Differential Equations › Section G · Higher order differential equations › Topic G.2

Homogeneous nth order equations

For constant coefficients, an equation of any order still comes down to the roots of one polynomial. The new features are roots repeated more than twice, complex roots that repeat, and the practical work of finding all the roots in the first place.

The same trial, a bigger polynomial

Take a homogeneous equation with constant coefficients,

a_n y^{(n)} + a_{n-1} y^{(n-1)} + \cdots + a_1 y' + a_0 y = 0

As in the second order case, try \(y = e^{rt}\). Each derivative brings out one factor of \(r\), so every term is a multiple of \(e^{rt}\), and \(e^{rt}\) solves the equation exactly when

a_n r^n + a_{n-1} r^{n-1} + \cdots + a_1 r + a_0 = 0

This is the characteristic equation. By the fundamental theorem of algebra, it has exactly \(n\) roots when each is counted as often as it repeats. The job is to turn that list of \(n\) roots into \(n\) independent solutions, a fundamental set in the sense of the topic Basic concepts for nth order equations.

Distinct real roots

Each real root \(r\) that appears only once contributes one solution, \(e^{rt}\). Exponentials with different exponents are always linearly independent, so if all \(n\) roots are real and different,

y = c_1 e^{r_1 t} + c_2 e^{r_2 t} + \cdots + c_n e^{r_n t}

Nothing here is new; the only extra effort is in finding the roots and in fitting \(n\) constants. For a third order equation, that means a system of three linear equations, one each from \(y\), \(y'\) and \(y''\) at the initial point.

Repeated real roots

If a real root \(r\) appears \(k\) times in the list, it has multiplicity \(k\), and \(e^{rt}\) alone accounts for only one of those \(k\) places. The rest come from multiplying by powers of \(t\):

e^{rt},\ t e^{rt},\ t^2 e^{rt},\ \ldots,\ t^{k-1} e^{rt}

These \(k\) functions are independent, and the highest power is \(t^{k-1}\), not \(t^k\). For example, \(y''' - 3y'' + 3y' - y = 0\) has characteristic polynomial \((r - 1)^3\), a triple root, so \(y = (c_1 + c_2 t + c_3 t^2)e^t\). Second order equations allowed at most a double root; now a root can repeat as often as the order allows.

Complex roots, single and repeated

With real coefficients, complex roots come in conjugate pairs \(\lambda \pm \mu i\), and each pair appearing once contributes the two real solutions \(e^{\lambda t}\cos \mu t\) and \(e^{\lambda t}\sin \mu t\), by Euler's formula as before. From fourth order on, a pair can itself repeat. If \(\lambda \pm \mu i\) has multiplicity \(k\), multiply both functions by \(1, t, \ldots, t^{k-1}\), giving \(2k\) solutions in all. Because complex roots pair up, a polynomial of odd degree always has at least one real root, a useful fact when hunting for the first one.

Finding the roots

With degree 3 or more, factoring is the hard part. Start with the rational root test: if the coefficients are integers, any rational root \(\frac{p}{q}\) in lowest terms has \(p\) dividing \(a_0\) and \(q\) dividing \(a_n\). Test those candidates; each root \(c\) you find lets you divide out \(r - c\) and continue with a smaller polynomial. Look for patterns too: \(r^4 - 16\) is a difference of squares, and \(r^4 + 8r^2 + 16\) is a quadratic in \(r^2\). When none of this works, the roots are found numerically with software, and the solution is written from those values.

Worked example: a cubic initial value problem

Solve \(y''' - 2y'' - 5y' + 6y = 0\) with \(y(0) = 3\), \(y'(0) = 2\), \(y''(0) = 14\). 1. Characteristic equation: \(r^3 - 2r^2 - 5r + 6 = 0\). The candidates are \(\pm 1, \pm 2, \pm 3, \pm 6\), and \(r = 1\) works: \(1 - 2 - 5 + 6 = 0\). 2. Dividing by \(r - 1\) leaves \(r^2 - r - 6 = (r - 3)(r + 2)\), so the roots are \(1, 3, -2\). 3. General solution: \(c_1e^t + c_2e^{3t} + c_3e^{-2t}\). 4. Conditions:

c_1 + c_2 + c_3 = 3, \quad c_1 + 3c_2 - 2c_3 = 2, \quad c_1 + 9c_2 + 4c_3 = 14

5. These give \(c_1 = c_2 = c_3 = 1\); check: \(1 + 3 - 2 = 2\) and \(1 + 9 + 4 = 14\). So \(y = e^t + e^{3t} + e^{-2t}\).

