbf{X}_k(t) = 0$ for all $t \in I$.
- Any two solution vectors $\mathbf{X}_1$ and $\mathbf{X}_2$ are linearly dependent if and only if one is a constant multiple of the other.
- For $k > 2$, if the set is linearly dependent, at least one vector can be expressed as a linear combination of the remaining vectors.
40.6 Linear Independence of Solution Vectors
🔑 Definition — Linearly independent: A set of solution vectors $\mathbf{X}_1, \mathbf{X}_2, \ldots, \mathbf{X}_k$ of $\frac{d\mathbf{X}}{dt} = \mathbf{A}\mathbf{X}$ on I is linearly independent if it is not linearly dependent on I, meaning $\mathbf{X}(t) = c_1\mathbf{X}_1(t) + c_2\mathbf{X}_2(t) + \cdots + c_k\mathbf{X}_k(t) = 0$ only when each $c_i = 0$.
📌 Example 5: Consider $\mathbf{X}_1 = \begin{pmatrix} 3e^t \ e^t \end{pmatrix}$ and $\mathbf{X}_2 = \begin{pmatrix} e^{-t} \ e^{-t} \end{pmatrix}$ for the system $\mathbf{X}' = \begin{pmatrix} 2 & -3 \ 1 & -2 \end{pmatrix}\mathbf{X}$.
Check: $\mathbf{X}_1' = \begin{pmatrix} 3e^t \ e^t \end{pmatrix} = \begin{pmatrix} 2 & -3 \ 1 & -2 \end{pmatrix}\begin{pmatrix} 3e^t \ e^t \end{pmatrix}$ and $\mathbf{X}_2' = \begin{pmatrix} -e^{-t} \ -e^{-t} \end{pmatrix} = \begin{pmatrix} 2 & -3 \ 1 & -2 \end{pmatrix}\begin{pmatrix} e^{-t} \ e^{-t} \end{pmatrix}$ ✓
To check linear dependence: Suppose $c_1\mathbf{X}_1 + c_2\mathbf{X}_2 = 0$. Then $c_1\begin{pmatrix} 3e^t \ e^t \end{pmatrix} + c_2\begin{pmatrix} e^{-t} \ e^{-t} \end{pmatrix} = \begin{pmatrix} 0 \ 0 \end{pmatrix}$ for all t. This gives $3c_1e^t + c_2e^{-t} = 0$ and $c_1e^t + c_2e^{-t} = 0$. Subtracting: $2c_1e^t = 0$ for all t, so $c_1 = 0$. Then $c_2e^{-t} = 0$ for all t, so $c_2 = 0$. Thus the vectors are linearly independent.
⭐ Key Takeaways
The matrix form $\mathbf{X}' = \mathbf{A}(t)\mathbf{X} + \mathbf{F}(t)$ provides a compact way to represent any system of linear first-order differential equations. For homogeneous systems, the superposition principle allows linear combinations of solutions to form new solutions—this is the fundamental building block for constructing general solutions. The existence and uniqueness theorem guarantees a unique solution when coefficient functions are continuous. Understanding linear dependence and independence of solution vectors is critical for determining when solutions form a basis for the solution space. Finally, the initial-value problem formally connects these systems to real-world applications with specific starting conditions.
🧠 Quick Revision Questions
- How do you write a system of n linear first-order differential equations in matrix form?
- What condition guarantees a unique solution to the initial-value problem $\mathbf{X}' = \mathbf{A}(t)\mathbf{X} + \mathbf{F}(t)$, $\mathbf{X}(t_0) = \mathbf{X}_0$?
- State the superposition principle for homogeneous linear systems. Why is it important?
- When is a set of solution vectors ${\mathbf{X}_1, \mathbf{X}_2, \ldots, \mathbf{X}_k}$ said to be linearly dependent on an interval I?
- For two solution vectors $\mathbf{X}_1$ and $\mathbf{X}_2$, what condition is equivalent to their linear dependence?
📘 Lecture 42 — Homogeneous Linear Systems
📖 Overview: This lecture focuses on solving homogeneous linear systems of first-order differential equations with constant coefficients. It introduces the method of finding solutions using eigenvalues and eigenvectors, building on matrix theory from previous lectures to determine fundamental sets and general solutions.
🗂️ Topics Covered
The lecture begins by revisiting fundamental matrices and sets of solutions, then introduces the eigenvalue-eigenvector method for solving homogeneous systems. It covers the characteristic equation, distinct real eigenvalues, and demonstrates how to construct the general solution from eigenvectors and eigenvalues through detailed examples.
📝 Lecture Summary
42.1 Theorem (Wronskian and Linear Independence)
A necessary and sufficient condition for the set of solution vectors X₁, X₂, ..., Xₙ of the homogeneous system X′ = AX to be linearly independent on an interval I is that the Wronskian of these solutions is non-zero for every t ∈ I.
🔑 Definition — Wronskian for Systems: For n solution vectors of dimension n, the Wronskian is the determinant of the matrix formed by these vectors as columns. 📐 Formula: W(X₁, X₂, ..., Xₙ) = det[X₁ X₂ ... Xₙ] ≠ 0, ∀ t ∈ I 📌 Example: For X₁ = (1, -1)ᵀe⁻²ᵗ and X₂ = (3, 5)ᵀe⁶ᵗ, W = |e⁻²ᵗ 3e⁶ᵗ; -e⁻²ᵗ 5e⁶ᵗ| = 8e⁴ᵗ ≠ 0, proving linear independence.
