and
on the domain
and then on the domain
.
What do you conclude ?

. Where can
we restrict
so that it has a well defined inverse
function? (We'd like as big a set as possible...)
Can you find an argument without the computer for
why this is true ?
. Estimate again
where we can restrict
so that it has a well defined inverse
function? (A non-computer assisted argument is much harder here ...)
.
Using the ``Draw Tangent'' feature of Analyzer*, estimate
those points x where
has the slope of its tangent
line equal to -1, 0, or 1. Now check your estimates by
graphing
.
and
.
. For which initial values of x
would one expect Newton's method to have difficulty at finding the
roots of this function ?
for
constants a, b, and c. How does the
graph look for different values of a, b, and c ? How does the
sign of the discriminant
affect the appearance of the
graph ?
whose exact value is easy to compute.
Some of the the numerical integration methods which Analyzer*
supports include left-hand, midpoint, the trapezoid rule, and
Simpson's rule. For each of these determine how many subdivisions
you need to estimate the answer with no more than a 1% error.