Skip to main content
corrected normalisation in "tends to 0"
Source Link
ARG
  • 4.4k
  • 1
  • 25
  • 46

There is no generating set for which this is possible (actually this is true for any graph with an infinite connected component). Furthermore, your claim that (in non-amenable groups) such $f_n$ cannot exist if $f_n \geq 0$ is false. Here is a detailed proof.

Lemma 1: If $0 \leq v,w \in \ell^1G$ then $\|v+w\|_{\ell^1} = \|v \|_{\ell^1} + \|w\|_{\ell^1}$.

Proof: for all $g \in G$, $|v(g) + w(g)| = v(g) + w(g) = |v(g)| + |w(g)|$ this passes to the $\ell^1$-norm (by summing both sides over all $g$). $\hspace{10.5cm} \square$

Lemma 2: For any function $v \in \ell^1G$ and for any $g \in G$, $\|gv\|_{\ell^1} = \|v\|_{\ell^1}$. If further $v \geq 0$, $gv \geq 0$.

Proof: $gv$ has the same values as $v$ but permuted by the group action, so the norm is the same. The same reasoning shows that $Pv$ is positive when $v$ is. $\hspace{7.5cm} \square$

Definition: Let $P$ be the random walk operator: $P = \frac{1}{|S|} \sum_{g \in S} g$ or $P = L+I$ (where $I$ is the identity).

Lemma 3: for any $0 \leq v \in \ell^1G$, $\|Pv\|_{\ell^1} = \|v\|_{\ell^1}$ and $Pv \geq 0$.

Proof: Note that for any $g \in G$, $gv \geq 0$. $$ \begin{array}{rll} \|Pv\|_{\ell^1} &= \| \tfrac{1}{|S|} \sum_{g \in S} gv \|_{\ell^1} & = \tfrac{1}{|S|} \| \sum_{g \in S} gv \|_{\ell^1} \\ &\overset{Lemma 1}{=} \tfrac{1}{|S|} \sum_{g \in S} \| gv \|_{\ell^1} &\overset{Lemma 2}{=} \tfrac{1}{|S|} \sum_{g \in S} \| v \|_{\ell^1} \\ &= \| v \|_{\ell^1} \end{array} $$ $Pv$ is a strictly positive scalar times a sum of positive function, so it is positive. $\hspace{3cm} \square$

Notation: $P^k$ is $P$ applied $k$ times (e.g. $P^3 v= PPPv$) and $P^0 = I$.

Lemma 4: Let $f_n = \displaystyle \sum_{i=0}^n P^i v$ where $v \geq 0$ and $\|v\|_{\ell^1} =1$. Then $\|f_n\|_{\ell^1}= n+1$.

Proof: $$ \begin{array}{rll} \|f_n\|_{\ell^1} &= \displaystyle \| \sum_{i=0}^n P^i v\|_{\ell^1} & \overset{Lemma 1}{=} \displaystyle \sum_{i=0}^n \| P^i v\|_{\ell^1}\\ & \overset{Lemma 3}{=} \displaystyle \sum_{i=0}^n \|v\|_{\ell^1} & = n+1. \\ \end{array} $$

Lemma 5: Let $f_n = \displaystyle \sum_{i=0}^n P^i v$ where $v \geq 0$ and $\|v\|_{\ell^1} =1$. Then $\displaystyle \frac{\|L f_n\|_{\ell^1}}{\| f_n \|_{\ell^1}} \to 0$.

Proof: Let's compute $Lf_n$ using $L=P-I$: $$ Lf_n = \displaystyle \sum_{i=0}^n LP^i v = \displaystyle \sum_{i=0}^n (P-I)P^i v = \displaystyle \sum_{i=0}^n (P^{i+1}-P^i) v = P^{n+1}v -v $$$$ Lf_n = \displaystyle \sum_{i=0}^n LP^i v = \displaystyle \sum_{i=0}^n (P-I)P^i v = \displaystyle \sum_{i=0}^n \big( P^{i+1}v-P^i v \big) = P^{n+1}v -v $$ Now $\|Lf_n\|_{\ell^1} = \|P^{n+1}v -v\|_{\ell^1} \overset{TI}{\leq} \|P^{n+1}v\|_{\ell^1} + \|v\|_{\ell^1} = 2$ where $TI$ stands for the triangle inequality.

