I want to solve this problem by using Laplace transform. But I don't know how to transform ${\cal L}[y^2]$. Does anyone has any idea by using Laplace or else? Thanks.

- 10,390
- 2
- 12
- 31

- 11
-
Welcome to MSE. Your question is phrased as an isolated problem, without any further information or context. This does not match many users' quality standards, so it may attract downvotes, or closed. To prevent that, please [edit] the question. This will help you recognise and resolve the issues. Concretely: please provide context, and include your work and thoughts on the problem. These changes can help in formulating more appropriate answers. – José Carlos Santos Aug 19 '22 at 08:49
-
6It's a nonlinear ODE, so Laplace doesn't seem like a good idea! – Hans Lundmark Aug 19 '22 at 09:03
-
1Does this answer your question? How do you solve the Initial value probelm $dp/dt = 10p(1-p), p(0)=0.1$? – Hans Lundmark Aug 19 '22 at 09:04
-
Hint. Try $y=\frac 1 z$ – Claude Leibovici Aug 19 '22 at 09:17
-
1As there are no terms involving $x$, separation of variables looks like a good approach here. – quarague Aug 19 '22 at 09:18
-
1Or try a Riccati approach, setting $y=-u'/u$. – Lutz Lehmann Aug 19 '22 at 09:35
4 Answers
$$y'+y=y^2 \implies y^{-2}y'+y^{-1}=1$$ Let $y^{-1}=v \implies y^{-2} y'=-v'$, thenwe get $$v'=v-1 \implies \int \frac{dv}{v-1}=\int dx+C$$ You may complete it now as $$\frac{1}{y(x)}=De^x+1 \implies y(x)=\frac{1}{De^{x}+1}\implies D=-2$$ Finally $$y(x)=\frac{1}{1-2e^x}$$

- 43,235
-
why are people downvoting this it's literally the textbook approach – algevristis Aug 19 '22 at 15:04
-
-
The OP says "solve using the Laplace transform" ...
As pointed out by Hans, the Laplace transform is no good for nonlinear DEs. We can rearrange this DE, though. Instead of solving for $y$ as a function of $x$, try to solve for $x$ as a function of $y$. $$ \frac{dy}{dx}+y=y^2 \\ \frac{1}{dx/dy} + y = y^2 \\ \frac{dx}{dy} = \frac{1}{y^2-y} $$ Now it is a linear DE, so susceptible for Laplace transform methods.
Next problem: because of the poles at $y=0$ and $y=1$, the Laplace transform of this right-hand side does not exist (as a function). You could try interpreting that Laplace transform as a Schwartz distribution; I won't do that here.
Let's translate the variable: $z=y-2$ so that the poles are at $z=-1, z=-2$ and the Laplace transform will exist. $$ \frac{dx}{dz} = \frac{1}{z^2+3z+2} = \frac{1}{z+1}-\frac{1}{z+2} $$ Now the Laplace transform $X(t) = \mathscr{L}(x(z))$ satisfies (using Laplace transform tables) $$ tX(t)-x(0) = e^t\operatorname{Ei}_1(t)-e^{2t}\operatorname{Ei}_1(2t) $$ where $x(0)$ is $x$ when $z=0$, that is, when $y=2$. Let's call this value $c$. Also note: $\operatorname{Ei}_1$ is an exponential integral function; something we find in our Laplace transform tables. Solve: $$ X(t) = \frac{e^t\operatorname{Ei}_1(t)-e^{2t}\operatorname{Ei}_1(2t)+c}{t} $$ Take inverse Laplace transform (again using tables): $$ x(z) = c-\ln\frac{z+2}{2(z+1)} $$ Although this is guaranteed only for $z>0$ or $y>2$, let's proceed anyway: with $z=y-2$, we get $$ x = c - \ln\frac{y}{2(y-1)} $$ The given initial condition says when $y=-1$ we have $x=0$. Solve to get $c=-2\ln 2$. Then $$ x = -2\ln 2 - \ln\frac{y}{2(y-1)} = \ln \frac{y-1}{2y} \\ e^x = \frac{y-1}{2y} \\ y=\frac{1}{1-2e^x} $$ Plug into the original DE to verify that this is, indeed, the solution.

- 111,679
I would do this a little differently. From $\frac{dy}{dx}+ y= y^2$ we have $\frac{dy}{dx}= y^2- y$. Separating variables, $\frac{dy}{y^2- y}= dx$
$y^2- y= y(y- 1)$ so we can write the right side as "partial fractions" $\frac{1}{y(y-1}= \frac{A}{y}+ \frac{B}{y- 1}$ Multiply on both sides by y(y- 1) $1= 0y+ 1= A(y- 1)+ By= (A+ B)y- A$.
So we have 0= A- B and 1= -A. A= -1 so -1- B= 0 and B= -1.
$\frac{dy}{y(y- 1)}= -\frac{dy}{y}- \frac{dy}{y- 1}$
Integrating -ln(y)- ln(y- 1)= -ln(y(y- 1))= -ln(y^2- y)= x+ C.
Taking the exponential of both sides, $\frac{1}{y^2- y}= C'e^x$ where $C'= e^C$.

- 1,497
$\frac{dy}{dx}=y^2-y \implies \int \frac{dy}{y^2-y}=\int dx$. We get $\ln \frac{y-1}{y}=x+C$. Upon simplification, we get $y=\frac{1}{1-2e^x}$.

- 43,235