Variation of Parameters: A Powerful Tool for Solving Non‑Homogeneous Linear Differential Equations
When a differential equation contains a non‑zero forcing term, the usual approach of guessing a particular solution can become cumbersome or even impossible. Now, variation of parameters offers a systematic, algebraic method that works for any linear differential equation with variable coefficients, provided we can solve the corresponding homogeneous equation. This article walks through the theory, step‑by‑step derivation, practical examples, and common pitfalls, ensuring you can confidently apply the technique to a wide range of problems.
Introduction
Consider a second‑order linear differential equation in the form
[ y'' + p(x),y' + q(x),y = g(x), ]
where (p(x)), (q(x)), and (g(x)) are continuous functions on an interval (I). The left‑hand side represents the homogeneous part, while (g(x)) is the non‑homogeneous forcing term. The general solution is the sum of the homogeneous solution (y_h) and a particular solution (y_p):
[ y(x) = y_h(x) + y_p(x). ]
The method of variation of parameters constructs (y_p) by allowing the constants in (y_h) to vary with (x). This seemingly simple idea unlocks a powerful algorithm that can handle arbitrary (g(x)).
Step 1: Solve the Homogeneous Equation
First, find two linearly independent solutions (y_1(x)) and (y_2(x)) to
[ y'' + p(x),y' + q(x),y = 0. ]
These form the basis of the solution space. The homogeneous solution is
[ y_h(x) = C_1,y_1(x) + C_2,y_2(x), ]
where (C_1) and (C_2) are constants determined by initial or boundary conditions.
Example
For (y'' - 3y' + 2y = 0), the characteristic equation (r^2 - 3r + 2 = 0) yields roots (r=1) and (r=2). Thus
[ y_1(x) = e^{x}, \qquad y_2(x) = e^{2x}. ]
Step 2: Assume a Particular Solution with Variable Coefficients
Replace the constants (C_1) and (C_2) by functions (u_1(x)) and (u_2(x)):
[ y_p(x) = u_1(x),y_1(x) + u_2(x),y_2(x). ]
The goal is to determine (u_1) and (u_2) such that (y_p) satisfies the original non‑homogeneous equation Worth knowing..
Step 3: Impose a Convenient Constraint
Direct substitution leads to second‑order derivatives of (u_1) and (u_2), complicating the algebra. To simplify, impose the condition
[ u_1'(x),y_1(x) + u_2'(x),y_2(x) = 0. ]
This eliminates the second‑order derivatives of (u_1) and (u_2) from the expression for (y_p''), leaving only first‑order terms Easy to understand, harder to ignore..
Step 4: Derive the System for (u_1') and (u_2')
Differentiate (y_p) once and twice, substitute into the original equation, and use the constraint to obtain:
[ \begin{cases} u_1'(x),y_1'(x) + u_2'(x),y_2'(x) = g(x),\[4pt] u_1'(x),y_1(x) + u_2'(x),y_2(x) = 0. \end{cases} ]
Basically a linear system for the unknowns (u_1') and (u_2'). Solving it yields:
[ u_1'(x) = -,\frac{y_2(x),g(x)}{W(y_1,y_2)(x)}, \qquad u_2'(x) = \frac{y_1(x),g(x)}{W(y_1,y_2)(x)}, ]
where (W(y_1,y_2)(x)) is the Wronskian:
[ W(y_1,y_2)(x) = y_1(x),y_2'(x) - y_1'(x),y_2(x). ]
The Wronskian never vanishes on (I) because (y_1) and (y_2) are linearly independent It's one of those things that adds up..
Step 5: Integrate to Find (u_1) and (u_2)
Integrate the expressions for (u_1') and (u_2'):
[ u_1(x) = -\int \frac{y_2(x),g(x)}{W(y_1,y_2)(x)},dx, \qquad u_2(x) = \int \frac{y_1(x),g(x)}{W(y_1,y_2)(x)},dx. ]
These integrals may be elementary or require special functions, depending on (g(x)). Once (u_1) and (u_2) are known, substitute back into (y_p):
[ y_p(x) = u_1(x),y_1(x) + u_2(x),y_2(x). ]
Finally, the general solution is
[ y(x) = C_1,y_1(x) + C_2,y_2(x) + y_p(x). ]
Worked Example 1: Polynomial Forcing
Solve
[ y'' - 3y' + 2y = x. ]
Homogeneous part: As before, (y_1=e^{x}), (y_2=e^{2x}).
Wronskian: (W = e^{x}\cdot 2e^{2x} - e^{x}\cdot e^{2x} = e^{3x}).
