Skip to main content
Since I manage to partially answer the question, I have to make some corrections in the question.
Source Link

I am reading Moser's paper On commuting circle mappings and simultaneous Diophantine approximations and I found it hard because it is my first time that I seriously have to read a paper. It is a local simultaneous conjugacy result for commuting diffeomorphisms in the neighborhood of rotations.

The main theorem is Theorem 1 which states that if $\phi_k$ are diffeomorphisms of $S^1$ in the neighborhood of rotations $R_{\alpha_k}$ in $C^{\infty}$-topology, $k=1,\ldots,n$, where $\alpha=(\alpha_1,\ldots,\alpha_n)$ is Diophantine vector in $\mathbb R^n$, and if $\phi_k \circ \phi_l=\phi_l \circ \phi_k$, then there is a smooth diffeomorphism of $S^1$ simultaneously conjugating each $\phi_k$ to $R_{\alpha_k}$.

Here is the set-up for the proof of the main theorem. Let $V_0=C^{\infty}_0(S^1,\mathbb R)$, $V_1=C^{\infty}_0(S^1,\mathbb R^n)$ and $V_2=C^{\infty}_0(S^1,so(n))$, where $0$ subscript means that functions have mean value $0$ and $so(n)$ is the Lie algebra of $SO(n)$, i.e. the vector space of all skew-symmetric matrices. Also, let $A \colon V_0 \to V_1$ and $B \colon V_1 \to V_2$ be linear maps given by $(Av)_k=L_kv$ and $(Bw)_{kl}=L_kw_l-L_lw_k$, where $L_kv(x)=v(x+\alpha_k)-v(x)$.

Write the formal adjoints (which I also don't know what it means) of $A$ and $B$ as $A^* \colon V_1 \to V_0$ and $B^* \colon V_2 \to V_1$ given by $A^*w=\sum_{k=1}^{n}L_k^*w_k$ and $(B^*z)_l=\sum_{k=1}^{n}L_k^*z_{kl}$, where $L^*_kv(x)=v(x-\alpha_k)-v(x)$. It was very easy to show that $BA=0$, $A^*B^*=0$ and $AA^*+B^*B=A^*A \otimes Id_{V_1}$ and also understand the proof of the Lemma 3.1 which states that the operator $M=A^*A \colon V_0 \to V_0$ is bijective if and only if $\alpha$ is Diophantine.

CouldIt is obvious that $im A \subset ker B$, since $BA=0$, but may one help me by telling me whyhow he uses Lemma 3.1 into obtain the proofexactness of the main theorem, i.esequence $0 \to V_0 \stackrel{A}{\to} V_1 \stackrel{B}{\to} V_2 \to 0$?

And one more thing. Theorem 1I am prety confused with a compatibility condition, and how identity $AA^*+B^*B=A^*A \otimes Id_{V_1}$ impliessince in the linearizes equation $Av-g=B^*BM^{-1}g$$g_k=\phi_k$ has to have zero mean value, where $v=A^*M^{-1}g$ for some $g \in V_1$ but in the nonlinearized equation it does not. So, how he solves that mean value "obstacle"?

Thank you a lot.

I am reading Moser's paper On commuting circle mappings and simultaneous Diophantine approximations and I found it hard because it is my first time that I seriously have to read a paper. It is a local simultaneous conjugacy result for commuting diffeomorphisms in the neighborhood of rotations.

The main theorem is Theorem 1 which states that if $\phi_k$ are diffeomorphisms of $S^1$ in the neighborhood of rotations $R_{\alpha_k}$ in $C^{\infty}$-topology, $k=1,\ldots,n$, where $\alpha=(\alpha_1,\ldots,\alpha_n)$ is Diophantine vector in $\mathbb R^n$, and if $\phi_k \circ \phi_l=\phi_l \circ \phi_k$, then there is a smooth diffeomorphism of $S^1$ simultaneously conjugating each $\phi_k$ to $R_{\alpha_k}$.

