Linear equations with constant coefficients

A differential equation ties a function to its derivatives. The linear ones with constant coefficients, a·y″ + b·y′ + c·y = f(x), describe springs, RLC and RC circuits and decay: they are the first met at school and university. The general solution is the sum of the complementary function, solving f = 0, and any particular integral.

For the homogeneous equation try y = e^(rx): substituting leaves the characteristic equation a·r² + b·r + c = 0. Two real roots give two exponentials, a double root r gives (C₁ + C₂x)·e^(rx), two complex roots α ± iβ give e^(αx)(C₁ cos βx + C₂ sin βx), an oscillation that decays or grows.

The particular integral is found by undetermined coefficients: try a function of the same form as the right-hand side with unknown coefficients. For y″ + 3y′ + 2y = 4x + e^(−x), try Ax + B and, since e^(−x) already solves the homogeneous equation, K·x·e^(−x): the result is 2x − 3 + x·e^(−x). The initial conditions then fix the constants.

Common mistakes

  • Forgetting to multiply by x at resonance: the trial Ke^(−x) would give 0 = e^(−x), impossible.
  • Applying the initial conditions to the complementary function alone instead of the full solution.
  • Writing complex roots as two real exponentials: sine and cosine of βx are needed.

Frequently asked questions

What is an initial value problem?

The equation together with initial conditions, the value of y and, for second order, of y′ at a point x₀. They fix the constants and make the solution unique.

What if the right-hand side is different, say ln x?

Undetermined coefficients works only for polynomials, exponentials, sines, cosines and their products. Other cases need variation of parameters, which often leads to non-elementary integrals.

What do complex roots represent?

An oscillation: the imaginary part β is the angular frequency, and the real part α says whether the amplitude decays (α < 0), grows (α > 0) or stays constant (α = 0).

How this calculation works

Homogeneous: a·r² + b·r + c = 0. Δ > 0: y = C₁e^(r₁x) + C₂e^(r₂x); Δ = 0: y = (C₁ + C₂x)e^(rx); Δ < 0: y = e^(αx)(C₁ cos βx + C₂ sin βx). Particular for f = xᵏe^(mx) cos(ωx) or sin(ωx): write f as the real part of e^(λx)G(x), λ = m + iω, and seek y = Re[e^(λx)R(x)] with a·R″ + (2aλ + b)R′ + p(λ)R = G, multiplying R by x^s when λ is a characteristic root of multiplicity s. The initial conditions give a linear system in C₁ and C₂.