💡 Why this matters: Unlike the Wronskian for single ODEs, this determinant involves no differentiation—it is simply the determinant of the solution matrix. If W ≠ 0 for any single t₀, then W ≠ 0 for all t in the interval.
42.2 Fundamental Set of Solution
A set {X₁, X₂, ..., Xₙ} of n solution vectors of the homogeneous system X′ = AX on an interval I is called a fundamental set of solutions if the solution vectors are linearly independent.
42.2.1 Theorem (Existence of a Fundamental Set) There always exists a fundamental set of solutions for the homogeneous system X′ = AX on any interval I.
42.3 General Solution (Homogeneous Systems)
If X₁, X₂, ..., Xₙ is a fundamental set of solutions of X′ = AX on I, then any linear combination X = c₁X₁ + c₂X₂ + ... + cₙXₙ with arbitrary constants cᵢ is the general solution of the system.
📌 Example: From Example 1, X₁ = (1, -1)ᵀe⁻²ᵗ and X₂ = (3, 5)ᵀe⁶ᵗ form a fundamental set. Their general solution is X = c₁(1, -1)ᵀe⁻²ᵗ + c₂(3, 5)ᵀe⁶ᵗ.
42.4 Non-homogeneous Systems
A system dX/dt = A(t)X + F(t) is non-homogeneous if F(t) ≠ 0 for all t. The general solution consists of a complementary function and a particular integral.
🔑 Definition — Particular Integral (Xₚ): Any vector, free of arbitrary parameters, whose entries satisfy each equation of the non-homogeneous system.
42.5 Theorem (Structure of Solutions) If X₁, X₂, ..., Xₖ are solutions of X′ = AX and Xₚ is a particular solution of X′ = AX + F(t), then X = c₁X₁ + c₂X₂ + ... + cₖXₖ + Xₚ is also a solution of the non-homogeneous system.
🔑 Definition — Complementary Function (X꜀): The general solution X꜀ = c₁X₁ + c₂X₂ + ... + cₙXₙ of the homogeneous system X′ = AX is called the complementary function of the non-homogeneous system.
42.5.2 General Solution of Non-homogeneous Systems The general solution of a non-homogeneous system is X = X꜀ + Xₚ.
📌 Example: Given Xₚ = (3t-4, -5t+6)ᵀ is a particular solution of X′ = [(1,3),(5,3)]X + (12t-11, -3)ᵀ, and X꜀ = c₁(1,-1)ᵀe⁻²ᵗ + c₂(3,5)ᵀe⁶ᵗ, the general solution is X = c₁(1,-1)ᵀe⁻²ᵗ + c₂(3,5)ᵀe⁶ᵗ + (3t-4, -5t+6)ᵀ.
42.6 Fundamental Matrix
If X₁, X₂, ..., Xₙ form a fundamental set of solutions of X′ = AX, the fundamental matrix is the n×n matrix whose columns are these solution vectors.
🔑 Definition — Fundamental Matrix Φ(t): Φ(t) = [X₁ X₂ ... Xₙ], where each Xᵢ is a column vector solution.
📐 Properties: Φ′(t) = A(t)Φ(t); det(Φ(t)) ≠ 0 (since it equals the Wronskian); Φ⁻¹(t) exists for all t in I.
📌 Example: For the system X′ = [(1,3),(5,3)]X with fundamental solutions X₁ = (e⁻²ᵗ, -e⁻²ᵗ)ᵀ and X₂ = (3e⁶ᵗ, 5e⁶ᵗ)ᵀ, the fundamental matrix is Φ(t) = [(e⁻²ᵗ, 3e⁶ᵗ); (-e⁻²ᵗ, 5e⁶ᵗ)]. The general solution can be written as X = Φ(t)C, where C = (c₁, c₂)ᵀ.
Homogeneous Linear Systems (Main Topic)
The lecture now focuses on solving homogeneous systems X′ = AX where A is a constant n×n matrix, by seeking solutions of the form X = Ke^(λt) where K is a constant vector and λ is a constant.
🔑 Derivation: Substituting X = Ke^(λt) into X′ = AX gives λKe^(λt) = AKe^(λt). Canceling e^(λt) (non-zero) yields AK = λK, or equivalently (A - λI)K = 0.
This is a system of linear algebraic equations. For a non-trivial solution K (K ≠ 0), we require: 🔑 Characteristic Equation: det(A - λI) = 0 This is an nth-degree polynomial in λ, whose roots λ are the eigenvalues of A. Corresponding vectors K are eigenvectors.
42.1.1 Case 1: Distinct Real Eigenvalues If A has n distinct real eigenvalues λ₁, λ₂, ..., λₙ with corresponding eigenvectors K₁, K₂, ..., Kₙ, the general solution on (-∞, ∞) is: X = c₁K₁e^(λ₁t) + c₂K₂e^(λ₂t) + ... + cₙKₙe^(λₙt)
📌 Example 2: Solve dx/dt = 2x + 3y, dy/dt = 2x + y.
- Matrix form: X′ = [(2,3),(2,1)]X, so A = [(2,3),(2,1)].
- Characteristic equation: det(A - λI) = |2-λ, 3; 2, 1-λ| = (2-λ)(1-λ) - 6 = λ² - 3λ - 4 = 0.
