Learning and Working With Generalized Functions Can Be Fun: F. Farassat NASA Langley Research Center Hampton, Virginia
Learning and Working With Generalized Functions Can Be Fun: F. Farassat NASA Langley Research Center Hampton, Virginia
Learning and Working With Generalized Functions Can Be Fun: F. Farassat NASA Langley Research Center Hampton, Virginia
With Generalized
Functions Can Be Fun
F. Farassat
NASA Langley Research Center
Hampton, Virginia
Talk Presented at
Penn State Engineering Seminar
April 10, 2006
F. Farassat
Outline of the Talk
What is an ordinary function?
What is a generalized function (GF)?
How do we extend the definition of a
function?
A brief definition of GFs
Differentiation of GFs
Some applications of GFs
Concluding remarks
F. Farassat 2 of 11
What is an Ordinary Function (OF)?
We think of an ordinary function such
as sin x or ex as a table of ordered
pairs of numbers (x, sin x) or (x, ex).
This means that we think as ordinary
functions pointwise. A graph is a way
is illustrating this pointwise view of a
function.
The functions engineers learn in
calculus are all ordinary functions.
These functions are not enough for
what the engineers need later
F. Farassat 3 of 11
What is a Generalized Function (GF)?
( x ) ( x ) dx = ( 0 )
We can show that there is no ordinary
function with this property. However,
this function is so important in
mathematics that mathematicians had
to find a way to define it rigorously. It
took about 25 years to do this!
F. Farassat 4 of 11
How Do We Extend the
Definition of a Function?
We must change the way we think of
functions from pointwise (ordered pair
of numbers) to one that allows new
objects or functions. This process of
extension of the definition of ordinary
functions is difficult and most
mathematicians never have to do this.
F. Farassat 5 of 11
How Do We Extend the Definition
of a Function? (Contd)
When we extend the definition of a
function, we require two things:
All ordinary functions must be
included in the extension
All the operations on ordinary
functions must also apply to the new
functions
The most important thing to know
about extension of ordinary functions
to GFs is that we can solve new
problems of engineering and science.
F. Farassat 6 of 11
A Brief Definition of GFs
We take some well-behaved functions
which we call test functions (TFs). For an
OF f ( x ) , we define the following mapping
(functional) of TFs to real numbers
F [] = f ( x ) ( x ) dx , ( x ) is a TF
This mapping is linear:
F [ 1 + 2 ] = F [ 1 ] + F [ 2 ]
and continuous: F [ n ] 0 if n 0
F. Farassat 7 of 11
Models of Functions
Old (Conventional) Model: We think of a
f(x)
function as a table of ordered pairs (x, f(x))
where for each x, f(x) is unique. This table (xo, f(xo))
can be graphed as shown and usually has an
uncountable number of ordered pairs.
xo x
( x ) ( x x ) dx = ( x ) ,
0 0
( x ) ( x ) dx = (0)
----------- , e 2ix ( x ) = 1
F. Farassat 9 of 11
Differentiation of Generalized Functions
All test functions are in D.
f(x) ordinary function, differentiable, F [ ] = f dx , we must identify F [ ]
with f dx . But F [ ] f dx = f dx = F [ ] since D .
Therefore, we use the relation:
F [ ] = F [ ]
as the definition of derivative of any generalized function F[]. Similarly
F ( n ) [ ] = ( 1 ) n F [ ( n ) ] , i.e., generalized functions have derivatives of all
orders.
Examples:
i) [ ] = [ ] = ( 0 ) or ( x ) ( x ) dx = ( 0 )
ii) [ ] = ( 1 ) 2 [ ] = ( 0 ) or ( x ) ( x ) dx = ( 0 )
Note: If an ordinary function is differentiable on real line, then f gen. = f .
However, generalized derivative of an ordinary function can be a singular
generalized function.
20 of 95, September 1996
F. Farassat, Aeroacoustics Branch
Fluid Mechanics and Acoustics Division
NASA Langley Research Center
Hampton, Virginia
Differentiation of Generalized Functions (Contd)
df
Notation: For ordinary (regular G.F.s) functions, we use f ( x ) or ------ for
f gen. to distinguish generalized from ordinary derivative. dx
x 0-
= + f dx = f dx + f ( x 0 )
x 0+
or symbolically
f ( x ) = f ( x ) + f ( x x 0 )
F. Farassat 10 of 11
Generalized Functions in Multidimensions (Contd)
Generalized functions in n dimensions are continuous linear functionals on n
dimensional test function space D.
Examples:
i) ( x ) ( x ) dx = (0)
ii) ------- ( x ) ( x ) dx = ------- ( 0 )
x i x i
From our point of view, the most important generalized functions are delta
functions whose supports are on open or closed surfaces, e.g., (f). We need to
interpret integrals of the form
I1 = ( f ) ( x ) dx and I 2 = ( f ) ( x ) dx .
g = g ( f = 0+ ) g ( f = 0 ) f(x) = 0
Since u 3 = f , we have g = g + gf ( f ) .
g = g + g f ( f )
Similarly
g = g + g f ( f )
31 of 95, September 1996
F. Farassat, Aeroacoustics Branch
Fluid Mechanics and Acoustics Division
NASA Langley Research Center
Hampton, Virginia
The Integration of (f) and (f)
We assume f ( x ) is defined such that f = 1 on the surface f = 0. This can
always be done. This means df = dn = du3
Parametrize the space in vicinity of surface f = 0 by variables (u1, u2, u3) as
shown. Then
u3 distance
from f = 0
= g ( 2 ) ( u 1, u 2, u 3 ) du 1 du 2 du 3
I1 = ( x ) ( u 3 ) g ( 2 ) du 1 du 2 du 3 = [ ( x ) ]u3 = 0 g ( 2 ) du 1 du 2
I1 = ( x ) ( f ) dx = ( x ) ds
f =0
F. Farassat 11 of 11