Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I have a problem:

$$\begin{cases} \sin(x) + 2y = 2 \\ \cos(y - 1) + x = 0.7 \end{cases} $$

with margin of error 0.00001

And I need to solve this by using Fixed-point iteration method.

Can someone help me with that? Algorithm for full solution might be the best thing ever, since I gave to solve this in Maple later..

Thanks is advance.

P.S. one more question. What is q in this method and how to find it?

share|improve this question
1  
newton-raphson method en.wikipedia.org/wiki/Newton%27s_method –  Ömer Oct 7 '13 at 17:37
    
@user1075413: Not sure what you mean by that, but you could solve the first equation for $y$, substitute into the second equation and then use Newton's Method (iteration) to find $x$. Lastly, substitute into equation one to find $y$. I get $((x, y) = (-0.289809, 1.14288)$. –  Amzoti Oct 7 '13 at 17:42
    
A simpler method than Newton to start with could be gradient descent –  Keeran Brabazon Oct 7 '13 at 17:49
    
This question can't really be answered without knowing which method you mean. In my opinion, "Simple Iteration Method" sounds like fixed-point iteration. Is any of the suggestions correct? –  JiK Oct 7 '13 at 18:06
    
@JiK it seems that it actually is "fixed-point iteration" method. Can you help with that? –  domakas Oct 7 '13 at 18:10

1 Answer 1

up vote 1 down vote accepted

Try this iteration:

$ \qquad x_{n+1}=0.7-\cos(y_n-1) $

$ \qquad y_{n+1}=(2-\sin(x_n))/2 $

Taking $x_0=y_0=0$ works fine for me. I get that

$\qquad x^*=-0.28980932884903, \quad y^*= 1.1428847552227$

is a fixed point of $F(x,y)=(0.7-\cos(y-1),2-\sin(x))/2)$ after 27 steps, but it only takes 11 steps to get an error below $10^{-6}$. The error is measured by $\| F(x,y)-(x,y) \|_1$, where $\|(x,y)\|_1 = \max(|x|,|y|)$.

share|improve this answer
    
What is the condition while checking for error? I use abs(x[n+1]-x[n]) < (1-q)*e/q, however I don't know how to calculate q and I only check for x (do I need to check y?) –  domakas Oct 7 '13 at 18:31
    
Can you explain what ∥F(x,y)−(x,y)∥ is? And few stupid questions: what ∥ ∥ means? And what F(x,y) is called in English? –  domakas Oct 7 '13 at 18:46
    
Still don't understand what condition should be checked while checking for precision. ∥F(x,y)−(x,y)∥ should be smaller that e or what? –  domakas Oct 7 '13 at 19:48
    
@user1075413, yes, compare that with say $10^{-6}$. –  lhf Oct 7 '13 at 19:51
    
Can't get it right.. after first iteration ∥F(x,y)−(x,y)∥ = 1.447760103... and after 10th iteration 1.432694311... it decreases very very slowly, so it's never lower than 10^-6 –  domakas Oct 7 '13 at 19:57

Your Answer

 
discard

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.