Saturday, 21 April 2012

Quadratic Formula

We all remember the quadratic formula from high school,  x = ( -b ± √ b² - 4ac ) / 2a, but if it's so familiar to us all of us, why do I keep coming across such bad implementations in code?

Polynomials

Lets start with a slightly more general problem, we're interested in finding all the real roots of these polynomials in x :


k0  +  k1 x  =  0

k0  +  k1 x  +  k2 x2  =  0

k0  +  k1 x  +  k2 x2  +  k3 x3  =  0

k0  +  k1 x  +  k2 x2  +  k3 x3  +  k4 x4   =  0

k0  +  k1 x  +  k2 x2  +  k3 x3  +  k4 x4  +  k5 x5  =  0

We know from the Fundamental Theorem Of Algebra that, for example, a quintic will have at most 5 real roots, so lets go ahead and write out the function prototypes :


(We also know from Galois Theory that quintics might require a numerical approximation, but that's a post for another day!)

The Quadratic Formula, Revisited

descriminant = kLinear² - 4*kQuadratic*kConstant

x = ( -kLinear ± √ descriminant ) / 2 / kQuadratic

So the first case that I hardly ever see handled, is when kQuadratic vanishes, leading to a divide-by-zero exception.  Actually, this case is very easy to handle if we have a family of functions:


Another important case is when kConstant is zero.  We can see graphically that the polynomial goes through the origin :
kConstant is zero means that 0.0 is a root.

Here's the code:
Note the sneaky 1+ and +1

Catastrophic Cancellation

The last problem we need to fix is when the square root of the descriminant is close in magnitude to kLinear.  Graphically :
This happens surprisingly often in physics calculations.
To fix this, consider the product of the roots :
rootProduct = root[0] * root[1]
rootProduct =  (-kLinear + √descriminant)/2/kQuadratic *
(-kLinear - √descriminant)/2/kQuadratic
rootProduct = ( -kLinear + √ descriminant ) * ( -kLinear - √ descriminant ) / 4 / kQuadratic²


Now something cool happens! Remember that (x+y) * (x-y) = x² - y²

rootProduct = ( (-kLinear)² - √ descriminant² ) / 4 / kQuadratic²


rootProduct = ( kLinear² - descriminant ) / 4 / kQuadratic²

Expanding:
rootProduct = ( kLinear² - kLinear² + 4*kQuadratic*kConstant) / 4 / kQuadratic²

rootProduct = ( kQuadratic*kConstant) / kQuadratic²


rootProduct = kConstant / kQuadratic

What this means is that we can use the larger root to calculate the smaller root by dividing by their product, and keep all the accuracy.


Appendix - Quadratic formula in C++

If you're just here for the code, the following two functions are hereby licensed under CC0


#include <math.h>
#define EPSILON (1e-10)

// From http://missingbytes.blogspot.com/2012/04/quadratic-formula.html
int SolveLinear(double rootDest[1],double kConstant,double kLinear){
    if(-EPSILON<kLinear&&kLinear<EPSILON){
        return 0;
    }
    rootDest[0]=-kConstant/kLinear;
    return 1;
}

// From http://missingbytes.blogspot.com/2012/04/quadratic-formula.html
int SolveQuadratic(double rootDest[2],double kConstant,double kLinear,double kQuadratic){
    if(-EPSILON<kQuadratic&&kQuadratic<EPSILON){
        return SolveLinear(rootDest,kConstant,kLinear);
    }
    if(-EPSILON<kConstant&&kConstant<EPSILON){
        rootDest[0]=0.0f;
        return 1+SolveLinear(rootDest+1,kLinear,kQuadratic);
    }
    double descriminant=kLinear*kLinear-4*kQuadratic*kConstant;
    if(descriminant<0){
        return 0;
    }
    if(descriminant==0){// no EPSILON check!!
        rootDest[0]=-kLinear/(2*
kQuadratic);
        return 1;
    }
    double adder=sqrt(descriminant);
    if(adder*kLinear>0){
        adder=-adder;
    }
    double rootLarge=(-kLinear+adder)/(2*kQuadratic);
    rootDest[0]=rootLarge;
    rootDest[1]=kConstant/(kQuadratic*rootLarge);
    //std::sort(rootDest,rootDest+2);
    return 2;
}


Even More Math

They tell me that each mathematical formula in your blog post exactly halves your potential readership.

  How does that make you feel?

Why not talk about your feelings in the comments below?

5 comments:

  1. Nice post Chris! When are you going to share your graphing software with us?

    ReplyDelete
  2. Sweet! Actually it's a work-in-progress research I cooked up in response to the 2001 SIGGRAPH paper, "Reliable two-dimensional graphing methods for mathematical formulae with two free variables."

    I added my own secret sauce for anti-aliasing, inequalities, and threading, and so the resultant code is a little hairy and not really fit for public eyes.

    Mine also doesn't handle branch cuts like in the original. e.g. sin(y)=x is fine, but it chokes on y=asin(x)

    If you're still keen, I could maybe send you a binary to play with? In any case, I totally recommend Jeff Tupper's paper, it's got the good stuff...

    ReplyDelete
  3. Your code screen shots remind me of Yate - do you still use it? The replacement of && with ∩ is cute; reminds me of pretty-mode for Emacs (http://www.emacswiki.org/emacs/pretty-mode.el).

    ReplyDelete
  4. Yup, it's the same code from YATE, with a few improvements and ports along the way.

    I didn't know about pretty-mode in Emacs, that's pretty sweet!

    I do a number of replacements, mostly to aid visual pattern matching :

    -> →
    && ∩
    || ∪
    != ≠
    == ≈ // !!! not correct at all !!!
    <= ≤
    >= ≥
    << «
    >> »
    ++ < sup>++< /sup>

    I also used to match "<=", and the ligatures "fi" and "fl". It took a while for me to figure out that matching non-coding sequences
    doesn't help you write code that is more correct. i.e. it doesn't improve productivity.

    It also helped me feel better about remapping "==" to a symbol which means approximately. It's wrongness is outweighed by it's usefulness.

    And a quick shout out to any budding language designers reading this, please consider using ":=" for the assignment operator!

    ReplyDelete
  5. That's right, Rust, I'm talking about you...

    http://www.rust-lang.org

    ReplyDelete