Compute (u_1') and (u_2'):
[ u_1' = -\frac{e^{2x},x}{e^{3x}} = -x,e^{-x}, \qquad u_2' = \frac{e^{x},x}{e^{3x}} = x,e^{-2x}. ]
Integrate:
[ u_1 = -\int x,e^{-x},dx = -\bigl(-x,e^{-x} - e^{-x}\bigr) = x,e^{-x} + e^{-x}, ] [ u_2 = \int x,e^{-2x},dx = -\frac{1}{2}x,e^{-2x} - \frac{1}{4}e^{-2x}. ]
Construct (y_p):
[ y_p = u_1,y_1 + u_2,y_2 = (x,e^{-x} + e^{-x})e^{x} + \left(-\frac{1}{2}x,e^{-2x} - \frac{1}{4}e^{-2x}\right)e^{2x} = x + 1 - \frac{1}{2}x - \frac{1}{4} = \frac{1}{2}x + \frac{3}{4}. ]
Thus the general solution:
[ y(x) = C_1,e^{x} + C_2,e^{2x} + \frac{1}{2}x + \frac{3}{4}. ]
Worked Example 2: Trigonometric Forcing
Solve
[ y'' + y = \sin(2x). ]
Homogeneous part: (y_1=\cos x), (y_2=\sin x).
Wronskian: (W = \cos x \cdot \cos x - (-\sin x)\cdot \sin x = \cos^2x + \sin^2x = 1).
Compute (u_1') and (u_2'):
[ u_1' = -y_2,g = -\sin x \cdot \sin(2x) = -\sin x \cdot 2\sin x\cos x = -2\sin^2x\cos x, ] [ u_2' = y_1,g = \cos x \cdot \sin(2x) = \cos x \cdot 2\sin x\cos x = 2\sin x\cos^2x. ]
Integrate:
[ u_1 = -\int 2\sin^2x\cos x,dx = -\int 2(1-\cos^2x)\cos x,dx = -\int 2\cos x,dx + \int 2\cos^3x,dx. ]
Using substitution (t=\sin x), (dt=\cos x,dx):
[ -\int 2\cos x,dx = -2\sin x, ] [ \int 2\cos^3x,dx = \int 2(1-\sin^2x),d(\sin x) = 2\sin x - \frac{2}{3}\sin^3x. ]
Thus (u_1 = -2\sin x + 2\sin x - \frac{2}{3}\sin^3x = -\frac{2}{3}\sin^3x) Surprisingly effective..
Similarly,
[ u_2 = \int 2\sin x\cos^2x,dx = \int 2\sin x(1-\sin^2x),dx = -2\cos x + \frac{2}{3}\cos^3x. ]
Now
[ y_p = u_1,y_1 + u_2,y_2 = \left(-\frac{2}{3}\sin^3x\right)\cos x + \left(-2\cos x + \frac{2}{3}\cos^3x\right)\sin x. ]
Simplify using (\sin^2x + \cos^2x = 1):
[ y_p = -\frac{2}{3}\sin^3x\cos x - 2\cos x\sin x + \frac{2}{3}\cos^3x\sin x = -\frac{2}{3}\sin x\cos x(\sin^2x - \cos^2x) - 2\sin x\cos x. ]
Recognizing (\sin^2x - \cos^2x = -\cos(2x)) and (\sin x\cos x = \frac{1}{2}\sin(2x)), we obtain
[ y_p = \frac{1}{3}\sin(2x)\cos(2x) - \sin(2x). ]
A cleaner form after algebraic manipulation is
[ y_p = \frac{1}{3}\sin(2x). ]
Hence the general solution:
[ y(x) = C_1\cos x + C_2\sin x + \frac{1}{3}\sin(2x). ]
Scientific Explanation
The core idea behind variation of parameters is that the homogeneous solution space is a vector space. Any particular solution can be expressed as a linear combination of the basis functions (y_1) and (y_2), but with coefficients that vary with (x). Think about it: by allowing the coefficients to change, we introduce enough flexibility to satisfy the non‑homogeneous term. In practice, the imposed constraint (u_1'y_1 + u_2'y_2 = 0) ensures that the second‑order derivatives of (u_1) and (u_2) cancel out, leaving a solvable first‑order system. The Wronskian encapsulates the linear independence of the basis functions and appears naturally in the denominators of (u_1') and (u_2').
Frequently Asked Questions
| Question | Answer |
|---|---|
| **When does variation of parameters fail?Here's the thing — | |
| **What if (g(x)) is a product of a function and a solution of the homogeneous equation? Here's the thing — ** | No. Still, the integrals may be elementary or require special functions. ** |
| **Can we use it for non‑linear equations?For simple forcing terms (polynomials, exponentials, sines, cosines) undetermined coefficients is quicker. Now, ** | Only once for each (u_i). In real terms, |
| **Do we need to integrate twice? On top of that, e. Practically speaking, variation of parameters is more general but can involve messy integrals. , perturbation methods) are needed. Think about it: | |
| Is it always easier than undetermined coefficients? , the homogeneous solutions are not independent), the method cannot be applied. So the method relies on linearity; for non‑linear equations, other techniques (e. g. | The method still works, but the integrals may simplify due to cancellation. |
Conclusion
Variation of parameters transforms the challenge of finding a particular solution into a systematic procedure that hinges on the homogeneous solutions and the Wronskian. While the algebra can be involved, the method’s generality makes it indispensable for tackling a wide array of linear differential equations, especially when the forcing term does not fit the patterns handled by undetermined coefficients. Mastery of this technique equips you with a versatile tool for both theoretical investigations and practical applications across physics, engineering, and applied mathematics.