Here is the set-up for the proof of the main theorem. Let $V_0=C^{\infty}_0(S^1,\mathbb R)$, $V_1=C^{\infty}_0(S^1,\mathbb R^n)$ and $V_2=C^{\infty}_0(S^1,so(n))$, where $0$ subscript means that functions have mean value $0$ and $so(n)$ is the Lie algebra of $SO(n)$, i.e. the vector space of all skew-symmetric matrices. Also, let $A \colon V_0 \to V_1$ and $B \colon V_1 \to V_2$ be linear maps given by $(Av)_k=L_kv$ and $(Bw)_{kl}=L_kw_l-L_lw_k$, where $L_kv(x)=v(x+\alpha_k)-v(x)$.

Write the formal adjoints (which I also don't know what it means) of $A$ and $B$ as $A^* \colon V_1 \to V_0$ and $B^* \colon V_2 \to V_1$ given by $A^*w=\sum_{k=1}^{n}L_k^*w_k$ and $(B^*z)_l=\sum_{k=1}^{n}L_k^*z_{kl}$, where $L^*_kv(x)=v(x-\alpha_k)-v(x)$. It was very easy to show that $BA=0$, $A^*B^*=0$ and $AA^*+B^*B=A^*A \otimes Id_{V_1}$ and also understand the proof of the Lemma 3.1 which states that the operator $M=A^*A \colon V_0 \to V_0$ is bijective if and only if $\alpha$ is Diophantine.

Could one help me by telling me why he uses Lemma 3.1 in the proof of the main theorem, i.e. Theorem 1, and how identity $AA^*+B^*B=A^*A \otimes Id_{V_1}$ implies $Av-g=B^*BM^{-1}g$, where $v=A^*M^{-1}g$ for some $g \in V_1$ . Thank you a lot.

I am reading Moser's paper On commuting circle mappings and simultaneous Diophantine approximations and I found it hard because it is my first time that I seriously have to read a paper. It is a local simultaneous conjugacy result for commuting diffeomorphisms in the neighborhood of rotations.

The main theorem is Theorem 1 which states that if $\phi_k$ are diffeomorphisms of $S^1$ in the neighborhood of rotations $R_{\alpha_k}$ in $C^{\infty}$-topology, $k=1,\ldots,n$, where $\alpha=(\alpha_1,\ldots,\alpha_n)$ is Diophantine vector in $\mathbb R^n$, and if $\phi_k \circ \phi_l=\phi_l \circ \phi_k$, then there is a smooth diffeomorphism of $S^1$ simultaneously conjugating each $\phi_k$ to $R_{\alpha_k}$.

Here is the set-up for the proof of the main theorem. Let $V_0=C^{\infty}_0(S^1,\mathbb R)$, $V_1=C^{\infty}_0(S^1,\mathbb R^n)$ and $V_2=C^{\infty}_0(S^1,so(n))$, where $0$ subscript means that functions have mean value $0$ and $so(n)$ is the Lie algebra of $SO(n)$, i.e. the vector space of all skew-symmetric matrices. Also, let $A \colon V_0 \to V_1$ and $B \colon V_1 \to V_2$ be linear maps given by $(Av)_k=L_kv$ and $(Bw)_{kl}=L_kw_l-L_lw_k$, where $L_kv(x)=v(x+\alpha_k)-v(x)$.

Write the formal adjoints (which I also don't know what it means) of $A$ and $B$ as $A^* \colon V_1 \to V_0$ and $B^* \colon V_2 \to V_1$ given by $A^*w=\sum_{k=1}^{n}L_k^*w_k$ and $(B^*z)_l=\sum_{k=1}^{n}L_k^*z_{kl}$, where $L^*_kv(x)=v(x-\alpha_k)-v(x)$. It was very easy to show that $BA=0$, $A^*B^*=0$ and $AA^*+B^*B=A^*A \otimes Id_{V_1}$ and also understand the proof of the Lemma 3.1 which states that the operator $M=A^*A \colon V_0 \to V_0$ is bijective if and only if $\alpha$ is Diophantine.

