Quadratic Equations
We now turn to the task of finding an iterator G(x) whose fixed points are the solutions of ax2 + bx + c = 0, ideally one with slope zero at those fixed points. Here we assume that b2 - 4ac > 0, thereby assuring the existence of real solutions.
In the special case
a = 1, b = 0, and c = -k
we have found the desired iterator to be
|
A simple calculation
|
shows that the square roots of k are in fact fixed points of F(x), while the A-G Mean inequality can be used to confirm that its graph is flat at these fixed points.
Against this background, we begin our efforts to solve ax2 + bx + c = 0 by setting b=0 and considering the quadratic equation
ax2 + c = 0.
With a little algebra, this reduces to the previous case . For by writing this equation as x2 - (-c/a) = 0 and setting k = -c/a, we are (for the case b = 0) led to the iterator
|
Exercise. By solving G(x) = x, confirm that x is a fixed point of G(x) whenever ax2 + c = 0.
Having found the desired iterator for the case b = 0, we turn to the general quadratic equation: ax2 + bx + c = 0. Our approach will be one of modifying the iterator G(x) as given above. Here it is readily confirmed that we could modify the numerator of G to arrive at the iterator
|
that has the "right fixed points" - i.e., if ax2 + bx + c = 0, then G1(x) = x. Similarly, we could modify the denominator of G to arrive at the iterator
|
|
Well, we have arrived at two iterators with the right fixed points (and there are other possibilities as well). To choose between them we could compare their effectiveness in approximating the roots of a particular quadratic equation, say x2 - x - 1 = 0. Another possibility is to look at the graphs of both iterators and determine if one of these has slope zero at its fixed points.
Go to the NEXT PAGE
Return to the PRECEDING PAGE