Differential Equations › Section F · Series solutions › Topic F.3
Euler (Cauchy-Euler) equations
One family of variable-coefficient equations can be solved exactly, in closed form, even at a point where the series method breaks down. Euler equations show what solutions do near a singular point, and they are the model for the more general method that follows.
Powers matched to derivatives
An Euler equation, also called a Cauchy-Euler or equidimensional equation, has the form
with constants \(a \ne 0\), \(b\) and \(c\). Each derivative is multiplied by the matching power of \(x\). Dividing by \(ax^2\) gives \(P = \frac{b}{ax}\) and \(Q = \frac{c}{ax^2}\), neither analytic at 0, so \(x = 0\) is a singular point and the method of the topic Series solutions near ordinary points cannot be used there. Yet these equations have exact solutions. They appear in many problems with radial symmetry, such as steady heat flow in a disk or stress in a thick-walled pipe, once the equation is written in terms of the distance from the centre.
Trying a power of x
In a constant-coefficient equation, \(e^{rx}\) works because each derivative returns the same exponential. Here a power of \(x\) plays that role. Assume \(x > 0\) and try \(y = x^r\). Then \(xy' = r x^r\) and \(x^2 y'' = r(r-1)x^r\), so every term is a multiple of \(x^r\):
Since \(x^r \ne 0\), the power works exactly when
This quadratic is the indicial equation of the Euler equation. Its two roots decide the solution, and, as for constant coefficients, there are three cases: distinct real roots, a repeated root and complex roots.
Case 1: distinct real roots
If the indicial equation has two different real roots \(r_1\) and \(r_2\), then \(x^{r_1}\) and \(x^{r_2}\) are two solutions. Neither is a constant multiple of the other, so they form a fundamental set and
The roots need not be whole numbers: \(2x^2y'' + 3xy' - y = 0\) has indicial equation \(2r^2 + r - 1 = 0\), roots \(\frac{1}{2}\) and \(-1\), and solutions \(\sqrt{x}\) and \(x^{-1}\). A fractional power like this is one reason the restriction \(x > 0\) is needed.
Worked example: an initial value problem
Solve \(x^2 y'' - 2xy' - 4y = 0\) with \(y(1) = 0\) and \(y'(1) = 5\). 1. Indicial equation:
so \(r = 4\) or \(r = -1\). 2. General solution: \(y = c_1 x^4 + c_2 x^{-1}\), so \(y' = 4c_1 x^3 - c_2 x^{-2}\). 3. At \(x = 1\): \(c_1 + c_2 = 0\) and \(4c_1 - c_2 = 5\). Adding gives \(5c_1 = 5\), so \(c_1 = 1\) and \(c_2 = -1\). 4. Answer: \(y = x^4 - x^{-1}\) for \(x > 0\). Check \(x^{-1}\): the three terms give \((2 + 2 - 4)x^{-1} = 0\). The initial conditions sit at \(x = 1\), because \(x = 0\) is the singular point.
Case 2: a repeated root
If the indicial equation has a double root \(r\), then \(x^r\) gives only one solution. The second is \(x^r \ln x\), so
Where the constant-coefficient case multiplied by \(x\), the Euler case multiplies by \(\ln x\); the next cards show why. For instance \(x^2y'' - 3xy' + 4y = 0\) has indicial equation \(r^2 - 4r + 4 = (r - 2)^2\), so its general solution is \(x^2(c_1 + c_2 \ln x)\).
Worked example: a double root
Solve \(x^2y'' + 5xy' + 4y = 0\) with \(y(1) = 2\) and \(y'(1) = -1\). 1. Indicial equation: \(r^2 + 4r + 4 = (r + 2)^2\), a double root \(r = -2\). 2. General solution: \(y = x^{-2}(c_1 + c_2 \ln x)\). 3. By the product rule,
4. At \(x = 1\), \(\ln 1 = 0\), so \(y(1) = c_1 = 2\) and \(y'(1) = -2c_1 + c_2 = -1\), giving \(c_2 = 3\). 5. Answer: \(y = x^{-2}(2 + 3\ln x)\). As \(x \to 0^+\) this blows up, which is typical of solutions near a singular point.
