Sequences in Hilbert space
Clash Royale CLAN TAG#URR8PPP
up vote
0
down vote
favorite
Let $A $ be an operator acting on an infinite dimensional complex Hibert space $ H $. Let $W(A)$ be
the set of all $langle Ax,x rangle $ with $xin H,~Vert x Vert=1$ and
$ w(A)=displaystyle sup_zin W(A)vert zvert. $
Let $(x_n)$ be a sequence of unit vectors in $H$ such that $$ displaystyle lim_nvertlangle Ax_n,x_n ranglevert=w(A).$$ Does exist a sequence $(y_n)$ of unit vectors in $H$ such that
$ langle x_n , y_n rangle=0 $ for all $n$ and
$$ displaystyle lim_nvert langle Ay_n,y_n ranglevert= w(A)?$$
Our approach is based on the following. By the Berberian-Orland trick,
there exist a complex Hilbert space $K supset H $ and an operator $A^o$ acting on $K$ with $A^o_vertH=A$ and $W(A^o)=adh(W(A))$, so we have $w(A)=w(A^o)$.
Put $$ displaystyle lim_nlangle Ax_n,x_n rangle=langle A^ox,x rangle_K quad and quad displaystyle lim_nlangle Ay_n,y_n rangle=langle A^oy,y rangle_K, $$ with $xin K$, $yin K$ and $Vert x Vert_K=Vert yVert_K=1$. Therefore, it suffices to show that $langle x,y rangle_K=0$.
If it is the case, since $langle A^ox,x rangle in W(A^o)cap partial W(A^o) $, then we get $vertlangle A^ox,x rangle_Kvert=vertlangle A^oy,y rangle_Kvert$. That is $ displaystyle lim_nvertlangle Ay_n,y_n ranglevert=w(A)$.
functional-analysis
add a comment |Â
up vote
0
down vote
favorite
Let $A $ be an operator acting on an infinite dimensional complex Hibert space $ H $. Let $W(A)$ be
the set of all $langle Ax,x rangle $ with $xin H,~Vert x Vert=1$ and
$ w(A)=displaystyle sup_zin W(A)vert zvert. $
Let $(x_n)$ be a sequence of unit vectors in $H$ such that $$ displaystyle lim_nvertlangle Ax_n,x_n ranglevert=w(A).$$ Does exist a sequence $(y_n)$ of unit vectors in $H$ such that
$ langle x_n , y_n rangle=0 $ for all $n$ and
$$ displaystyle lim_nvert langle Ay_n,y_n ranglevert= w(A)?$$
Our approach is based on the following. By the Berberian-Orland trick,
there exist a complex Hilbert space $K supset H $ and an operator $A^o$ acting on $K$ with $A^o_vertH=A$ and $W(A^o)=adh(W(A))$, so we have $w(A)=w(A^o)$.
Put $$ displaystyle lim_nlangle Ax_n,x_n rangle=langle A^ox,x rangle_K quad and quad displaystyle lim_nlangle Ay_n,y_n rangle=langle A^oy,y rangle_K, $$ with $xin K$, $yin K$ and $Vert x Vert_K=Vert yVert_K=1$. Therefore, it suffices to show that $langle x,y rangle_K=0$.
If it is the case, since $langle A^ox,x rangle in W(A^o)cap partial W(A^o) $, then we get $vertlangle A^ox,x rangle_Kvert=vertlangle A^oy,y rangle_Kvert$. That is $ displaystyle lim_nvertlangle Ay_n,y_n ranglevert=w(A)$.
functional-analysis
Where is the question? Is this a "proof verification" question?
â b00n heT
Aug 15 at 12:42
Might be stupid consideration, but in the simplest case i was able to immagine $A : mathbbC to mathbbC$ such that $A(z) = 2z$ this should be impossible because there are no couple of ortogonal vector with unitary norm
â JayTuma
Aug 15 at 12:44
We ask if the sequence $(y_n)$ exists!