Worked example: a repeated complex pair

Solve \(y^{(4)} + 8y'' + 16y = 0\). 1. Characteristic equation: \(r^4 + 8r^2 + 16 = 0\), a quadratic in \(r^2\) that is a perfect square, \((r^2 + 4)^2 = 0\). 2. So \(r = \pm 2i\), each with multiplicity 2: here \(\lambda = 0\) and \(\mu = 2\). 3. The single-pair solutions are \(\cos 2t\) and \(\sin 2t\); the repeat adds \(t\cos 2t\) and \(t\sin 2t\). 4. General solution:

y = (c_1 + c_2 t)\cos 2t + (c_3 + c_4 t)\sin 2t

Four functions for a fourth order equation, as required. The \(t\cos 2t\) terms oscillate with an amplitude that grows in proportion to \(t\), even though the equation has no forcing and no negative damping.

Worked example: every case at once

One equation can mix all the cases. Suppose a fifth order equation has characteristic polynomial

(r + 1)^2 (r - 3)(r^2 + 2r + 5)

1. The factor \((r + 1)^2\) is a double root \(-1\): \(e^{-t}\) and \(te^{-t}\). 2. The factor \(r - 3\) is a simple root: \(e^{3t}\). 3. The quadratic \(r^2 + 2r + 5\) has roots \(-1 \pm 2i\), found with the quadratic formula: \(e^{-t}\cos 2t\) and \(e^{-t}\sin 2t\). 4. Count: \(2 + 1 + 2 = 5\) functions, matching the order. 5. So

y = (c_1 + c_2 t)e^{-t} + c_3 e^{3t} + e^{-t}(c_4\cos 2t + c_5\sin 2t)

Similarly, \(y^{(4)} - 16y = 0\) factors as \((r - 2)(r + 2)(r^2 + 4)\), giving \(e^{2t}\), \(e^{-2t}\), \(\cos 2t\) and \(\sin 2t\).

Common mistakes to avoid

First, losing roots: always count, because the number of functions must equal the order. Second, stopping the powers of \(t\) too early or too late; multiplicity \(k\) gives powers \(t^0\) up to \(t^{k-1}\). Third, taking a square root where a fourth root is needed: \(r^4 = 16\) has real roots \(\pm 2\), not \(\pm 4\), plus the complex pair \(\pm 2i\). Fourth, treating a repeated complex pair like a single one and missing \(t\cos \mu t\) and \(t\sin \mu t\). Fifth, testing only whole-number candidates when \(a_n \ne 1\): fractions \(\frac{p}{q}\) can be roots too, as in \(2r^2 + r - 1\), with root \(\frac{1}{2}\).

Key takeaways

Trying \(e^{rt}\) turns a constant-coefficient equation of order \(n\) into a characteristic polynomial of degree \(n\) with \(n\) roots counted with multiplicity. A simple real root gives \(e^{rt}\); a root of multiplicity \(k\) gives \(e^{rt}\) through \(t^{k-1}e^{rt}\); a complex pair \(\lambda \pm \mu i\) gives \(e^{\lambda t}\cos \mu t\) and \(e^{\lambda t}\sin \mu t\), multiplied by the same powers of \(t\) if it repeats. Find roots with the rational root test, division and patterns, or software. Always check that the count of functions equals the order.

Flashcards

Flashcard

Characteristic equation

aₙrⁿ + … + a₁r + a₀ = 0

From trying y = eʳᵗ.

Test your understanding

Question 1 of 5

The characteristic polynomial of y‴ + y″ - y′ - y = 0 is (r - 1)(r + 1)². What is the general solution?

Question 2 of 5

What is the general solution of y⁽⁴⁾ - 16y = 0?

Question 3 of 5

Which function belongs to the general solution of y⁽⁴⁾ + 18y″ + 81y = 0?

Question 4 of 5

A constant-coefficient equation has characteristic roots 0, 0, 0 and 5. What is its general solution?

Question 5 of 5

What does a real root of multiplicity 4 contribute to the general solution?

← Previous topicNext topic →