Case 3: complex roots
If the roots are \(\lambda \pm \mu i\), write \(x^{\mu i} = e^{i\mu \ln x}\) and apply Euler's formula:
The real and imaginary parts are real solutions, so
For \(x^2y'' - xy' + 5y = 0\), the indicial equation \(r^2 - 2r + 5 = 0\) gives \(r = 1 \pm 2i\), so \(\lambda = 1\) and \(\mu = 2\): the solutions are \(x\cos(2\ln x)\) and \(x\sin(2\ln x)\). The oscillation runs in \(\ln x\), so it speeds up without limit as \(x \to 0^+\).
Why logarithms appear
The substitution \(x = e^t\), so \(t = \ln x\), explains all three cases at once. Let \(Y(t) = y(e^t)\). The chain rule gives \(Y' = xy'\) and \(Y'' = xy' + x^2y''\), so \(x^2y'' = Y'' - Y'\). The Euler equation becomes
a constant-coefficient equation in \(t\). Its characteristic equation \(ar^2 + (b - a)r + c = 0\) is the indicial equation rearranged. Translating back: \(e^{rt} = x^r\), the repeated-root solution \(te^{rt}\) becomes \(x^r \ln x\), and \(e^{\lambda t}\cos \mu t\) becomes \(x^{\lambda}\cos(\mu \ln x)\).
Negative x and other centres
For \(x < 0\), replace \(x\) by \(|x|\) everywhere, so the distinct-root case reads \(c_1|x|^{r_1} + c_2|x|^{r_2}\), and \(\ln x\) becomes \(\ln|x|\). These formulas hold on any interval that does not contain 0. The same ideas cover the shifted form
whose singular point is \(x_0\): the indicial equation is unchanged and \(x\) becomes \(x - x_0\). For example,
gives \((r + 1)^2 = 0\), so on any interval avoiding \(x = 1\)
Common mistakes to avoid
First, using the constant-coefficient equation \(ar^2 + br + c = 0\): for \(x^2y'' - 2xy' - 4y = 0\) it would give \(r^2 - 2r - 4\) instead of the correct \(r^2 - 3r - 4\). Remember the \(r(r-1)\). Second, writing exponentials \(e^{rx}\) instead of powers \(x^r\). Third, using \(x \cdot x^r\) for a repeated root; the second solution is \(x^r \ln x\). Fourth, writing \(\cos(\mu x)\) instead of \(\cos(\mu \ln x)\) for complex roots, and forgetting the factor \(x^{\lambda}\). Fifth, imposing initial conditions at the singular point \(x = 0\).
Key takeaways
An Euler equation \(ax^2y'' + bxy' + cy = 0\) has a singular point at 0 but exact solutions on \(x > 0\). Trying \(y = x^r\) gives the indicial equation \(ar(r - 1) + br + c = 0\). Distinct real roots give \(c_1x^{r_1} + c_2x^{r_2}\); a double root gives \(x^r(c_1 + c_2\ln x)\); complex roots \(\lambda \pm \mu i\) give \(x^{\lambda}\cos(\mu\ln x)\) and \(x^{\lambda}\sin(\mu\ln x)\). The substitution \(x = e^t\) turns the equation into a constant-coefficient one and explains every case. Use \(|x|\), or \(|x - x_0|\) for the shifted form, on intervals avoiding the singular point.
Flashcards
Flashcard
ax²y″ + bxy′ + cy = 0
Test your understanding
Question 1 of 5
What is the general solution of 2x²y″ + 3xy′ - y = 0 for x > 0?
Question 2 of 5
What is the general solution of x²y″ - 3xy′ + 4y = 0 for x > 0?
Question 3 of 5
What is the general solution of x²y″ + xy′ + 9y = 0 for x > 0?
Question 4 of 5
With x = eᵗ and Y(t) = y(eᵗ), what does x²y″ + bxy′ + cy = 0 become?
Question 5 of 5
Solve x²y″ + xy′ - 4y = 0 with y(1) = 3 and y′(1) = 2.
