Doubt in derivation of Regular falsi method

349 Views Asked by At

I have doubt in derivation. Pls clarify

Let points be $(a, f(a)), (c,0), (b,f(b))$

$y-f(a) = m (x-a) = \frac{f(b)-f(a)}{b-a}(x-a)$

Substituting y=0, x=c,

$0-f(a) = \frac{f(b)-f(a)}{b-a}(c-a) \implies \fbox {c=a-f(a) ( $\frac{b-a}{f(b)-f(a)}$)}$ ---->(1)

But if i use geometry in uploaded figure

$c = a+ f(a) cot \theta$

where $\cot \theta = (1/slope ) = (1/tan \theta) = \frac{(f(b)-f(a))}{(b-a)}$

here $\theta = \angle ACB =$ slope of line AC = $ \tan^{-1}\frac{(f(b)-f(a))}{(b-a)}$

this means $c=a+f(a)/\tan \theta = \fbox {$a+f(a) \frac{b-a}{f(b)-f(a)}$}$ ---> (2)

If we compare (1) and (2) equations, i am getting sign change i..e, a+() and a-()

But in textbook regular falsi method final equation is given by (1) and not (2).

Pls point me to where i went wrong.enter image description here

1

There are 1 best solutions below

4
On

You did not consider that $f(a)$ is negative. Thus your formula should be $$ c=a+(-f(a))\cot θ $$ which gives back the first variant of the midpoint formula.


This is not only the regula falsi formula, this secant root formula is also used in the secant method, Dekker's, Brent's method etc.