Taylor Polynomial & Max-Min Problems
104003 Differential and Integral Calculus I
Technion International School of Engineering 2010-11
Tutorial Summary – February 16, 2011 – Kayla Jacobs
Taylor Polynomial
If function
f(x) can be differentiated (at least)
n times in the neighborhood of point
x =
a,
then the n
th
-degree Taylor polynomial of f(x) at x = a is:
ܶ
ሺݔ, ܽሻ ൌ ݂ሺܽሻ
݂
ᇱ
ሺܽሻ
1! · ሺݔ െ ܽሻ
݂
ᇱᇱ
ሺܽሻ
2! · ሺݔ െ ܽሻ
ଶ
݂
ሺଷሻ
ሺܽሻ
3! · ሺݔ െ ܽሻ
ଷ
…
݂
ሺሻ
ሺܽሻ
݊! · ሺݔ െ ܽሻ
ൌ
݂
ሺሻ
ሺܽሻ
݇! · ሺݔ െ ܽሻ
ୀ
This is the best possible n-degree approximation of f(x) “near”
x =
a
.
The more terms you include (the higher n is), and/or the closer to x = a… the better the approximation.
As
݊ ՜ ∞
, the Taylor polynomial
coverges to the exact function f.
It’s then is called the “Taylor series of f.” (Note then it doesn’t matter how “near” you are to x = a.)
For
݊ ൌ 1, you simply get the
linear approximation we’ve already learned about!
ࢌሺ࢞ሻ ൎ ࢌሺࢇሻ ࢌ
ᇱ
ሺࢇሻ · ሺ࢞ െ ࢇሻ when ݔ ൎ ܽ
Remainder:
ܴ
ሺݔ, ܽሻ is the
n
th
-degree remainder for f(x) at x = a.
This is the error made by the approximation of f as a Taylor polynomial:
ܴ
ሺݔ, ܽሻ ൌ ݂ሺݔሻ െ ܶ
ሺݔ, ܽሻ
ൌ
݂
ሺାଵሻ
ሺݖሻ
ሺ݊ 1ሻ! ሺݔ െ ܽሻ
ାଵ
Maclaurin Polynomial
Special case of Taylor polynomial, when
a = 0.
ܶ
ሺݔ, 0ሻ ൌ ∑
ሺೖሻ
ሺሻ
!
· ݔ
ୀ
Common Maclaurin series (as
݊ ՜ ∞ሻ
:
•
ଵ
ଵି௫
ൌ 1 ݔ ݔ
ଶ
ݔ
ଷ
ڮ ൌ ∑
ݔ
ஶ
ୀ
(for
|ݔ| ൏ 1ሻ
•
ଵ
ଵା௫
ൌ 1 െ ݔ ݔ
ଶ
െ ݔ
ଷ
ڮ ൌ ∑ ሺെ1ሻ
· ݔ
ஶ
ୀ
(for
|ݔ| ൏ 1ሻ
•
sinሺݔሻ ൌ ݔ െ
௫
య
ଷ!
௫
ఱ
ହ!
െ ڮ ൌ ∑
ሺିଵሻ
ೖ
ሺଶାଵሻ!
ஶ
ୀ
· ݔ
ଶାଵ
(for all real
ݔ)
•
cosሺݔሻ ൌ 1 െ
௫
మ
ଶ!
௫
ర
ସ!
െ ڮ ൌ ∑
ሺିଵሻ
ೖ
ሺଶሻ!
ஶ
ୀ
· ݔ
ଶ
(for all real
ݔ)
•
݁
௫
ൌ 1 ݔ
௫
మ
ଶ!
௫
య
ଷ!
ڮ ൌ ∑
௫
ೖ
!
ஶ
ୀ
(for all real
ݔ)
DeMoivre’s Theorem
݁
௫
ൌ cosሺݔሻ ݅ · sin ሺݔሻ
(derivable from the Maclaurin series for sin(x), cos(x), and
݁
௫
)
… where z is a number between a and x.
Calculus – Tutorial Summary – February 16, 2011
2
Maclaurin Polynomial Example: sin(x)
The graphs below show the actual function,
sinሺݔሻ, and its Maclaurin polynomial for various values of
n.
As
݊ ՜ ∞, the Maclaurin series is
:
sinሺݔሻ ൌ ݔ െ
௫
య
ଷ!
௫
ఱ
ହ!
െ ڮ ൌ ∑
ሺିଵሻ
ೖ
ሺଶାଵሻ!
ஶ
ୀ
· ݔ
ଶାଵ
Note that there are no even-n terms, so
ܶ
ଶ
ሺݔሻ ൌ ܶ
ଶିଵ
ሺݔሻ
n = 1
ܶ
ଵ
ሺݔሻ ൌ ݔ
n = 3
ܶ
ଷ
ሺݔሻ ൌ ݔ െ
ݔ
ଷ
3!
n = 5
ܶ
ହ
ሺݔሻ ൌ ݔ െ
ݔ
ଷ
3!
ݔ
ହ
5!
… n = 15
ܶ
ଵହ
ሺݔሻ ൌ ݔ െ
ݔ
ଷ
3!
ݔ
ହ
5! െ
ݔ
7!
ݔ
ଽ
9! െ
ݔ
ଵଵ
11!
ݔ
ଵଷ
13! െ
ݔ
ଵହ
15!
Now take a look at n = 99. It agrees very well with the actual
sinሺݔሻ curve until you get out to about x = 35,
when it goes crazy. Here is a picture of the x-interval (34, 39):
(Remember, the polynomial is “centered” at
a = 0, so x = 35 is quite far away from the center.)
Max-Min Problems
x=33
x=39
Calculus – Tutorial Summary – February 16, 2011
3
(Adapted from
http://www.math.ucdavis.edu/~kouba/CalcOneDIRECTORY/maxmindirectory/MaxMin.html
)
1. Read each problem slowly and carefully. If you misread the problem or hurry through it, you
have NO chance of solving it correctly.
2. If appropriate, draw a sketch or diagram of the problem to be solved. Pictures are a great help in
organizing and sorting out your thoughts.
3. Define variables to be used and carefully label your picture or diagram with these variables. This
step is very important because it leads directly or indirectly to the creation of mathematical
equations.
4. Write down all equations which are related to your problem or diagram. Clearly denote that
equation which you are asked to maximize or minimize.
• Experience will show you that MOST optimization problems will begin with two equations.
One equation is a "constraint" equation and the other is the "optimization" equation.
• The "constraint" equation is used to solve for one of the variables. This is then
substituted into the "optimization" equation before differentiation occurs.
• Some problems may have NO constraint equation. Some problems may have two or more
constraint equations.
5. Before differentiating, make sure that the optimization equation is a function of only one
variable.
6. Differentiate using the well-known rules of differentiation. Solve for the variable value(s) that
satisfy the derivative being set to 0.
6. Verify that your result is a maximum or minimum value using the first or second derivative
test for extrema.
7. If appropriate, don’t forget to check the endpoints, which might be the global
maximum/minimum.