Non-Homogeneous Systems: Variation of Parameters
You'll learn how to solve non-homogeneous linear systems of ODEs using variation of parameters with the fundamental matrix.
The Core Problem
You've solved homogeneous systems like x' = Ax using eigenvalues and the fundamental matrix Φ(t). But what about non-homogeneous systems of the form:
x' = Ax + g(t)
where g(t) is a vector of forcing functions? Just as with second-order ODEs, variation of parameters provides a systematic way to find a particular solution when the forcing term isn't convenient for other methods.
The Method
The general solution structure is x(t) = x_h(t) + x_p(t), where x_h is the homogeneous solution and x_p is a particular solution.
Variation of parameters gives us:
x_p(t) = Φ(t) ∫ Φ⁻¹(s) g(s) ds
Here, Φ(t) is your fundamental matrix (built from linearly independent solutions of the homogeneous system), and Φ⁻¹(t) is its inverse.
How It Works
Think of the fundamental matrix as a "translator" between coordinate systems. The term Φ⁻¹(s) g(s) converts the forcing function into the "language" of the homogeneous solutions. You integrate this over time, then Φ(t) translates back to produce the particular solution.
For initial value problems x(t₀) = x₀, use:
x_p(t) = Φ(t) ∫_{t₀}^t Φ⁻¹(s) g(s) ds
Then add to the homogeneous solution adjusted for initial conditions.
Why It Matters
Unlike undetermined coefficients, variation of parameters handles any forcing function g(t)—even discontinuous or non-standard ones. The price is more computation (matrix inversion and integration), but you gain complete generality.
Key Takeaway: Variation of parameters for systems uses the fundamental matrix to systematically construct particular solutions as x_p = Φ(t) ∫ Φ⁻¹(s) g(s) ds, handling any non-homogeneous forcing term.