- Factor: (λ + 1)(λ - 4) = 0, so eigenvalues are λ₁ = -1, λ₂ = 4.
- Eigenvector for λ₁ = -1: Solve (A + I)K = 0 → [(3,3),(2,2)]K = 0 → 3k₁ + 3k₂ = 0 → k₁ = -k₂. Choose k₂ = 1 → K₁ = (-1, 1)ᵀ.
- Eigenvector for λ₂ = 4: Solve (A - 4I)K = 0 → [(-2,3),(2,-3)]K = 0 → -2k₁ + 3k₂ = 0 → 2k₁ = 3k₂. Choose k₂ = 2 → k₁ = 3, so K₂ = (3, 2)ᵀ.
- General solution: X = c₁(-1, 1)ᵀe⁻ᵗ + c₂(3, 2)ᵀe⁴ᵗ, or in component form: x(t) = -c₁e⁻ᵗ + 3c₂e⁴ᵗ, y(t) = c₁e⁻ᵗ + 2c₂e⁴ᵗ.
⭐ Key Takeaways
The most critical concept is the eigenvalue-eigenvector method for solving homogeneous linear systems with constant coefficients: assume X = Ke^(λt), substitute to get (A-λI)K = 0, and find λ from det(A-λI) = 0. For n distinct real eigenvalues, the general solution is a linear combination of n independent solutions, each formed by multiplying an eigenvector by e^(λt). The Wronskian of solution vectors (a simple determinant) tests linear independence. For non-homogeneous systems, the complete solution is the sum of the complementary function (homogeneous general solution) and a particular integral, while the fundamental matrix Φ(t) organizes all fundamental solutions into one matrix satisfying Φ′ = AΦ.
🧠 Quick Revision Questions
- What condition on the Wronskian guarantees that n solution vectors of X′ = AX are linearly independent?
- What is the form assumed for a solution when applying the eigenvalue-eigenvector method to X′ = AX?
- How do you find the eigenvalues of the coefficient matrix A in a homogeneous linear system?
- In the system X′ = [(2,3),(2,1)]X, what are the eigenvalues and one eigenvector for each?
- What is the general solution structure for a non-homogeneous system X′ = AX + F(t) in terms of complementary function and particular integral?
📘 Lecture 43 — Eigenvalues and Eigenvectors for Homogeneous Systems
📖 Overview: This lecture continues the study of solving homogeneous linear systems of differential equations using the eigenvalue-eigenvector method. It covers Case 1 (real and distinct eigenvalues) with a 3×3 example, and introduces Case 2 (complex eigenvalues), providing a theorem for obtaining real-valued solutions and working through multiple examples. Mastering both cases is essential for solving any homogeneous system ( \mathbf{X}' = A\mathbf{X} ).
🗂️ Topics Covered
The lecture begins by completing a 2×2 system example with real distinct eigenvalues (Example 2), then works through a 3×3 system (Example 3) using Gauss-Jordan elimination to find eigenvectors. It then formally introduces Case 2 for complex eigenvalues, stating the theorem for obtaining two linearly independent real solutions from a complex eigenvalue/eigenvector pair. Examples 4, 5, and 6 demonstrate the complete process for systems with complex eigenvalues, including finding complex eigenvectors and converting to real solutions using the real and imaginary parts of the eigenvector. The lecture concludes with a set of 10 practice exercises.
📝 Lecture Summary
Example 2 (Continued)
The lecture continues from a previous example, solving a 2×2 homogeneous system where the characteristic equation yields real and distinct roots ( \lambda = -1 ) and ( \lambda = 4 ).
For ( \lambda = -1 ): The system ( (A - \lambda I)K = 0 ) gives the equations ( 3k_1 + 3k_2 = 0 ) and ( 2k_1 + 2k_2 = 0 ). These are equivalent, representing ( k_1 + k_2 = 0 ) or ( k_1 = -k_2 ). Choosing ( k_2 = -1 ) gives ( k_1 = 1 ). Thus the first eigenvector is: [ K_1 = \begin{pmatrix} 1 \ -1 \end{pmatrix} ]
For ( \lambda = 4 ): The system ( (A - \lambda I)K = 0 ) gives ( -2k_1 + 3k_2 = 0 ) and ( 2k_1 - 3k_2 = 0 ). This represents ( 2k_1 - 3k_2 = 0 ) or ( k_1 = \frac{3}{2}k_2 ). Choosing ( k_2 = 2 ) gives ( k_1 = 3 ). Thus the second eigenvector is: [ K_2 = \begin{pmatrix} 3 \ 2 \end{pmatrix} ]
Therefore, two linearly independent solution vectors are: [ X_1 = \begin{pmatrix} 1 \ -1 \end{pmatrix} e^{-t}, \quad X_2 = \begin{pmatrix} 3 \ 2 \end{pmatrix} e^{4t} ]
The general solution of the system is ( X = c_1 X_1 + c_2 X_2 ), which expands to: [ \begin{pmatrix} x(t) \ y(t) \end{pmatrix} = \begin{pmatrix} c_1 e^{-t} + 3c_2 e^{4t} \ -c_1 e^{-t} + 2c_2 e^{4t} \end{pmatrix} ] Meaning ( x(t) = c_1 e^{-t} + 3c_2 e^{4t} ) and ( y(t) = -c_1 e^{-t} + 2c_2 e^{4t} ).
