o-nauts, before I begin I must stress that I am self-taught and in college I will be taking Calculus I this semester. Nonetheless I have studied on my own up to the Divergence Theorem and Introductory Complex Analysis. In my venture, I now find myself completing an introductory course of Ordinary Differential Equations (which up until now has mostly been Calcu-Algebra to me).

I have now encountered a particular second-order differential equation that can be reduced to a linear, first-order equation of which the solution I have found is not in agreement with what my shows to be the solution, so here goes:

Given $$\frac{d^2y}{dx^2} + \frac{dy}{dx} = 0$$

Let $p = \frac{dx}{dy}$ and $p\frac{dp}{dy} = \frac{d^2y}{dx^2}$

$$\frac{d^2y}{dx^2} + \frac{dy}{dx} = 0 \Rightarrow p\frac{dp}{dy} + p =0$$

Which simplifies to:

$$\frac{dp}{dy} + 1 = 0$$ which when evaluated with separation of variables, we get: $$p + y = 0$$ which is $$\frac{dy}{dx} + y + C_1 = 0$$

I go on to do this:

$$\frac{dy}{dx} + y + C_1 = 0$$ $$\frac{dy}{dx} + y = -C_1$$ which to me is a first order linear equation and can be solved using the canonical integrating factor.

Thus, given the standard for of a first-order linear equation: $$ \frac{dy}{dx} + Py = Q$$ Let: $$P = 1, Q = -C_1 \quad and \quad \rho = e^{\int Pdx}$$

The equation $$\frac{dy}{dx} + y = -C_1$$ evaluates to $$\rho y = \int \rho (-C_1) dx + C_2$$

Which when evaluted (atleast for me), we get:

$$y = C_2e^{-x} - C_1$$

Whereas my textbook's solution is: $$y = C_1e^{-x} + C_2$$

Now, what the !@#$%^& am I doing wrong here?

Oh, and I am using Addison-Wesley's "Calculus and Analytic Geometry" printed in 1968.

Thank You!

  • $\begingroup$ $-C_1=K_1$ thats it your solution is correct to me Its just a constant after all so writing $-C_1$ or $C_1$ its the same $\endgroup$ – Isham Jul 13 '18 at 14:59

Your answer is correct, $C_1$ is a constant so writing $-C_1$ or $K_1$ is the same..

Note that you can integrate directly $$y''+y'=0$$ Integrate $$y'+y=K_1$$ Multiply by $e^x$ $$(ye^x)'=K_1e^x$$ Integrate $$ye^x=K_1e^x+K_2$$ Finally $$y(x)=K_1+K_2e^{-x}$$


Your answer and your book's answer are equivalent because C_1 and C_2 are arbitrary constants to be found by initial information.

Note that your solution could have been shorter as follows:

$$y'' +y' =0$$

$$ u=y' \implies u'+u=0$$ $$ u'=-u \implies u=C_1 e^{-x} $$ $$ y'= C_1 e^{-x} \implies y=-C_1e^{-x}+C_2 $$

  • $\begingroup$ What I don't understand is that clearly $-C_1& is not $C_1$. I am limited to what I can learn as differential equations are beyond the scope of my bachelor's degree program. Thank you nonetheless!! $\endgroup$ – SSBASE Jul 13 '18 at 15:23
  • $\begingroup$ @SSBASE C_1 is an arbitrary constant, you may call it K or C or -C_1 or whatever else. Once the initial values are given then the constants will be uniquely determined. Sort of like the constant C in an integral. $\endgroup$ – Mohammad Riazi-Kermani Jul 13 '18 at 15:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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