â A. Bag
Aug 15 at 12:51
add a comment |Â
up vote
0
down vote
favorite
up vote
0
down vote
favorite
Let $A $ be an operator acting on an infinite dimensional complex Hibert space $ H $. Let $W(A)$ be
the set of all $langle Ax,x rangle $ with $xin H,~Vert x Vert=1$ and
$ w(A)=displaystyle sup_zin W(A)vert zvert. $
Let $(x_n)$ be a sequence of unit vectors in $H$ such that $$ displaystyle lim_nvertlangle Ax_n,x_n ranglevert=w(A).$$ Does exist a sequence $(y_n)$ of unit vectors in $H$ such that
$ langle x_n , y_n rangle=0 $ for all $n$ and
$$ displaystyle lim_nvert langle Ay_n,y_n ranglevert= w(A)?$$
Our approach is based on the following. By the Berberian-Orland trick,
there exist a complex Hilbert space $K supset H $ and an operator $A^o$ acting on $K$ with $A^o_vertH=A$ and $W(A^o)=adh(W(A))$, so we have $w(A)=w(A^o)$.
Put $$ displaystyle lim_nlangle Ax_n,x_n rangle=langle A^ox,x rangle_K quad and quad displaystyle lim_nlangle Ay_n,y_n rangle=langle A^oy,y rangle_K, $$ with $xin K$, $yin K$ and $Vert x Vert_K=Vert yVert_K=1$. Therefore, it suffices to show that $langle x,y rangle_K=0$.
If it is the case, since $langle A^ox,x rangle in W(A^o)cap partial W(A^o) $, then we get $vertlangle A^ox,x rangle_Kvert=vertlangle A^oy,y rangle_Kvert$. That is $ displaystyle lim_nvertlangle Ay_n,y_n ranglevert=w(A)$.
functional-analysis
Let $A $ be an operator acting on an infinite dimensional complex Hibert space $ H $. Let $W(A)$ be
the set of all $langle Ax,x rangle $ with $xin H,~Vert x Vert=1$ and
$ w(A)=displaystyle sup_zin W(A)vert zvert. $
Let $(x_n)$ be a sequence of unit vectors in $H$ such that $$ displaystyle lim_nvertlangle Ax_n,x_n ranglevert=w(A).$$ Does exist a sequence $(y_n)$ of unit vectors in $H$ such that
$ langle x_n , y_n rangle=0 $ for all $n$ and
$$ displaystyle lim_nvert langle Ay_n,y_n ranglevert= w(A)?$$
Our approach is based on the following. By the Berberian-Orland trick,
there exist a complex Hilbert space $K supset H $ and an operator $A^o$ acting on $K$ with $A^o_vertH=A$ and $W(A^o)=adh(W(A))$, so we have $w(A)=w(A^o)$.
Put $$ displaystyle lim_nlangle Ax_n,x_n rangle=langle A^ox,x rangle_K quad and quad displaystyle lim_nlangle Ay_n,y_n rangle=langle A^oy,y rangle_K, $$ with $xin K$, $yin K$ and $Vert x Vert_K=Vert yVert_K=1$. Therefore, it suffices to show that $langle x,y rangle_K=0$.
If it is the case, since $langle A^ox,x rangle in W(A^o)cap partial W(A^o) $, then we get $vertlangle A^ox,x rangle_Kvert=vertlangle A^oy,y rangle_Kvert$. That is $ displaystyle lim_nvertlangle Ay_n,y_n ranglevert=w(A)$.
functional-analysis
edited Aug 15 at 12:52
asked Aug 15 at 12:35
A. Bag
694
694
Where is the question? Is this a "proof verification" question?
â b00n heT
Aug 15 at 12:42
Might be stupid consideration, but in the simplest case i was able to immagine $A : mathbbC to mathbbC$ such that $A(z) = 2z$ this should be impossible because there are no couple of ortogonal vector with unitary norm
â JayTuma
Aug 15 at 12:44
We ask if the sequence $(y_n)$ exists!
