differentiation 2 first principles - university of...
TRANSCRIPT
Differentiation 2first principles
J A Rossiter
1
Slides by Anthony Rossiter
Introduction
• The previous video introduces the concept of differentiation and the term derivative.
• Next we need to look at how differentiation is performed and the derivative computed.
• The focus here is on 1st principles, that is to show, briefly, how the main results are derived.
• Students who are happy to go straight to core results without understanding the origins can skip this resource and go straight to resources 6 to get into some computations.
Slides by Anthony Rossiter
2
Recap
• Differentiation means to find the gradient; in general this involves some mathematical operations.
• A derivative is the result of differentiation, that is a function defining the gradient of a curve.
• The notation of derivative uses the letter ‘d’ and is not a fraction!
Slides by Anthony Rossiter
3
dx
dfderivative
dx
dyxfy )(
Spoken as ‘d f d x’.
)( fdx
d Spoken as ‘d d x of f’.The action of differentiation.
What is differentiation?Differentiation is a process which finds the gradient of a curve, precisely, at any point along the curve.
Slides by Anthony Rossiter
4
-2 -1.5 -1 -0.5 0 0.5 1-4
-3
-2
-1
0
1
2
x-2 -1.5 -1 -0.5 0 0.5 1
-4
-3
-2
-1
0
1
2
x
dy/dx= 1.75 at this point.
12 23 xxxy
dy/dx = -0.25 at this point.
dy/dx = 3.75 at this point.
First principles – gradient estimation
For a general curve, the gradient can be estimated using the formulae:
Slides by Anthony Rossiter
5
0.4 0.5 0.6 0.7 0.8 0.9 1-1.5
-1
-0.5
0
0.5
1
1.5
2
x
xinchange
yinchangegradient
12
12 )()(
xx
xyxygradient
This is close, if difference between the x-values is small.
Clearly not exact gradient.
0 0.5 1 1.5 2 2.5 3-10
0
10
20
30
40
50
x
First principles – gradient estimation
For a general curve, the gradient can be estimated using the formulae:
Slides by Anthony Rossiter
6
12
12 )()(
xx
xyxygradient
This is close, if difference between the x-values is small.
Clearly not exact gradient.
As difference gets smaller, the approximation
becomes more accurate.
0 0.5 1 1.5 2 2.5 3-10
0
10
20
30
40
50
x
First principles – gradient estimationFor a general curve, the gradient can be computed as a limiting value:
Slides by Anthony Rossiter
7
xxx
xyxxy
dx
dyx
)(
)()(lim 0
Clearly, the smaller δx, the more accurate the
gradient estimate.
)](,[ xxyxx
x)](,[ xyx
CaviatWe will not dwell on mathematical subtleties, but users need to assume the limit exists and is well defined.
For many curves, this limit is not unique or well defined at some points and consequently, at those points differentiation is not uniquely defined.
Slides by Anthony Rossiter
8
xxx
xyxxy
dx
dyx
)(
)()(lim 0
EXAMPLES OF USING FIRST PRINCIPLES TO DERIVE DERIVATIVES OF SOME COMMON FUNCTIONS
Slides by Anthony Rossiter
9
xxx
xyxxy
dx
dyx
)(
)()(lim 0
Example 1
Simply substitute into the formula from the previous page.
Slides by Anthony Rossiter
10
2xy
xxx
xyxxy
dx
dyx
)(
)()(lim 0
xxx
xxx
dx
dyx
)(
)(lim
22
0
xx
xxx
dx
dyx 2
))(2(lim 0
-2 -1 0 1 2 30
1
2
3
4
5
6
7
8
9
xVisual inspection validates this answer
is sensible.
x
xxxxx
dx
dyx
222
0
))(2(lim
Example 2
Simply substitute into the formula.
Slides by Anthony Rossiter
11
3xy
xxx
xyxxy
dx
dyx
)(
)()(lim 0
x
xxxxxxx
dx
dy
xxx
xxx
dx
dy
x
x
33223
0
33
0
))()(33(lim
)(
)(lim
222
0 3))()(33(
lim xx
xxxxx
dx
dyx
Visual inspection validates
this answer is sensible.
Example 3
Simply substitute into the formula.
Slides by Anthony Rossiter
12
nxy
x
xxnxx
dx
dy
xxx
xxx
dx
dy
nnn
x
nn
x
)(lim
)(
)(lim
1
0
0
11
0
)(lim
nn
x nxx
xnx
dx
dy
Ignore higher order terms in δx as these
go to zero.
Example 4
Simply substitute into the formula.
Slides by Anthony Rossiter
13
)sin(axy
x
axxaaxxaax
dx
dy
x
axxaax
dx
dy
x
x
)sin()sin()cos()cos()sin(lim
)sin()sin(lim
0
0
;)sin(lim;1)cos(lim 00 xaxaxa xx
)cos()cos(
lim 0 axax
xaax
dx
dyx
x
axxaaxax
dx
dyx
)sin()cos(1)sin(lim 0
Example 5
Simply substitute into the formula.
Slides by Anthony Rossiter
14
bxey
x
ee
dx
dy
x
ee
dx
dy
xbbx
x
bxxxb
x
)1(lim
lim
0
)(
0
;)1(lim 0 xbe xb
x
bxbx
x bex
xbe
dx
dy
0lim
Example 6
This one is easiest handled by recognising the following relationship; this is obvious as it amounts to a simple swapping of the axis.
Slides by Anthony Rossiter
15
axy log
y
y
e
a
dx
dye
ady
dx
1
yea
xaxy1
)log(
xax
a
e
a
dx
dyy
1
11
dy
dx
dx
dyor
dydxdx
dy
Table of some common results
Slides by Anthony Rossiter
16
adx
dyaxy 1 nn nax
dx
dyaxy
)cos()sin( bxbdx
dybxy )sin()cos( bxb
dx
dybxy
)(sec)tan( 2 bxbdx
dybxy cxcx ce
dx
dyey
xdx
dyxy
1log
)cosh()sinh( bxbdx
dybxy )sinh()cosh( bxb
dx
dybxy
)cot()sin(
1
)sin(
1x
bxb
dx
dy
bxy
Summary
• This video has introduced differentiation using first principles derivations.
• The derivatives of a few common functions have been given.
• Readers can use the same procedures to find derivatives for other functions but in general it is more sensible to access a table of answers which have been derived for you.
• Later videos will gradually introduce known formulae and their application.
Slides by Anthony Rossiter
17
© 2016 University of Sheffield
This work is licensed under the Creative Commons Attribution 2.0 UK: England & Wales Licence. To view a copy of this licence, visit http://creativecommons.org/licenses/by/2.0/uk/ or send a letter to: Creative Commons, 171 Second Street, Suite 300, San Francisco, California 94105, USA.
It should be noted that some of the materials contained within this resource are subject to third party rights and any copyright notices must remain with these materials in the event of reuse or repurposing.
If there are third party images within the resource please do not remove or alter any of the copyright notices or website details shown below the image.
(Please list details of the third party rights contained within this work.
If you include your institutions logo on the cover please include reference to the fact that it is a trade mark and all copyright in that image is reserved.)
Anthony RossiterDepartment of Automatic Control and
Systems EngineeringUniversity of Sheffieldwww.shef.ac.uk/acse