MHB Finding an Expression for Y in a Linear Multistep Method

Click For Summary
The discussion focuses on deriving an expression for \( y_n \) in a linear multistep method (LMM) applied to the initial value problem \( y'(t) = y(t) \) with \( y(0) = 1 \). The LMM equation is given as \( y_n - \frac{3}{2}y_{n-1} + \frac{1}{2}y_{n-2} = h\left(\frac{1}{2}f_n + \frac{1}{4}f_{n-1} - \frac{1}{4}f_{n-2}\right) \), which simplifies to express \( y_n \) in terms of previous values. The initial conditions are set as \( y_0 = 1 \) and \( y_1 = 1 \), but using \( y_1 = y_0(1+h) \) is suggested for a better approximation. This adjustment leads to a more accurate integration of the LMM, improving the overall solution quality. The conversation emphasizes the importance of selecting appropriate starting values in multistep methods.
shen07
Messages
54
Reaction score
0
consider the following LMM

$$y_n-\frac{3}{2}y_{n-1}+\frac{1}{2}y_{n-2}=h(\frac{1}{2}f_n+\frac{1}{4}f_{n-1}-\frac{1}{4}f_{n-2})$$

which is applied to the initial value problem $$y'(t)=y(t),0\leq{t}\leq{1}\\ and\\ y(0)=1$$

How do i find an expression for $$y_n$$, if the starting values are $$y_0=1\\ and\\ y_1=1$$
 
Physics news on Phys.org
shen07 said:
consider the following LMM

$$y_n-\frac{3}{2}y_{n-1}+\frac{1}{2}y_{n-2}=h(\frac{1}{2}f_n+\frac{1}{4}f_{n-1}-\frac{1}{4}f_{n-2})$$

which is applied to the initial value problem $$y'(t)=y(t),0\leq{t}\leq{1}\\ and\\ y(0)=1$$

How do i find an expression for $$y_n$$, if the starting values are $$y_0=1\\ and\\ y_1=1$$

Usually you have a problem in the form:

$$\frac{dy}{dt}=f(t,y)$$

with an initial condition $y(0)=y_0$. So here $f_n=f(t_n,y_n)=y_n$ and your method reduces to:

$$y_n-\frac{3}{2}y_{n-1}+\frac{1}{2}y_{n-2}=h\left(\frac{1}{2}y_n+\frac{1}{4}y_{n-1}-\frac{1}{4}y_{n-2}\right)$$

which can be rearranged to give $y_n$ in terms of $y_{n-1}$ and $y_{n-2}$

Reassuringly this integrates up nicely to give a reasonable approximation to the analytic solution after allowance is made for the poor choice of $y_1$. A much better choice for $y_1$ would be $y_0+h\times f(t_0,y_0)=y_0\times (1+h)$, effectively taking an Euler step to get your multi-step method started.

.
 
Last edited:

Similar threads

  • · Replies 7 ·
Replies
7
Views
4K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 2 ·
Replies
2
Views
794
  • · Replies 5 ·
Replies
5
Views
3K
  • · Replies 5 ·
Replies
5
Views
2K
Replies
2
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 1 ·
Replies
1
Views
348