Using Lemma 4 $\displaystyle \frac{ \|L f_n\|_{\ell^1}}{\|f_n\|_{\ell^1}} \leq \frac{2}{n+1}$ which obviously tends to 0.

Corollary: 0 is in the spectrum of $L_1$ and the Laplacian cannot be inverted $\ell^1$.

Note the argument is true in any graph. If $v$ is supported on a finite component, then $f_n$$\tfrac{1}{n+1} f_n$ tends to the constant function, which is indeed in the kernel. Otherwise $\tfrac{1}{n+1} f_n$ tends weak$^*$ to 0 but not in norm.

As a complementary remark the image of the Laplacian is not dense (in a graph with an infinite connected component). By taking $v = \delta_x$, the Dirac mass at some vertex $x$, the sequence $f_n$ above shows its image is weak$^*$ dense. It's easy to check that its image lies in $\ell^1_0X = \lbrace f \in \ell^1X \mid \sum_{x \in X} f(x) =0 \rbrace$. One can further check that $\overline{\mathrm{Im} \Delta} = \ell^1_0 X$ if and only if the graph has no non-constant bounded harmonic functions. Non-amenable groups always have non-constant bounded harmonic functions so that the (norm) closure of the image of $\Delta$ is strict subspace of $\ell^1_0 X$.

There is no generating set for which this is possible (actually this is true for any graph with an infinite connected component). Furthermore, your claim that (in non-amenable groups) such $f_n$ cannot exist if $f_n \geq 0$ is false. Here is a detailed proof.

Lemma 1: If $0 \leq v,w \in \ell^1G$ then $\|v+w\|_{\ell^1} = \|v \|_{\ell^1} + \|w\|_{\ell^1}$.

Proof: for all $g \in G$, $|v(g) + w(g)| = v(g) + w(g) = |v(g)| + |w(g)|$ this passes to the $\ell^1$-norm (by summing both sides over all $g$). $\hspace{10.5cm} \square$

Lemma 2: For any function $v \in \ell^1G$ and for any $g \in G$, $\|gv\|_{\ell^1} = \|v\|_{\ell^1}$. If further $v \geq 0$, $gv \geq 0$.

Proof: $gv$ has the same values as $v$ but permuted by the group action, so the norm is the same. The same reasoning shows that $Pv$ is positive when $v$ is. $\hspace{7.5cm} \square$

Definition: Let $P$ be the random walk operator: $P = \frac{1}{|S|} \sum_{g \in S} g$ or $P = L+I$ (where $I$ is the identity).

Lemma 3: for any $0 \leq v \in \ell^1G$, $\|Pv\|_{\ell^1} = \|v\|_{\ell^1}$ and $Pv \geq 0$.

Proof: Note that for any $g \in G$, $gv \geq 0$. $$ \begin{array}{rll} \|Pv\|_{\ell^1} &= \| \tfrac{1}{|S|} \sum_{g \in S} gv \|_{\ell^1} & = \tfrac{1}{|S|} \| \sum_{g \in S} gv \|_{\ell^1} \\ &\overset{Lemma 1}{=} \tfrac{1}{|S|} \sum_{g \in S} \| gv \|_{\ell^1} &\overset{Lemma 2}{=} \tfrac{1}{|S|} \sum_{g \in S} \| v \|_{\ell^1} \\ &= \| v \|_{\ell^1} \end{array} $$ $Pv$ is a strictly positive scalar times a sum of positive function, so it is positive. $\hspace{3cm} \square$

Notation: $P^k$ is $P$ applied $k$ times (e.g. $P^3 v= PPPv$) and $P^0 = I$.

Lemma 4: Let $f_n = \displaystyle \sum_{i=0}^n P^i v$ where $v \geq 0$ and $\|v\|_{\ell^1} =1$. Then $\|f_n\|_{\ell^1}= n+1$.

