Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

In a book the author says the solution $u(x)$ for differential equation

$-\frac{d^2u}{dx^2} + \frac{du}{dx} = 1$

is the particular solution $u = x$ plus any $A + Be^{x}$. The $A$ and $B$ can be calculated from boundary conditions. I checked this and it works. I am confused however on how this solution is reached. Did he guess? Or did he find solutions for $\frac{du}{dx}$ and $-\frac{d^2u}{dx^2}$ seperately somehow add them up?

Any ideas?

share|cite|improve this question
See example 1 of section Solving First Order Linear Equations here:… – Did Aug 12 '11 at 16:17
Hi Didier, the solution from that example came out in the form of $y = constant + constant / e^{mx}$. The solution above is in the form of $y = x + Ax + Be^{x}$. What is the reason for this difference? – BB_ML Aug 12 '11 at 16:22
Function $du/dx$ vs function $y$. – Did Aug 12 '11 at 16:25
up vote 4 down vote accepted

The simplest thing to do here is to treat this as a first order equation in $y=\frac{du}{dx}$ first, and solve it. Note that the equation becomes $$-y' + y = 1,$$ which is a first order linear separable equation. We can solve it the usual way: if $y=1$ is constant, then the equation is satisfied; this is a particular/stable solution. Assume then that $y$ is not the constant function $1$. Then: $$\begin{align*} -\frac{dy}{dx} + y &= 1\\ y-1 &= \frac{dy}{dx}\\ dx &= \frac{dy}{y-1} &&\text{(this is where we use that }y\neq 1\text{)}\\ \int \,dx &= \int \frac{dy}{y-1}\\ x +C &= \ln|y-1| &&C\text{ arbitrary;}\\ e^{x+C} &= |y-1|\\ Ae^x &= |y-1| &&A\text{ a positive constant;}\\ Be^x &= y-1 &&B\text{ a nonzero constant;}\\ y &= 1+Be^x &&B\text{ a nonzero constant.} \end{align*}$$ So $y(x) = 1+Be^x$ where $B$ is a nonzero constant. Since $B=0$ gives the stable solution, we have that the general solution is $y(x) = 1+Be^x$ with $B$ an arbitrary constant.

But $y(x) = \frac{du}{dx}$, and what we are really interested in is $u(x)$. But now we have the very easy differential equation $$\frac{du}{dx} = 1 + Be^x,$$ which can be solved directly by integration. We get $$\begin{align*} \frac {du}{dx} &= 1 + Be^x\\ \int \frac{du(x)}{dx}\,dx &= \int(1+Be^x)\,dx\\ u(x) &= x + Be^x + C, \end{align*}$$ where $B$ and $C$ are arbitrary.

If you don't realize you can use the trick of replacing $u'$ with $y$ and solving two first order equations instead of a single second order equation, you can still solve this very easily because this is a (nonhomogeneous) second order equation with constant coefficients.

First, we solve the associated homogeneous equation, $$-u'' + u' = 0.$$ The associated equation is $-r^2 + r = 0$; the roots are $0$ and $1$, so this says that $e^{0x}=1$ and $e^{x}$ generate the space of solutions to the homogeneous equation.

So the general solution will be given by $$\mathbf{s} + A + Be^x$$ where $A$ and $B$ are arbitrary constants, and $\mathbf{s}$ is any particular solution to the equation $-u''+u'=1$. It is easy to find that $u(x)=x$ is a solution to this equation, so the general solution to the differential equation is then given by $$u(x) = x + A + Be^x,$$ with $A$ and $B$ arbitrary constants. Same solution as above.

share|cite|improve this answer
Hi Arturo. The points on that question I asked on dy/dx not being a ratio (and you answered) is through the roof! :) Thanks for this great answer too. – BB_ML Aug 12 '11 at 17:05

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.