Higher-Order Linear Differential Equations

TL;DR

Higher-order linear differential equations involve derivatives beyond the first, combining them linearly. You solve them by finding the roots of a characteristic polynomial, which dictate the form of the homogeneous solution. Non-homogeneous equations require an additional particular solution found through methods like undetermined coefficients or variation of parameters.

1. The Mental Model

Think of these equations as a puzzle where you're trying to find a function whose derivatives, when combined in a specific way, equal another function. The "order" just tells you the highest derivative involved, and "linear" means the function and its derivatives aren't multiplied by each other or raised to powers.

2. The Core Material

Higher-order linear differential equations have the general form:

$a_n \frac{d^n y}{dx^n} + a_{n-1} \frac{d^{n-1} y}{dx^{n-1}} + \dots + a_1 \frac{dy}{dx} + a_0 y = g(x)$

where $a_n, \dots, a_0$ are constants or functions of $x$, and $g(x)$ is a given function. If $g(x)=0$, it's a homogeneous equation; otherwise, it's non-homogeneous.

Solving Homogeneous Equations (Constant Coefficients)

Photo by Monstera Production on Pexels

For homogeneous equations with constant coefficients ($a_i$ are constants), you assume a solution of the form $y = e^{mx}$. Substituting this into the equation turns it into a characteristic equation (a polynomial equation).

$a_n m^n + a_{n-1} m^{n-1} + \dots + a_1 m + a_0 = 0$

The roots of this polynomial determine the form of the general solution $y_c$.

Case 1: Distinct Real Roots

If you have distinct real roots $m_1, m_2, \dots, m_n$, the general solution is:
$y_c = C_1 e^{m_1 x} + C_2 e^{m_2 x} + \dots + C_n e^{m_n x}$

Case 2: Repeated Real Roots

If a real root $m$ has multiplicity $k$ (appears $k$ times), then its contribution to the solution is:
$C_1 e^{mx} + C_2 x e^{mx} + \dots + C_k x^{k-1} e^{mx}$

Case 3: Complex Conjugate Roots

If you have a pair of complex conjugate roots $m = \alpha \pm i\beta$, their contribution to the solution is:
$e^{\alpha x}(C_1 \cos(\beta x) + C_2 \sin(\beta x))$
If these complex roots are repeated, say $k$ times, then:
$e^{\alpha x}[(C_1 \cos(\beta x) + C_2 \sin(\beta x)) + x(C_3 \cos(\beta x) + C_4 \sin(\beta x)) + \dots + x^{k-1}(C_{2k-1} \cos(\beta x) + C_{2k} \sin(\beta x))]$

The general solution for the homogeneous part is the sum of all these contributions.

Solving Non-Homogeneous Equations

Photo by Monstera Production on Pexels

For non-homogeneous equations, the general solution is $y = y_c + y_p$, where $y_c$ is the complementary (homogeneous) solution you just found, and $y_p$ is a particular solution.

Method of Undetermined Coefficients

This method works well when $g(x)$ is a polynomial, exponential, sine, cosine, or a combination of these. You make an "educated guess" for the form of $y_p$ based on $g(x)$ and its derivatives.

graph TD
    A["Identify Equation Type"] --> B{"Is it Homogeneous?"}
    B -- Yes --> C["Form Characteristic Equation"]
    B -- No --> D["Find Homogeneous Solution (y_c)"]

    C --> E["Find Roots (m)"]
    E -- Distinct Real --> F["y_c = C₁e^(m₁x) + ..."]
    E -- Repeated Real --> G["y_c = C₁e^(mx) + C₂xe^(mx) + ..."]
    E -- Complex Conjugate --> H["y_c = e^(αx)(C₁cos(βx) + C₂sin(βx))"]

    D --> I["Determine g(x)"]
    I -- Simple forms (polynomial, exponential, sin/cos) --> J["Use Undetermined Coefficients"]
    J --> K["Guess Form of y_p (watch for duplication with y_c)"]
    K --> L["Substitute y_p into Original Equation"]
    L --> M["Solve for Unknown Coefficients in y_p"]

    I -- More complex g(x) or non-constant coefficients --> N["Use Variation of Parameters"]
    N --> P["Calculate Wronskian and Integrals"]
    P --> Q["Form y_p"]

    (F, G, H) --> R["Final Solution: y = y_c (if homogeneous)"]
    (M, Q) --> S["Final Solution: y = y_c + y_p (if non-homogeneous)"]

Important Note for Undetermined Coefficients: If your initial guess for $y_p$ contains terms that are already part of $y_c$, you must multiply your guess by $x^k$, where $k$ is the smallest positive integer that eliminates the duplication.

