Jump to content
  • Advertisement
Sign in to follow this  
udvat

Pre search for root

This topic is 3606 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

Hello, For a function, I want to calculate the range where one root exists. For example, f(x) has a root in subrange [a1,b1], another root in subrange[a2,b2] in a total range of [X,Y]. Is any such algorithm available to calculate the ranges [a1,b1],[a2,b2]? Thanks in advance.

Share this post


Link to post
Share on other sites
Advertisement
I remember asking about this when I took numerical analysis. It really does seem like an important part of the problem.

Unfortunately, even if f is differentiable on (a,b), it can have infinitely many roots.

What you might try is making some more constraining assumptions about your function. Something like differentiable on [a,b] with the derivative bounded by a constant. That enables you to calculate neighborhoods around a non-root point where you know there can't be any roots. I'm guessing you can slowly fill up the interval that way, finding most roots in the process. Even still, imagine the function f(x)=0. How would you know it doesn't have a little bump somewhere in your interval?

I wish I knew of a paper about this. I would imagine someone has 'solved' this problem before.

Share this post


Link to post
Share on other sites
I found the method "bracketing method" in the following link

http://civilwares.free.fr/27102005/Applied%20Numerical%20Methods%20with%20MATLAB%20for%20Engineers%20and%20Scientists/cha92657_ch05.pdf

It works fine for my purpose. Thanks to all.

Share this post


Link to post
Share on other sites
If the function f is an analytic function some powerful tools from complex analysis can be used. Specially Rouche's theorem can be used to determine the number of roots inside some closed contour. If there is only one root inside the contour, even the position of the root can be determined by a numerical integration.
More information and some samples can be found in chapter 3 of the book Complex Analysis by Freitag and Busam .
Unfortunately there aren't such elegant theorems available for non analytic function's. But for continuous (and smooth enough) functions one can take many samples and use the fact that if f(a) and f(b) have different signs then there exists at least one root in the interval (a,b) .

Share this post


Link to post
Share on other sites
Sign in to follow this  

  • Advertisement
×

Important Information

By using GameDev.net, you agree to our community Guidelines, Terms of Use, and Privacy Policy.

We are the game development community.

Whether you are an indie, hobbyist, AAA developer, or just trying to learn, GameDev.net is the place for you to learn, share, and connect with the games industry. Learn more About Us or sign up!

Sign me up!