I can't quite put a finger as to why using 'AngleBracket' is so uncomfortable. For one part, one can't really assigned some arbitrary variable to any definition of an expression associated with 'AngleBracket'. It's frustrating.
Unlike any arbitrary function defined by, say, f[x_]:=x
, you can define another by g[x_]:=2x
You can't do the same for an expression defined with AngleBracket
.
I.e.,
p=AngleBracket[f_,g_]:=Integrate[f g,{x,-1,1}]
which makes it possible only to have one expression defined with AngleBracket
possible for if you have
y=AngleBracket[f_,g_]:=Integrate[f g,{x,-2,2}]
you wouldn't be able to evaluate
y=AngleBracket[f_,g_]:=Integrate[f g,{x,-2,2}]
for different f an g variable.
A set of polynomial is $\left(1-x^{2}\right)x^{n-1}$.
A basis set of linearly independent polynomial $p_n\left(x\right)$ for $n=0,1,2,3,4$ that satisfies $p'_{n}\left(\pm1\right)=0$
is
$\left\{p\left(x\right)\right\}_{n=0}^{4}$=Table[Integrate[(1 - x^2) x^(n - 1), x], {n, 0, 4}]
.
I seek to construct an orthonormal set with respect to the inner product defined by
AngleBracket[f_, g_] := Integrate[f g, {x, -1, 1}]
from the above linearly independent set of orthogonal polynomial in the vector space.
To do this, use Orthogonalize
:
Orthogonalize[Table[Integrate[(1 - x^2) x^(n - 1), x], {n, 0, 4}],
p] // Factor
.
However, the result does not follow.
Any input to illuminate what is wrong would be appreciated.
f[x_] := x
andg[x_] := 2x
, you could/should definedAngleBracket[f_, g_] := Integrate[f[x] g[x], {x, -1, 1}]
. Then the callAngleBracket[f, g]
would work for whatever functionsf
andg
were and whatever symbol was used in their definitions, but it would not work for expressions in terms of a variable such asAngleBracket[x, 2x]
. Now it is not really possible, nor should it be, forAngleBracket[]
to represent two different inner products (such as one over the interval{-1, 1}
and one over{-2, 2}
). $\endgroup$