Variation of Parameters

This method is more general and works for any $g(x)$, but it's often more computationally intensive. If $y_1, y_2, \dots, y_n$ are linearly independent solutions to the homogeneous equation, then the particular solution is:

$y_p = u_1(x) y_1(x) + u_2(x) y_2(x) + \dots + u_n(x) y_n(x)$

where $u_i'(x)$ are found by solving a system of equations involving the Wronskian. For a second-order equation $y'' + P(x)y' + Q(x)y = f(x)$ (after dividing by $a_n$), with homogeneous solutions $y_1, y_2$:

$u_1'(x) = -\frac{y_2(x) f(x)}{W(y_1, y_2)}$
$u_2'(x) = \frac{y_1(x) f(x)}{W(y_1, y_2)}$

where $W(y_1, y_2) = y_1 y_2' - y_2 y_1'$ is the Wronskian. You then integrate $u_1'$ and $u_2'$ to find $u_1$ and $u_2$.

3. Worked Example

Let's solve the non-homogeneous equation: $y'' - 3y' + 2y = 4e^{3x}$

Step 1: Solve the homogeneous equation ($y'' - 3y' + 2y = 0$)
The characteristic equation is $m^2 - 3m + 2 = 0$.
Factoring gives $(m-1)(m-2) = 0$, so the roots are $m_1 = 1$ and $m_2 = 2$.
The homogeneous solution is $y_c = C_1 e^x + C_2 e^{2x}$.

Step 2: Find a particular solution ($y_p$)
Since $g(x) = 4e^{3x}$, we guess $y_p = A e^{3x}$.
Calculate its derivatives:
$y_p' = 3A e^{3x}$
$y_p'' = 9A e^{3x}$

Substitute these into the original non-homogeneous equation:
$9A e^{3x} - 3(3A e^{3x}) + 2(A e^{3x}) = 4e^{3x}$
$9A e^{3x} - 9A e^{3x} + 2A e^{3x} = 4e^{3x}$
$2A e^{3x} = 4e^{3x}$

Equating coefficients of $e^{3x}$:
$2A = 4 \Rightarrow A = 2$.

So, the particular solution is $y_p = 2e^{3x}$.

Step 3: Form the general solution
The general solution is $y = y_c + y_p$.
$y = C_1 e^x + C_2 e^{2x} + 2e^{3x}$.

4. Key Takeaways

• Higher-order linear differential equations are solved by combining homogeneous and particular solutions.
• The homogeneous solution depends on the roots of the characteristic equation.
• Real distinct roots give exponential terms $e^{mx}$.
• Repeated real roots involve $x^{k-1}e^{mx}$ terms.
• Complex conjugate roots give sinusoidal terms $e^{\alpha x}(C_1 \cos(\beta x) + C_2 \sin(\beta x))$.
• Undetermined coefficients is efficient for specific forms of $g(x)$ but requires careful guessing.
• Always check for overlap between your $y_p$ guess and $y_c$ terms when using undetermined coefficients, and multiply by $x^k$ if needed.
• Variation of parameters is more general but more complex, suitable when undetermined coefficients isn't applicable.

Common Mistakes to Avoid

Photo by KATRIN BOLOVTSOVA on Pexels

• Not forming the characteristic equation correctly from the homogeneous part.
• Incorrectly handling repeated roots or complex roots when writing $y_c$.
• Forgetting to multiply your $y_p$ guess by $x^k$ when there's duplication with $y_c$.
• Algebraic errors when substituting $y_p$ derivatives back into the non-homogeneous equation.
• Not dividing the entire non-homogeneous equation by $a_n$ before applying variation of parameters (if $a_n$ is not 1).

5. Now Try It

Solve the following differential equation: $y''' - y'' = 3$.
What to do: First, find the homogeneous solution ($y_c$). Then, use the method of undetermined coefficients to find the particular solution ($y_p$). Finally, write the complete general solution.
What success looks like: You should arrive at a general solution of the form $y = C_1 + C_2 x + C_3 e^x - \frac{3}{2}x^2$.