It is obvious that $im A \subset ker B$, since $BA=0$, but may one help me by telling me how he uses Lemma 3.1 to obtain the exactness of the sequence $0 \to V_0 \stackrel{A}{\to} V_1 \stackrel{B}{\to} V_2 \to 0$?

And one more thing. I am prety confused with a compatibility condition, since in the linearizes equation $g_k=\phi_k$ has to have zero mean value, but in the nonlinearized equation it does not. So, how he solves that mean value "obstacle"?

Thank you a lot.

I am reading Moser's paper On commuting circle mappings and simultaneous Diophantine approximations and I found it hard because it is my first time that I seriously have to read a paper. It is a local simultaneous conjugacy result for commuting diffeomorphisms in the neighborhood of rotations.

The main theorem is Theorem 1 which states that if $\phi_k$ are diffeomorphisms of $S^1$ in the neighborhood of rotations $R_{\alpha_k}$ in $C^{\infty}$-topology, $k=1,\ldots,n$, where $\alpha=(\alpha_1,\ldots,\alpha_n)$ is Diophantine vector in $\mathbb R^n$, and if $\phi_k \circ \phi_l=\phi_l \circ \phi_k$, then there is a smooth diffeomorphism of $S^1$ simultaneously conjugating each $\phi_k$ to $R_{\alpha_k}$.

Here is the set-up for the proof of the main theorem. Let $V_0=C^{\infty}_0(S^1,\mathbb R)$, $V_1=C^{\infty}_0(S^1,\mathbb R^n)$ and $V_2=C^{\infty}_0(S^1,so(n))$, where $0$ subscript means that functions have mean value $0$ and $so(n)$ is the Lie algebra of $SO(n)$, i.e. the vector space of all skew-symmetric matrices. Also, let $A \colon V_0 \to V_1$ and $B \colon V_1 \to V_2$ be linear maps given by $(Av)_k=L_kv$ and $(Bw)_{kl}=L_kw_l-L_lw_k$, where $L_kv(x)=v(x+\alpha_k)-v(x)$.

Write the formal adjoints (which I also don't know what it means) of $A$ and $B$ as $A^* \colon V_1 \to V_0$ and $B^* \colon V_2 \to V_1$ given by $A^*w=\sum_{k=1}^{n}L_k^*w_k$ and $(B^*z)_l=\sum_{k=1}^{n}L_k^*z_{kl}$, where $L^*_kv(x)=v(x-\alpha_k)-v(x)$. Then it isIt was very easy to show that $BA=0$, $A^*B^*=0$ and $AA^*+B^*B=A^*A \otimes Id_{V_1}$ and also proveunderstand the proof of the Lemma 3.1 which states that the operator $M=A^*A \colon V_0 \to V_0$ is bijective if and only if $\alpha$ is Diophantine.

Could one help me by telling me why he uses Lemma 3.1 in the proof of the main theorem, i.e. Theorem 1, and how identity $AA^*+B^*B=A^*A \otimes Id_{V_1}$ implies $Av-g=B^*BM^{-1}g$, where $v=A^*M^{-1}g$ for some $g \in V_1$ . Thank you a lot.

I am reading Moser's paper On commuting circle mappings and simultaneous Diophantine approximations and I found it hard because it is my first time that I seriously have to read a paper. It is a local simultaneous conjugacy result for commuting diffeomorphisms in the neighborhood of rotations.

The main theorem is Theorem 1 which states that if $\phi_k$ are diffeomorphisms of $S^1$ in the neighborhood of rotations $R_{\alpha_k}$ in $C^{\infty}$-topology, $k=1,\ldots,n$, where $\alpha=(\alpha_1,\ldots,\alpha_n)$ is Diophantine vector in $\mathbb R^n$, and if $\phi_k \circ \phi_l=\phi_l \circ \phi_k$, then there is a smooth diffeomorphism of $S^1$ simultaneously conjugating each $\phi_k$ to $R_{\alpha_k}$.