Example 3 — Solving a 3×3 Homogeneous System
Solve the system: [ \frac{dx}{dt} = -4x + y + z, \quad \frac{dy}{dt} = x + 5y - z, \quad \frac{dz}{dt} = y - 3z ]
The coefficient matrix is: [ A = \begin{pmatrix} -4 & 1 & 1 \ 1 & 5 & -1 \ 0 & 1 & -3 \end{pmatrix} ] The characteristic equation is ( \det(A - \lambda I) = 0 ). Expanding the determinant yields: [ -(\lambda + 3)(\lambda + 4)(\lambda - 5) = 0 ] Thus the eigenvalues are ( \lambda = -3, -4, 5 ).
To find eigenvectors, the system ( (A - \lambda I)K = 0 ) is solved using Gauss-Jordan elimination on the augmented matrix.
For ( \lambda = -3 ): The augmented matrix is reduced to row-echelon form, giving ( k_1 = k_3 ) and ( k_2 = 0 ). Choosing ( k_3 = 1 ) gives ( k_1 = 1 ). The eigenvector is: [ K_1 = \begin{pmatrix} 1 \ 0 \ 1 \end{pmatrix} ] The corresponding solution is ( X_1 = \begin{pmatrix} 1 \ 0 \ 1 \end{pmatrix} e^{-3t} ).
For ( \lambda = -4 ): Row reduction gives ( k_1 = 10k_3 ) and ( k_2 = -k_3 ). Choosing ( k_3 = 1 ) gives ( k_1 = 10, k_2 = -1 ). The eigenvector is: [ K_2 = \begin{pmatrix} 10 \ -1 \ 1 \end{pmatrix} ] The corresponding solution is ( X_2 = \begin{pmatrix} 10 \ -1 \ 1 \end{pmatrix} e^{-4t} ).
For ( \lambda = 5 ): Row reduction gives ( k_1 = k_3 ) and ( k_2 = 8k_3 ). Choosing ( k_3 = 1 ) gives ( k_1 = 1, k_2 = 8 ). The eigenvector is: [ K_3 = \begin{pmatrix} 1 \ 8 \ 1 \end{pmatrix} ] The corresponding solution is ( X_3 = \begin{pmatrix} 1 \ 8 \ 1 \end{pmatrix} e^{5t} ).
The general solution of the 3×3 system is: [ X = c_1 \begin{pmatrix} 1 \ 0 \ 1 \end{pmatrix} e^{-3t} + c_2 \begin{pmatrix} 10 \ -1 \ 1 \end{pmatrix} e^{-4t} + c_3 \begin{pmatrix} 1 \ 8 \ 1 \end{pmatrix} e^{5t} ]
42.1.2 Case 2 (Complex Eigenvalues)
When the coefficient matrix ( A ) has complex eigenvalues, the roots of the characteristic equation are imaginary. The process still uses the eigenvalue-eigenvector method, but the solutions will involve complex exponentials, which can be converted to real solutions using trigonometric functions.
Theorem (Solutions corresponding to complex eigenvalues)
If ( \lambda_1 = \alpha + i\beta ) is a complex eigenvalue of ( A ), and ( K_1 ) is its corresponding eigenvector, then the vectors: [ X_1 = K_1 e^{\lambda_1 t}, \quad X_2 = \overline{K_1} e^{\overline{\lambda_1} t} ] are solutions of the homogeneous system ( \mathbf{X}' = A\mathbf{X} ). (Note the conjugate relationship.)
Example 4
Solve the system: [ \frac{dx}{dt} = 6x - y, \quad \frac{dy}{dt} = 5x + 4y ] The coefficient matrix is ( A = \begin{pmatrix} 6 & -1 \ 5 & 4 \end{pmatrix} ). The characteristic equation is ( \lambda^2 - 10\lambda + 29 = 0 ). Using the quadratic formula, the eigenvalues are: [ \lambda_1 = 5 + 2i, \quad \lambda_2 = 5 - 2i ]
For ( \lambda_1 = 5 + 2i ), the system ( (A - \lambda I)K = 0 ) gives the equation ( (1 - 2i)k_1 - k_2 = 0 ), or ( k_2 = (1 - 2i)k_1 ). Choosing ( k_1 = 1 ) gives the eigenvector: [ K_1 = \begin{pmatrix} 1 \ 1 - 2i \end{pmatrix} ] The conjugate eigenvector for ( \lambda_2 ) is ( K_2 = \begin{pmatrix} 1 \ 1 + 2i \end{pmatrix} ).
Two complex solutions are ( X_1 = \begin{pmatrix} 1 \ 1 - 2i \end{pmatrix} e^{(5+2i)t} ) and ( X_2 = \begin{pmatrix} 1 \ 1 + 2i \end{pmatrix} e^{(5-2i)t} ). The general complex solution is a linear combination of these.