â A. Bag
Aug 15 at 12:51
add a comment |Â
Where is the question? Is this a "proof verification" question?
â b00n heT
Aug 15 at 12:42
Might be stupid consideration, but in the simplest case i was able to immagine $A : mathbbC to mathbbC$ such that $A(z) = 2z$ this should be impossible because there are no couple of ortogonal vector with unitary norm
â JayTuma
Aug 15 at 12:44
We ask if the sequence $(y_n)$ exists!
â A. Bag
Aug 15 at 12:51
Where is the question? Is this a "proof verification" question?
â b00n heT
Aug 15 at 12:42
Where is the question? Is this a "proof verification" question?
â b00n heT
Aug 15 at 12:42
Might be stupid consideration, but in the simplest case i was able to immagine $A : mathbbC to mathbbC$ such that $A(z) = 2z$ this should be impossible because there are no couple of ortogonal vector with unitary norm
â JayTuma
Aug 15 at 12:44
Might be stupid consideration, but in the simplest case i was able to immagine $A : mathbbC to mathbbC$ such that $A(z) = 2z$ this should be impossible because there are no couple of ortogonal vector with unitary norm
â JayTuma
Aug 15 at 12:44
We ask if the sequence $(y_n)$ exists!
â A. Bag
Aug 15 at 12:51
We ask if the sequence $(y_n)$ exists!
â A. Bag
Aug 15 at 12:51
add a comment |Â
1 Answer
1
active
oldest
votes
up vote
0
down vote
This is not possible. Take $H=l^2$, define
$$
Ax = x_1e_1,
$$
then $langle Ax,xrangle = |x_1|^2$. Let $(x_n)$ be a sequence of unit vectors such that $langle Ax_n,x_nrangle to 1$, which implies $|x_n,1|to 1$.
Since $|x_n|=1$, it follows $|x_n -x_n,1e_1|to0$. Let $(y_n)$ be another sequence with these properties. Then clearly $|langle x_n,y_nrangle|to1$:
$$
langle x_n,y_nrangle = x_n,1y_n,1 +langle x_n -x_n,1e_1,x_n -x_n,1e_1rangle,
$$
which converges to a limit with absolute value $1$, as the second addend is the inner product of two sequences converging to zero.
Why $ |langle x_n,y_nrangle|to1$? to have $langle x,y rangle_K=0 $ in my suggestion is a sufficient condition not necessary.
â A. Bag
Aug 15 at 14:43
add a comment |Â
1 Answer
1
active
oldest
votes
1 Answer
1
active
oldest
votes
active
oldest
votes
active
oldest
votes
up vote
0
down vote
This is not possible. Take $H=l^2$, define
$$
Ax = x_1e_1,
$$
then $langle Ax,xrangle = |x_1|^2$. Let $(x_n)$ be a sequence of unit vectors such that $langle Ax_n,x_nrangle to 1$, which implies $|x_n,1|to 1$.
Since $|x_n|=1$, it follows $|x_n -x_n,1e_1|to0$. Let $(y_n)$ be another sequence with these properties. Then clearly $|langle x_n,y_nrangle|to1$:
$$
langle x_n,y_nrangle = x_n,1y_n,1 +langle x_n -x_n,1e_1,x_n -x_n,1e_1rangle,
$$
which converges to a limit with absolute value $1$, as the second addend is the inner product of two sequences converging to zero.
Why $ |langle x_n,y_nrangle|to1$? to have $langle x,y rangle_K=0 $ in my suggestion is a sufficient condition not necessary.
â A. Bag
Aug 15 at 14:43
add a comment |Â
up vote
0
down vote
This is not possible. Take $H=l^2$, define
$$
Ax = x_1e_1,
$$
then $langle Ax,xrangle = |x_1|^2$. Let $(x_n)$ be a sequence of unit vectors such that $langle Ax_n,x_nrangle to 1$, which implies $|x_n,1|to 1$.