Here is the set-up for the proof of the main theorem. Let $V_0=C^{\infty}_0(S^1,\mathbb R)$, $V_1=C^{\infty}_0(S^1,\mathbb R^n)$ and $V_2=C^{\infty}_0(S^1,so(n))$, where $0$ subscript means that functions have mean value $0$ and $so(n)$ is the Lie algebra of $SO(n)$, i.e. the vector space of all skew-symmetric matrices. Also, let $A \colon V_0 \to V_1$ and $B \colon V_1 \to V_2$ be linear maps given by $(Av)_k=L_kv$ and $(Bw)_{kl}=L_kw_l-L_lw_k$, where $L_kv(x)=v(x+\alpha_k)-v(x)$.

Write the formal adjoints (which I also don't know what it means) of $A$ and $B$ as $A^* \colon V_1 \to V_0$ and $B^* \colon V_2 \to V_1$ given by $A^*w=\sum_{k=1}^{n}L_k^*w_k$ and $(B^*z)_l=\sum_{k=1}^{n}L_k^*z_{kl}$, where $L^*_kv(x)=v(x-\alpha_k)-v(x)$. Then it is easy to show that $BA=0$, $A^*B^*=0$ and $AA^*+B^*B=A^*A \otimes Id_{V_1}$ and also prove the Lemma 3.1 which states that the operator $M=A^*A \colon V_0 \to V_0$ is bijective if and only if $\alpha$ is Diophantine.

Could one help me by telling me why he uses Lemma 3.1 in the proof of the main theorem, i.e. Theorem 1, and how identity $AA^*+B^*B=A^*A \otimes Id_{V_1}$ implies $Av-g=B^*BM^{-1}g$, where $v=A^*M^{-1}g$ for some $g \in V_1$ . Thank you a lot.

I am reading Moser's paper On commuting circle mappings and simultaneous Diophantine approximations and I found it hard because it is my first time that I seriously have to read a paper. It is a local simultaneous conjugacy result for commuting diffeomorphisms in the neighborhood of rotations.

The main theorem is Theorem 1 which states that if $\phi_k$ are diffeomorphisms of $S^1$ in the neighborhood of rotations $R_{\alpha_k}$ in $C^{\infty}$-topology, $k=1,\ldots,n$, where $\alpha=(\alpha_1,\ldots,\alpha_n)$ is Diophantine vector in $\mathbb R^n$, and if $\phi_k \circ \phi_l=\phi_l \circ \phi_k$, then there is a smooth diffeomorphism of $S^1$ simultaneously conjugating each $\phi_k$ to $R_{\alpha_k}$.

Here is the set-up for the proof of the main theorem. Let $V_0=C^{\infty}_0(S^1,\mathbb R)$, $V_1=C^{\infty}_0(S^1,\mathbb R^n)$ and $V_2=C^{\infty}_0(S^1,so(n))$, where $0$ subscript means that functions have mean value $0$ and $so(n)$ is the Lie algebra of $SO(n)$, i.e. the vector space of all skew-symmetric matrices. Also, let $A \colon V_0 \to V_1$ and $B \colon V_1 \to V_2$ be linear maps given by $(Av)_k=L_kv$ and $(Bw)_{kl}=L_kw_l-L_lw_k$, where $L_kv(x)=v(x+\alpha_k)-v(x)$.

Write the formal adjoints (which I also don't know what it means) of $A$ and $B$ as $A^* \colon V_1 \to V_0$ and $B^* \colon V_2 \to V_1$ given by $A^*w=\sum_{k=1}^{n}L_k^*w_k$ and $(B^*z)_l=\sum_{k=1}^{n}L_k^*z_{kl}$, where $L^*_kv(x)=v(x-\alpha_k)-v(x)$. It was very easy to show that $BA=0$, $A^*B^*=0$ and $AA^*+B^*B=A^*A \otimes Id_{V_1}$ and also understand the proof of the Lemma 3.1 which states that the operator $M=A^*A \colon V_0 \to V_0$ is bijective if and only if $\alpha$ is Diophantine.