Theorem (Real solutions corresponding to a complex eigenvalue)
To obtain real-valued solutions, we use the real and imaginary parts of a complex eigenvector. Let ( \lambda_1 = \alpha + i\beta ) be a complex eigenvalue and ( K_1 ) its eigenvector. Define: [ B_1 = \text{Re}(K_1) = \frac{1}{2}(K_1 + \overline{K_1}), \quad B_2 = \text{Im}(K_1) = \frac{1}{2i}(-K_1 + \overline{K_1}) ] Then two linearly independent real solutions on ( (-\infty, \infty) ) are: [ X_1 = (B_1 \cos \beta t - B_2 \sin \beta t) e^{\alpha t}, \quad X_2 = (B_2 \cos \beta t + B_1 \sin \beta t) e^{\alpha t} ]
Example 5
Solve ( \mathbf{X}' = \begin{pmatrix} 2 & 8 \ -1 & -2 \end{pmatrix} \mathbf{X} ).
The characteristic equation is ( \lambda^2 + 4 = 0 ), so eigenvalues are ( \lambda_1 = 2i ) and ( \lambda_2 = -2i ) (hence ( \alpha = 0, \beta = 2 )).
For ( \lambda_1 = 2i ), solving ( (A - 2iI)K = 0 ) gives ( k_1 = -(2+2i)k_2 ). Choosing ( k_2 = -1 ) yields ( k_1 = 2+2i ). The eigenvector is: [ K_1 = \begin{pmatrix} 2+2i \ -1 \end{pmatrix} = \begin{pmatrix} 2 \ -1 \end{pmatrix} + i \begin{pmatrix} 2 \ 0 \end{pmatrix} ] Thus ( B_1 = \begin{pmatrix} 2 \ -1 \end{pmatrix} ) and ( B_2 = \begin{pmatrix} 2 \ 0 \end{pmatrix} ).
Since ( \alpha = 0 ), the real general solution is: [ X = c_1 \left[ \begin{pmatrix} 2 \ -1 \end{pmatrix} \cos 2t - \begin{pmatrix} 2 \ 0 \end{pmatrix} \sin 2t \right] + c_2 \left[ \begin{pmatrix} 2 \ 0 \end{pmatrix} \cos 2t + \begin{pmatrix} 2 \ -1 \end{pmatrix} \sin 2t \right] ] Which simplifies to: [ X = c_1 \begin{pmatrix} 2\cos 2t - 2\sin 2t \ -\cos 2t \end{pmatrix} + c_2 \begin{pmatrix} 2\cos 2t + 2\sin 2t \ -\sin 2t \end{pmatrix} ]
Example 6
Solve ( \mathbf{X}' = \begin{pmatrix} 1 & 2 \ -1/2 & 1 \end{pmatrix} \mathbf{X} ).
The characteristic equation is ( \lambda^2 - 2\lambda + 2 = 0 ), giving eigenvalues ( \lambda_1 = 1 + i ) and ( \lambda_2 = 1 - i ) (so ( \alpha = 1, \beta = 1 )).
For ( \lambda_1 = 1 + i ), solving ( (A - (1+i)I)K = 0 ) yields the eigenvector: [ K_1 = \begin{pmatrix} 2 \ i \end{pmatrix} = \begin{pmatrix} 2 \ 0 \end{pmatrix} + i \begin{pmatrix} 0 \ 1 \end{pmatrix} ] Thus ( B_1 = \begin{pmatrix} 2 \ 0 \end{pmatrix} ) and ( B_2 = \begin{pmatrix} 0 \ 1 \end{pmatrix} ).
The two linearly independent real solutions are: [ X_1 = \left[ \begin{pmatrix} 2 \ 0 \end{pmatrix} \cos t - \begin{pmatrix} 0 \ 1 \end{pmatrix} \sin t \right] e^t, \quad X_2 = \left[ \begin{pmatrix} 0 \ 1 \end{pmatrix} \cos t + \begin{pmatrix} 2 \ 0 \end{pmatrix} \sin t \right] e^t ] The general real solution is: [ X = c_1 \begin{pmatrix} 2\cos t \ -\sin t \end{pmatrix} e^t + c_2 \begin{pmatrix} 2\sin t \ \cos t \end{pmatrix} e^t ]
📐 Key Pattern: When ( \alpha = 0 ), the solutions are purely oscillatory (sine/cosine). When ( \alpha \neq 0 ), the oscillations are multiplied by an exponential growth/decay factor ( e^{\alpha t} ).
42.4 Exercise
The lecture concludes with a set of 10 practice problems for the student to solve, covering both real and complex eigenvalue cases for 2×2 and 3×3 systems.
⭐ Key Takeaways
For Case 1 (real, distinct eigenvalues), you solve ( \det(A - \lambda I) = 0 ) to find eigenvalues, then for each ( \lambda ) solve ( (A - \lambda I)K = 0 ) to find a corresponding eigenvector. For a 3×3 system, you may need Gauss-Jordan elimination to solve for the eigenvector components. For Case 2 (complex eigenvalues), after finding a complex eigenvector ( K_1 = B_1 + iB_2 ), you generate two real solutions using the formulas ( X_1 = (B_1 \cos \beta t - B_2 \sin \beta t) e^{\alpha t} ) and ( X_2 = (B_2 \cos \beta t + B_1 \sin \beta t) e^{\alpha t} ). The general solution of the homogeneous system ( \mathbf{X}' = A\mathbf{X} ) is always a linear combination of n linearly independent solution vectors, where n is the dimension of the system. For exam success, you must be able to systematically find eigenvalues, compute eigenvectors (including using row reduction for 3×3), and correctly construct both real and complex general solutions.
