4
$\begingroup$

Having a bit of difficulty with this question:

Convert the initial value problem,

$$\frac{d^2y}{dx^2}-\frac{dy}{dx}+2y=x+1$$ where $$y(0)=2,\ \frac{dy}{dx}(0)=-1,$$

into a set of two coupled first-order initial value problems.

Firstly I let $$\frac{dy}{dx}=z$$ giving me the equation:

$$\frac{dz}{dx}=z-2y+x+1$$ where $$y(0)=2,\ z(0)=-1,\ x(0)=0$$

Use a step length h = 0.1 to find numerical approximations to y at x = 0.2 using generalizations to two coupled equations of the following numerical schemes:

  1. Trapezoidal method using the explicit Euler method as a predictor and one iteration of the corrector.

for this method I'm really confused and don't know where to start so any help would be great.

  1. the third-order Runge-Kutta scheme: $$k_1 = hf (x_n, y_n)\\ k_2=hf(x_n+\frac{h}{2},y_n+\frac{k_1}{2})\\ k_3=hf(x_n+h,y_n-k_1+2k_2)\\ y_{n+1}=y_n+\frac{1}{6}(k_1+4k_2+k_3) $$

for this method I'm a little confused I don't know where to start because of the 3 different variables z,y,x I tried starting by doing $$k_1=hz_0$$ and then $$l_1=hf(x_0,y_0)$$ but I found it difficult because of the 3 variables, any help is appreciated thanks.

$\endgroup$
  • $\begingroup$ Note: I've split this into 3 variables using $$\frac{dy}{dx}=f(x,y,z)=z$$ and got to the solution of $$k_2\ and \ l_2$$ but cant seem to figure out $$k_3$$ $\endgroup$ – Reety Jan 15 '18 at 23:48
1
$\begingroup$
  1. Trapezoidal method with Euler as predictor and one corrector step can be implemented as a 3-stage Runge-Kutta method for the vector $u=(y,z)$ \begin{align} k_1 &= hf(x,u)\\ k_2 &= hf(x+h,u+k_1)\\ k_3 &= hf(x+h,u+0.5(k_1+k_2))\\ u_+ &= u+0.5(k_1+k_3) \end{align}

    I assume that the involved description is to distinguish the method from the explicit trapezoidal method and to compare the 2nd order 3-stage method to the 3rd order 3 stage method of the second point.

  2. Yes, depending on notation you can expand the scalar method to the vector method that way. \begin{align} k_{1y} &= hz \\ k_{1z} &= hf(x,y,z) \\ k_{2y} &= h(z+0.5k_{1z}) \\ k_{2z} &= hf(x+0.5h, y+0.5k_{1y}, z+0.5k_{1z}) \end{align} etc.

There are only 2 components to this differential equation, the independent variable $x$ has explicitly available updates.

$\endgroup$

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.