Could one help me by telling me why he uses Lemma 3.1 in the proof of the main theorem, i.e. Theorem 1, and how identity $AA^*+B^*B=A^*A \otimes Id_{V_1}$ implies $Av-g=B^*BM^{-1}g$, where $v=A^*M^{-1}g$ for some $g \in V_1$ . Thank you a lot.

improved formatting
Source Link
Yemon Choi
  • 25.8k
  • 9
  • 69
  • 156

I am reading Moser's paper On commuting circle mappings and simultaneous Diophantine approximations and I found it hard because it is my first time that I seriously have to read a paper. It is a local simultaneous conjugacy result for commuting diffeomorphisms in the neighborhood of rotations. 

The main theorem is the Theorem 1 which states that if $\phi_k$ are diffeomorphisms of $S^1$ in the neighborhood of rotations $R_{\alpha_k}$ in $C^{\infty}$-topology, $k=1,\ldots,n$, where $\alpha=(\alpha_1,\ldots,\alpha_n)$ is Diophantine vector in $\mathbb R^n$, and if $\phi_k \circ \phi_l=\phi_l \circ \phi_k$, then there is a smooth diffeomorphism of $S^1$ simultaneously conjugating each $\phi_k$ to $R_{\alpha_k}$. 

Here is the set-up for the proof of the main theorem. Let $V_0=C^{\infty}_0(S^1,\mathbb R)$, $V_1=C^{\infty}_0(S^1,\mathbb R^n)$ and $V_2=C^{\infty}_0(S^1,so(n))$, where $0$ subscript means that functions have mean value $0$ and $so(n)$ is the Lie algebra of $SO(n)$, i.e. the vector space of all skew-symmetric matrices. Also, and let $A \colon V_0 \to V_1$ and $B \colon V_1 \to V_2$ be linear maps given by $(Av)_k=L_kv$ and $(Bw)_{kl}=L_kw_l-L_lw_k$, where $L_kv(x)=v(x+\alpha_k)-v(x)$. 

Write the formal adjoints (which I also don't know what it means) of $A$ and $B$ as $A^* \colon V_1 \to V_0$ and $B^* \colon V_2 \to V_1$ given by $A^*w=\sum_{k=1}^{n}L_k^*w_k$ and $(B^*z)_l=\sum_{k=1}^{n}L_k^*z_{kl}$, where $L^*_kv(x)=v(x-\alpha_k)-v(x)$. Then it is easy to show that $BA=0$, $A^*B^*=0$ and $AA^*+B^*B=A^*A \otimes Id_{V_1}$ and also prove the Lemma 3.1 which states that the operator $M=A^*A \colon V_0 \to V_0$ is bijective if and only if $\alpha$ is Diophantine. 

Could one help me by telling me why he uses Lemma 3.1 in the proof of the main theorem, i.e. Theorem 1, and how identity $AA^*+B^*B=A^*A \otimes Id_{V_1}$ implies $Av-g=B^*BM^{-1}g$, where $v=A^*M^{-1}g$ for some $g \in V_1$ . Thank you a lot.