🧠 Quick Revision Questions
- For a 2×2 system, how do you construct the general solution after finding eigenvalues ( \lambda_1 = 2 ) and ( \lambda_2 = -3 ) and their eigenvectors ( K_1 ) and ( K_2 )?
- In Example 3 (3×3 system), what row operations were used to reduce the augmented matrix for ( \lambda = -3 ), and what was the final relationship between ( k_1, k_2, k_3 )?
- What is the formula for converting a complex eigenvalue solution into two linearly independent real solutions?
- In Example 5, why did the exponential factor ( e^{\alpha t} ) disappear from the final solution, and what does that imply about the system's behavior?
- In Example 6, what are the real and imaginary parts of the complex eigenvector ( K_1 = \begin{pmatrix} 2 \ i \end{pmatrix} ), and how are they used to form the real solutions?
📘 Lecture 44 — Non-Homogeneous System
📖 Overview: This lecture addresses how to solve non-homogeneous systems of linear first-order differential equations, where the system includes a forcing function vector F(t). It introduces two methods—Undetermined Coefficients and Variation of Parameters—providing systematic procedures to find particular solutions, which are then combined with the complementary function for the general solution.
🗂️ Topics Covered
The lecture begins by defining non-homogeneous systems and their matrix notation. It then explains the two-part method of solution: finding the complementary function (solution of the homogeneous system) and a particular solution. The method of undetermined coefficients is presented with its applicability conditions, including handling duplication of terms, followed by several worked examples. Finally, the more general method of variation of parameters is developed using the fundamental matrix.
📝 Lecture Summary
44.1 Definition
Consider the system of linear first order differential equations where the coefficients (a_{ij}) and functions (f_i) are continuous on a common interval (I). The system is said to be non-homogeneous when (f_i(t) \neq 0) for all (i = 1, 2, ..., n). Otherwise it is called a homogeneous system.
44.2 Matrix Notation
In matrix notation, the system can be written as: [ X' = AX + F(t) ] where (X) is the column vector of unknown functions, (A) is the coefficient matrix, and (F(t)) is the column vector of forcing functions.
44.3 Method of Solution
To find the general solution of a non-homogeneous system, we need to find:
- The complementary function (X_c), which is the general solution of the corresponding homogeneous system (X' = AX).
- Any particular solution (X_p) of the non-homogeneous system (X' = AX + F(t)), found by the method of undetermined coefficients or the variation of parameters.
The general solution (X) of the system is then given by the sum of the complementary function and the particular solution: [ X = X_c + X_p ]
44.4 Method of Undetermined Coefficients
44.4.1 The form of (F(t))
The method of undetermined coefficients can be applied when the entries in the matrix (F(t)) have one of the following forms: constant functions, polynomial functions, exponential functions, (\sin(\beta x), \cos(\beta x)), or finite sums and products of these functions.
44.4.2 Duplication of Terms
The assumption for the particular solution (X_p) must be based on the prior knowledge of the complementary function (X_c) to avoid duplication of terms between (X_c) and (X_p).
Example 1
Solve the system on the interval ((-\infty, \infty)): [ X' = \begin{pmatrix} -1 & 2 \ -1 & 1 \end{pmatrix} X + \begin{pmatrix} -8 \ 3 \end{pmatrix} ]
Solution: First, find (X_c) by solving the homogeneous system. The characteristic equation is (\lambda^2 + 1 = 0), giving eigenvalues (\lambda = \pm i) ((\alpha = 0, \beta = 1)). The eigenvector for (\lambda_1 = i) is (K_1 = \begin{pmatrix} 1-i \ 1 \end{pmatrix}), leading to matrices (B_1 = \begin{pmatrix} 1 \ 1 \end{pmatrix}) and (B_2 = \begin{pmatrix} -1 \ 0 \end{pmatrix}). The two linearly independent solutions are: [ X_1 = \begin{pmatrix} \cos t + \sin t \ \cos t \end{pmatrix}, \quad X_2 = \begin{pmatrix} -\cos t + \sin t \ \sin t \end{pmatrix} ] Thus, the complementary function is: [ X_c = c_1 \begin{pmatrix} \cos t + \sin t \ \cos t \end{pmatrix} + c_2 \begin{pmatrix} -\cos t + \sin t \ \sin t \end{pmatrix} ] Since (F(t)) is a constant vector, assume a constant particular solution (X_p = \begin{pmatrix} a_1 \ b_1 \end{pmatrix}). Substituting into the original system and using (X'_p = 0) gives: [ \begin{pmatrix} 0 \ 0 \end{pmatrix} = \begin{pmatrix} -a_1 + 2b_1 - 8 \ -a_1 + b_1 + 3 \end{pmatrix} ] Solving this system yields (b_1 = 11) and (a_1 = 14). Therefore, the particular solution is: [ X_p = \begin{pmatrix} 14 \ 11 \end{pmatrix} ] Hence, the general solution of the non-homogeneous system is: [ X = c_1 \begin{pmatrix} \cos t + \sin t \ \cos t \end{pmatrix} + c_2 \begin{pmatrix} -\cos t + \sin t \ \sin t \end{pmatrix} + \begin{pmatrix} 14 \ 11 \end{pmatrix} ] 💡 Why this matters: If (F(t)) were a constant vector and the coefficient matrix had an eigenvalue (\lambda = 0), then (X_c) would contain a constant vector, requiring a modified assumption for (X_p) (e.g., (X_p = \begin{pmatrix} a_2 \ b_2 \end{pmatrix} t + \begin{pmatrix} a_1 \ b_1 \end{pmatrix})).
