Sunteți pe pagina 1din 1

Inversion and Modelling of Geophysical Data A Note on Norms

October 31, 2003

Norms
In the MathReview we briey talked about norms, and introduced norms in a perhaps rather theoretical context. Due to a question of one of the students, lets try to clarify norms, in the sense how to compute a norm, and how to solve question 3 of exercise # 1. Norms measure length of a vector; how we dene length is a separate issue, but let us think about length as measuring distance between two points. The most general form of a norm is called p-norm:
M
1 p

=
i=1

|xi |p

1 p .

(1)

The notation L1 -norm then means calculating the norm using Eq. [1] with p = 1; similarly, for the L2 -norm p = 2. The way we dene norms in Eq. [1] seems applicable only to vectors where we have a discrete set of values that can be summed. However, one can easily extend the above denition to general functions, dened over some interval [a, b]. The summation in Eq. [1] then becomes an integration, and the functional norm becomes 1
b
p

f (x)

=
a

[|f (x)|]p dx

1 p .

(2)

For question # 3 in the rst homework, you simply have to carry out the integrations for the given functions di (x) in the interval [a, b], i.e. L1 : d(x) L2 : d(x) L : d(x)
1

= =

b a

| f (x) | dx
b [| f (x) |]2 a

dx

1 2

(3)

= maxaxb [| f (x) |]

I hope that helps! Sorry for the confusion, and not being explicit enough in my explanations, and thanks to the one who pointed out that something was missing!

S-ar putea să vă placă și