I am reading Moser's paper On commuting circle mappings and simultaneous Diophantine approximations and I found it hard because it is my first time that I seriously have to read a paper. It is a local simultaneous conjugacy result for commuting diffeomorphisms in the neighborhood of rotations. The main theorem is the Theorem 1 which states that if $\phi_k$ are diffeomorphisms of $S^1$ in the neighborhood of rotations $R_{\alpha_k}$ in $C^{\infty}$-topology, $k=1,\ldots,n$, where $\alpha=(\alpha_1,\ldots,\alpha_n)$ is Diophantine vector in $\mathbb R^n$, and if $\phi_k \circ \phi_l=\phi_l \circ \phi_k$, then there is a smooth diffeomorphism of $S^1$ simultaneously conjugating each $\phi_k$ to $R_{\alpha_k}$. Here is the set-up for the proof of the main theorem. Let $V_0=C^{\infty}_0(S^1,\mathbb R)$, $V_1=C^{\infty}_0(S^1,\mathbb R^n)$ and $V_2=C^{\infty}_0(S^1,so(n))$, where $0$ subscript means that functions have mean value $0$ and $so(n)$ is the Lie algebra of $SO(n)$, i.e. the vector space of all skew-symmetric matrices, and let $A \colon V_0 \to V_1$ and $B \colon V_1 \to V_2$ be linear maps given by $(Av)_k=L_kv$ and $(Bw)_{kl}=L_kw_l-L_lw_k$, where $L_kv(x)=v(x+\alpha_k)-v(x)$. Write the formal adjoints (which I also don't know what it means) of $A$ and $B$ as $A^* \colon V_1 \to V_0$ and $B^* \colon V_2 \to V_1$ given by $A^*w=\sum_{k=1}^{n}L_k^*w_k$ and $(B^*z)_l=\sum_{k=1}^{n}L_k^*z_{kl}$, where $L^*_kv(x)=v(x-\alpha_k)-v(x)$. Then it is easy to show that $BA=0$, $A^*B^*=0$ and $AA^*+B^*B=A^*A \otimes Id_{V_1}$ and also prove the Lemma 3.1 which states that the operator $M=A^*A \colon V_0 \to V_0$ is bijective if and only if $\alpha$ is Diophantine. Could one help me by telling me why he uses Lemma 3.1 in the proof of the main theorem, i.e. Theorem 1, and how identity $AA^*+B^*B=A^*A \otimes Id_{V_1}$ implies $Av-g=B^*BM^{-1}g$, where $v=A^*M^{-1}g$ for some $g \in V_1$ . Thank you a lot.

I am reading Moser's paper On commuting circle mappings and simultaneous Diophantine approximations and I found it hard because it is my first time that I seriously have to read a paper. It is a local simultaneous conjugacy result for commuting diffeomorphisms in the neighborhood of rotations. 

The main theorem is Theorem 1 which states that if $\phi_k$ are diffeomorphisms of $S^1$ in the neighborhood of rotations $R_{\alpha_k}$ in $C^{\infty}$-topology, $k=1,\ldots,n$, where $\alpha=(\alpha_1,\ldots,\alpha_n)$ is Diophantine vector in $\mathbb R^n$, and if $\phi_k \circ \phi_l=\phi_l \circ \phi_k$, then there is a smooth diffeomorphism of $S^1$ simultaneously conjugating each $\phi_k$ to $R_{\alpha_k}$. 

Here is the set-up for the proof of the main theorem. Let $V_0=C^{\infty}_0(S^1,\mathbb R)$, $V_1=C^{\infty}_0(S^1,\mathbb R^n)$ and $V_2=C^{\infty}_0(S^1,so(n))$, where $0$ subscript means that functions have mean value $0$ and $so(n)$ is the Lie algebra of $SO(n)$, i.e. the vector space of all skew-symmetric matrices. Also, let $A \colon V_0 \to V_1$ and $B \colon V_1 \to V_2$ be linear maps given by $(Av)_k=L_kv$ and $(Bw)_{kl}=L_kw_l-L_lw_k$, where $L_kv(x)=v(x+\alpha_k)-v(x)$. 

Write the formal adjoints (which I also don't know what it means) of $A$ and $B$ as $A^* \colon V_1 \to V_0$ and $B^* \colon V_2 \to V_1$ given by $A^*w=\sum_{k=1}^{n}L_k^*w_k$ and $(B^*z)_l=\sum_{k=1}^{n}L_k^*z_{kl}$, where $L^*_kv(x)=v(x-\alpha_k)-v(x)$. Then it is easy to show that $BA=0$, $A^*B^*=0$ and $AA^*+B^*B=A^*A \otimes Id_{V_1}$ and also prove the Lemma 3.1 which states that the operator $M=A^*A \colon V_0 \to V_0$ is bijective if and only if $\alpha$ is Diophantine. 

Could one help me by telling me why he uses Lemma 3.1 in the proof of the main theorem, i.e. Theorem 1, and how identity $AA^*+B^*B=A^*A \otimes Id_{V_1}$ implies $Av-g=B^*BM^{-1}g$, where $v=A^*M^{-1}g$ for some $g \in V_1$ . Thank you a lot.

I fill in some missing details so the newcomers can at least understand the statment of the main theorem.
Source Link
Loading
Source Link
Boris
  • 39
  • 1
Loading