Since $|x_n|=1$, it follows $|x_n -x_n,1e_1|to0$. Let $(y_n)$ be another sequence with these properties. Then clearly $|langle x_n,y_nrangle|to1$:
$$
langle x_n,y_nrangle = x_n,1y_n,1 +langle x_n -x_n,1e_1,x_n -x_n,1e_1rangle,
$$
which converges to a limit with absolute value $1$, as the second addend is the inner product of two sequences converging to zero.
Why $ |langle x_n,y_nrangle|to1$? to have $langle x,y rangle_K=0 $ in my suggestion is a sufficient condition not necessary.
â A. Bag
Aug 15 at 14:43
add a comment |Â
up vote
0
down vote
up vote
0
down vote
This is not possible. Take $H=l^2$, define
$$
Ax = x_1e_1,
$$
then $langle Ax,xrangle = |x_1|^2$. Let $(x_n)$ be a sequence of unit vectors such that $langle Ax_n,x_nrangle to 1$, which implies $|x_n,1|to 1$.
Since $|x_n|=1$, it follows $|x_n -x_n,1e_1|to0$. Let $(y_n)$ be another sequence with these properties. Then clearly $|langle x_n,y_nrangle|to1$:
$$
langle x_n,y_nrangle = x_n,1y_n,1 +langle x_n -x_n,1e_1,x_n -x_n,1e_1rangle,
$$
which converges to a limit with absolute value $1$, as the second addend is the inner product of two sequences converging to zero.
This is not possible. Take $H=l^2$, define
$$
Ax = x_1e_1,
$$
then $langle Ax,xrangle = |x_1|^2$. Let $(x_n)$ be a sequence of unit vectors such that $langle Ax_n,x_nrangle to 1$, which implies $|x_n,1|to 1$.
Since $|x_n|=1$, it follows $|x_n -x_n,1e_1|to0$. Let $(y_n)$ be another sequence with these properties. Then clearly $|langle x_n,y_nrangle|to1$:
$$
langle x_n,y_nrangle = x_n,1y_n,1 +langle x_n -x_n,1e_1,x_n -x_n,1e_1rangle,
$$
which converges to a limit with absolute value $1$, as the second addend is the inner product of two sequences converging to zero.
edited Aug 15 at 16:28
answered Aug 15 at 14:09
daw
22k1542
22k1542
Why $ |langle x_n,y_nrangle|to1$? to have $langle x,y rangle_K=0 $ in my suggestion is a sufficient condition not necessary.
â A. Bag
Aug 15 at 14:43
add a comment |Â
Why $ |langle x_n,y_nrangle|to1$? to have $langle x,y rangle_K=0 $ in my suggestion is a sufficient condition not necessary.
â A. Bag
Aug 15 at 14:43
Why $ |langle x_n,y_nrangle|to1$? to have $langle x,y rangle_K=0 $ in my suggestion is a sufficient condition not necessary.
â A. Bag
Aug 15 at 14:43
Why $ |langle x_n,y_nrangle|to1$? to have $langle x,y rangle_K=0 $ in my suggestion is a sufficient condition not necessary.
â A. Bag
Aug 15 at 14:43
add a comment |Â
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2883547%2fsequences-in-hilbert-space%23new-answer', 'question_page');
);
Post as a guest
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Where is the question? Is this a "proof verification" question?
â b00n heT
Aug 15 at 12:42
Might be stupid consideration, but in the simplest case i was able to immagine $A : mathbbC to mathbbC$ such that $A(z) = 2z$ this should be impossible because there are no couple of ortogonal vector with unitary norm
â JayTuma
Aug 15 at 12:44
We ask if the sequence $(y_n)$ exists!
â A. Bag
Aug 15 at 12:51