The Student Room Group

Scroll to see replies

Reply 20
Well there's no other c3 coursework for any other boards.
Tallon
It's not as if I need help from an expert. I only wanted some questions answered by people who have done it before. I don't see what the problem is actually because the title is obviously good enough and every single board does the same c3 coursework.
This forum is for discussing mathematics, and the guidelines specifically say to

Guide to Posting
post your question clearly, with no inaccuracies, and nothing relevant missed out.


If, instead, you want to post in such a way that it only makes sense to people doing the exam, then the Maths Exams forum would be more appropriate.

As far as "needing an expert", I confess I have no idea what the expectations are for C3, but in the "real world" where people use N-R, the most common failure cases have nothing to do with the derivative equalling 0. (And at least one link for the C3 coursework also has failures other than the derivative = 0 one).
Reply 22
DFranklin
This forum is for discussing mathematics, and the guidelines specifically say to



If, instead, you want to post in such a way that it only makes sense to people doing the exam, then the Maths Exams forum would be more appropriate.

As far as "needing an expert", I confess I have no idea what the expectations are for C3, but in the "real world" where people use N-R, the most common failure cases have nothing to do with the derivative equalling 0. (And at least one link for the C3 coursework also has failures other than the derivative = 0 one).



The only other failures I can think of would be where the derivative is close to 0 and so it ends up finding a different root than the starting value would suggest it would find, or an asymptote problem where it looks like it's converging but really it's getting nowhere, or maybe the estimates would somehow get stuck in a loop going to and from two different places back and forth. Or something you can't differentiate obviously.

Sorry if I've seemed to annoyed you or something, I was just looking for some answers, as are a lot of people I think. There's even another coursework thread that's been made today by somebody else.




btw - does anybody know if it's ok if you actually find the root with NR? I've been trying with loads of different functinos are after about 5 or 6 iterations they all seem to have found it. Not sure if that's allowed though. And do you really have to tediuously explain all the formulae you've used in excel as well?
Tallon
The only other failures I can think of would be where the derivative is close to 0 and so it ends up finding a different root than the starting value would suggest it would find, or an asymptote problem where it looks like it's converging but really it's getting nowhere, or maybe the estimates would somehow get stuck in a loop going to and from two different places back and forth. Or something you can't differentiate obviously.
Fair enough. They are all much more common failure cases (other than looking like it's converging when it isn't, which doesn't happen very often) than the derivative being zero (or even near zero).

You might find http://en.wikipedia.org/wiki/Newton_fractal interesting - lots of pictures of N-R convergence (or failure) in the complex plain.

Sorry if I've seemed to annoyed you or something, I was just looking for some answers, as are a lot of people I think. There's even another coursework thread that's been made today by somebody else.
The issue isn't so much posting a coursework thread, it's that someone asked for context, and you decided not to bother to provide it because {paraphrased} "everyone doing C3 will know what I'm talking about".

Not only is that attitude unhelpful, but one of the most common reasons people refuse to post details is because they know if they did, it would show they were breaking the rules of their course. (I'm not saying this is the case here, but I am saying why it makes people suspicious).

btw - does anybody know if it's ok if you actually find the root with NR? I've been trying with loads of different functinos are after about 5 or 6 iterations they all seem to have found it. Not sure if that's allowed though. And do you really have to tediuously explain all the formulae you've used in excel as well?
I don't understand what you mean, to be honest. But looking at the docs in the other thread, it's fairly explicit that you are supposed to find a sign change (e.g. f(.99995) < 0 < f(1.00005)) in order to demonstrate there really is a root where you say there is.

Latest