Lecture 22
Lecture 22
Lecture 22
Recall that the Newton quotient is the slope of the secant between two nearby points (a, f (a)) and
(a + h, f (a + h)) on the graph of a function.
Such a secant is probably a reasonable approximation to the graph near these two points but if you want to
approximate f near x = a, what value of h should you choose?
It makes sense to find the limit as h → 0, which results in the tangent line to the graph at x = a.
The function defining the tangent line is a linear approximation to f at x = a.
Question
What is the slope of the tangent line?
What is the formula for the tangent line?
What is special about the tangent line compared with all other lines?
Definition and theorem about linear approximations
Definition 1
Let f be a function that is differentiable at a. Then the linear approximation to f based at a (or centered at
a) is denoted Lfa and is defined by
Lfa (x) = f (a) + f 0 (a)(x − a).
The point a is referred to as the basepoint or center of the linear approximation.
Remark: If it is clear from context which function we are approximating, we may write simply La for Lfa .
Theorem 1
Let I be an open interval containing the point a and suppose that the function f : I → R is differentiable at a.
Let La be the linear approximation to f based at a. Then there exists a function Qa satisfying lim Qa (h) = 0
h→0
such that
f (a + h) = La (a + h) + hQa (h)
for all h such that (a + h) ∈ I .
Remark: The term hQa (h) is the remainder or error in approximating f by La . Theorem 1 says that the error is
small in magnitude when |h| is small.
Examples of linear approximations
The well known small angle approximations in trigonometry, sin(x) ≈ x, cos(x) ≈ 1, tan(x) ≈ x, are linear
approximations based at x = 0.
These approximations are very useful in physics, etc.
Exercise
√
Find an approximate value for 17.
Another theorem about linear approximations
The following theorem tells us that if we can find a linear function that is a “good enough” approximation to f
near a, then f must be differentiable at a.
Theorem 2
Let I be an open interval containing the point a and let f be defined on I . Let L be the function defined by
L(x) = f (a) + m(x − a) for some real number m. If there exists a function Qa satisfying lim Qa (h) = 0 such that
h→0
We have seen that a linear approximation is a good way to (locally) approximate differentiable functions.
If we represent the change in function value (relative to the value at a point x) by ∆f = f (x + ∆x) − f (x), then
for small ∆x, we should expect ∆f ≈ Lx (x + ∆x) − Lx (x). This is what we call the differential.
Definition 2
Let f be differentiable on an open interval I . The differential of f at the basepoint x ∈ I and with increment
∆x is denoted df (x, ∆x) and is defined by
Theorem 3
If f is differentiable at x, then