Example 2
Solve the system: [ \frac{dx}{dt} = 6x + y + 6t, \quad \frac{dy}{dt} = 4x + 3y - 10t + 4 ]
Solution: In matrix form: (X' = \begin{pmatrix} 6 & 1 \ 4 & 3 \end{pmatrix} X + \begin{pmatrix} 6 \ -10 \end{pmatrix} t + \begin{pmatrix} 0 \ 4 \end{pmatrix}). The homogeneous system has eigenvalues (\lambda_1 = 2) and (\lambda_2 = 7), with corresponding eigenvectors (K_1 = \begin{pmatrix} 1 \ -4 \end{pmatrix}) and (K_2 = \begin{pmatrix} 1 \ 1 \end{pmatrix}). Thus, [ X_c = c_1 \begin{pmatrix} 1 \ -4 \end{pmatrix} e^{2t} + c_2 \begin{pmatrix} 1 \ 1 \end{pmatrix} e^{7t} ] Assume a particular solution of the form (X_p = \begin{pmatrix} a_2 \ b_2 \end{pmatrix} t + \begin{pmatrix} a_1 \ b_1 \end{pmatrix}). Substituting into the system and equating coefficients gives:
- From the (t) terms: (6a_2 + b_2 + 6 = 0) and (4a_2 + 3b_2 - 10 = 0), solving to (a_2 = -2, b_2 = 6).
- From the constant terms: (6a_1 + b_1 - a_2 = 0) and (4a_1 + 3b_1 - b_2 + 4 = 0), solving to (a_1 = -\frac{4}{7}, b_1 = \frac{10}{7}). Thus, the particular solution is: [ X_p = \begin{pmatrix} -2 \ 6 \end{pmatrix} t + \begin{pmatrix} -4/7 \ 10/7 \end{pmatrix} ] The general solution is: [ X = c_1 \begin{pmatrix} 1 \ -4 \end{pmatrix} e^{2t} + c_2 \begin{pmatrix} 1 \ 1 \end{pmatrix} e^{7t} + \begin{pmatrix} -2 \ 6 \end{pmatrix} t + \begin{pmatrix} -4/7 \ 10/7 \end{pmatrix} ]
Example 3
Determine the form of the particular solution vector (X_p) for: [ \frac{dx}{dt} = 5x + 3y - 2e^{-t} + 1, \quad \frac{dy}{dt} = -x + y + e^{-t} - 5t + 7 ]
Solution: In matrix form: (X' = \begin{pmatrix} 5 & 3 \ -1 & 1 \end{pmatrix} X + \begin{pmatrix} -2 \ 1 \end{pmatrix} e^{-t} + \begin{pmatrix} 0 \ -5 \end{pmatrix} t + \begin{pmatrix} 1 \ 7 \end{pmatrix}). The homogeneous system has eigenvalues (\lambda = 2, 4), with eigenvectors (K_1 = \begin{pmatrix} 1 \ -1 \end{pmatrix}) for (\lambda=2) and (K_2 = \begin{pmatrix} 3 \ -1 \end{pmatrix}) for (\lambda=4). So, [ X_c = c_1 \begin{pmatrix} 1 \ -1 \end{pmatrix} e^{2t} + c_2 \begin{pmatrix} 3 \ -1 \end{pmatrix} e^{4t} ] Since (e^{-t}) in (F(t)) does not match any eigenvalue, we assume: [ X_p = \begin{pmatrix} a_3 \ b_3 \end{pmatrix} e^{-t} + \begin{pmatrix} a_2 \ b_2 \end{pmatrix} t + \begin{pmatrix} a_1 \ b_1 \end{pmatrix} ] Note: If (e^{-t}) were replaced by (e^{2t}) (matching eigenvalue (\lambda=2)), the correct form would be: [ X_p = \begin{pmatrix} a_4 \ b_4 \end{pmatrix} t e^{2t} + \begin{pmatrix} a_3 \ b_3 \end{pmatrix} e^{2t} + \begin{pmatrix} a_2 \ b_2 \end{pmatrix} t + \begin{pmatrix} a_1 \ b_1 \end{pmatrix} ]
44.5 Variation of Parameters
Variation of parameters is a more powerful technique than the method of undetermined coefficients. It provides a systematic procedure for finding a particular solution of the non-homogeneous linear vector differential equation: [ \frac{dX}{dt} = AX + F(t) ]
Let (\phi(t)) be a fundamental matrix of the homogeneous system (X' = AX). Then the general solution of the homogeneous system is (X_c = \phi(t)C), where (C) is an arbitrary constant vector. We replace the constant vector (C) by a column matrix of functions (U(t)) so that: [ X_p = \phi(t) U(t) ] is a particular solution of the non-homogeneous system.
Differentiating using the product rule: [ X'_p = \phi(t) U'(t) + \phi'(t) U(t) ] Substituting into (X' = AX + F(t)) and using (\phi'(t) = A\phi(t)) gives: [ \phi(t) U'(t) = F(t) ] Multiplying by (\phi^{-1}(t)) (the inverse of the fundamental matrix) yields: [ U'(t) = \phi^{-1}(t) F(t) ] Integrating: [ U(t) = \int \phi^{-1}(t) F(t) dt ] Therefore, a particular solution is: [ X_p = \phi(t) \int \phi^{-1}(t) F(t) dt ] The general solution of the non-homogeneous system is then: [ X = X_c + X_p = \phi(t) C + \phi(t) \int \phi^{-1}(t) F(t) dt ]
Example
Find the general solution of the non-homogeneous system on ((-\infty, \infty)): [ X' = \begin{pmatrix} -3 & 1 \ 2 & -4 \end{pmatrix} X + \begin{pmatrix} 3t \ e^{-t} \end{pmatrix} ]
Solution: Solve the homogeneous system first. The characteristic equation is (\lambda^2 + 7\lambda + 10 = 0), giving eigenvalues (\lambda_1 = -2) and (\lambda_2 = -5). Eigenvectors are (K_1 = \begin{pmatrix} 1 \ 1 \end{pmatrix}) and (K_2 = \begin{pmatrix} 1 \ -2 \end{pmatrix}). The solution vectors are: [ X_1 = \begin{pmatrix} 1 \ 1 \end{pmatrix} e^{-2t}, \quad X_2 = \begin{pmatrix} 1 \ -2 \end{pmatrix} e^{-5t} ] The fundamental matrix is: [ \phi(t) = \begin{pmatrix} e^{-2t} & e^{-5t} \ e^{-2t} & -2e^{-5t} \end{pmatrix} ] Its inverse is: [ \phi^{-1}(t) = \begin{pmatrix} \frac{2}{3} e^{2t} & \frac{1}{3} e^{2t} \ \frac{1}{3} e^{5t} & -\frac{1}{3} e^{5t} \end{pmatrix} ] Now, compute (X_p = \phi(t) \int \phi^{-1}(t) F(t) dt): [ \phi^{-1}(t) F(t) = \begin{pmatrix} \frac{2}{3} e^{2t} & \frac{1}{3} e^{2t} \ \frac{1}{3} e^{5t} & -\frac{1}{3} e^{5t} \end{pmatrix} \begin{pmatrix} 3t \ e^{-t} \end{pmatrix} = \begin{pmatrix} 2t e^{2t} + \frac{1}{3} e^{t} \ t e^{5t} - \frac{1}{3} e^{4t} \end{pmatrix} ] Integrating: [ \int \phi^{-1}(t) F(t) dt = \begin{pmatrix} \int (2t e^{2t} + \frac{1}{3} e^{t}) dt \ \int (t e^{5t} - \frac{1}{3} e^{4t}) dt \end{pmatrix} = \begin{pmatrix} t e^{2t} - \frac{1}{2} e^{2t} + \frac{1}{3} e^{t} \ \frac{1}{5} t e^{5t} - \frac{1}{25} e^{5t} - \frac{1}{12} e^{4t} \end{pmatrix} ] Then: [ X_p = \phi(t) \begin{pmatrix} t e^{2t} - \frac{1}{2} e^{2t} + \frac{1}{3} e^{t} \ \frac{1}{5} t e^{5t} - \frac{1}{25} e^{5t} - \frac{1}{12} e^{4t} \end{pmatrix} ] Multiplying and simplifying yields: [ X_p = \begin{pmatrix} \frac{6}{5} t - \frac{27}{50} + \frac{1}{4} e^{-t} \ \frac{3}{5} t - \frac{21}{50} + \frac{1}{2} e^{-t} \end{pmatrix} ] The general solution is: [ X = X_c + X_p = c_1 \begin{pmatrix} 1 \ 1 \end{pmatrix} e^{-2t} + c_2 \begin{pmatrix} 1 \ -2 \end{pmatrix} e^{-5t} + \begin{pmatrix} \frac{6}{5} t - \frac{27}{50} + \frac{1}{4} e^{-t} \ \frac{3}{5} t - \frac{21}{50} + \frac{1}{2} e^{-t} \end{pmatrix} ]
44.6 Exercise
(The lecture text provides a list of exercises for the student to practice the methods covered, but no solution is given in the lecture.)
⭐ Key Takeaways
- The general solution of a non-homogeneous linear system is the sum of the complementary function (solution to the homogeneous system) and any particular solution. The complementary function is found by solving (X' = AX) using eigenvalues and eigenvectors. The method of undetermined coefficients is a standard technique for finding a particular solution when the forcing function (F(t)) is of a specific form (polynomial, exponential, sine/cosine, or their products), but it requires adjusting the assumed form to avoid duplication with the complementary function. Variation of parameters is a more general method that always works, provided the fundamental matrix of the homogeneous system is known; it involves integrating the product of the inverse fundamental matrix and the forcing function. The ability to correctly identify which method to use and how to handle duplication of terms is critical for exam success.
🧠 Quick Revision Questions
- What is the general form of the solution to a non-homogeneous system (X' = AX + F(t))?
- When using the method of undetermined coefficients, what condition must be satisfied by the entries of (F(t))?
- If (F(t) = \begin{pmatrix} 8 \ -3 \end{pmatrix}) and the coefficient matrix (A) has an eigenvalue (\lambda = 0), what form should the particular solution (X_p) take to avoid duplication?
- In the variation of parameters method, what is the fundamental matrix (\phi(t)) and how is it used to find a particular solution?
- In example 3, why would the assumed form for (X_p) change if the exponential term in (F(t)) matched an eigenvalue?