Proof: $$ \begin{array}{rll} \|f_n\|_{\ell^1} &= \displaystyle \| \sum_{i=0}^n P^i v\|_{\ell^1} & \overset{Lemma 1}{=} \displaystyle \sum_{i=0}^n \| P^i v\|_{\ell^1}\\ & \overset{Lemma 3}{=} \displaystyle \sum_{i=0}^n \|v\|_{\ell^1} & = n+1. \\ \end{array} $$

Lemma 5: Let $f_n = \displaystyle \sum_{i=0}^n P^i v$ where $v \geq 0$ and $\|v\|_{\ell^1} =1$. Then $\displaystyle \frac{\|L f_n\|_{\ell^1}}{\| f_n \|_{\ell^1}} \to 0$.

Proof: Let's compute $Lf_n$ using $L=P-I$: $$ Lf_n = \displaystyle \sum_{i=0}^n LP^i v = \displaystyle \sum_{i=0}^n (P-I)P^i v = \displaystyle \sum_{i=0}^n (P^{i+1}-P^i) v = P^{n+1}v -v $$ Now $\|Lf_n\|_{\ell^1} = \|P^{n+1}v -v\|_{\ell^1} \overset{TI}{\leq} \|P^{n+1}v\|_{\ell^1} + \|v\|_{\ell^1} = 2$ where $TI$ stands for the triangle inequality.

Using Lemma 4 $\displaystyle \frac{ \|L f_n\|_{\ell^1}}{\|f_n\|_{\ell^1}} \leq \frac{2}{n+1}$ which obviously tends to 0.

Corollary: 0 is in the spectrum of $L_1$ and the Laplacian cannot be inverted $\ell^1$.

Note the argument is true in any graph. If $v$ is supported on a finite component, then $f_n$ tends to the constant function, which is indeed in the kernel.

As a complementary remark the image of the Laplacian is not dense (in a graph with an infinite connected component). By taking $v = \delta_x$, the Dirac mass at some vertex $x$, the sequence $f_n$ above shows its image is weak$^*$ dense. It's easy to check that its image lies in $\ell^1_0X = \lbrace f \in \ell^1X \mid \sum_{x \in X} f(x) =0 \rbrace$. One can further check that $\overline{\mathrm{Im} \Delta} = \ell^1_0 X$ if and only if the graph has no non-constant bounded harmonic functions. Non-amenable groups always have non-constant bounded harmonic functions so that the (norm) closure of the image of $\Delta$ is strict subspace of $\ell^1_0 X$.

There is no generating set for which this is possible (actually this is true for any graph with an infinite connected component). Furthermore, your claim that (in non-amenable groups) such $f_n$ cannot exist if $f_n \geq 0$ is false. Here is a detailed proof.

Lemma 1: If $0 \leq v,w \in \ell^1G$ then $\|v+w\|_{\ell^1} = \|v \|_{\ell^1} + \|w\|_{\ell^1}$.

Proof: for all $g \in G$, $|v(g) + w(g)| = v(g) + w(g) = |v(g)| + |w(g)|$ this passes to the $\ell^1$-norm (by summing both sides over all $g$). $\hspace{10.5cm} \square$

Lemma 2: For any function $v \in \ell^1G$ and for any $g \in G$, $\|gv\|_{\ell^1} = \|v\|_{\ell^1}$. If further $v \geq 0$, $gv \geq 0$.

Proof: $gv$ has the same values as $v$ but permuted by the group action, so the norm is the same. The same reasoning shows that $Pv$ is positive when $v$ is. $\hspace{7.5cm} \square$

Definition: Let $P$ be the random walk operator: $P = \frac{1}{|S|} \sum_{g \in S} g$ or $P = L+I$ (where $I$ is the identity).

Lemma 3: for any $0 \leq v \in \ell^1G$, $\|Pv\|_{\ell^1} = \|v\|_{\ell^1}$ and $Pv \geq 0$.

Proof: Note that for any $g \in G$, $gv \geq 0$. $$ \begin{array}{rll} \|Pv\|_{\ell^1} &= \| \tfrac{1}{|S|} \sum_{g \in S} gv \|_{\ell^1} & = \tfrac{1}{|S|} \| \sum_{g \in S} gv \|_{\ell^1} \\ &\overset{Lemma 1}{=} \tfrac{1}{|S|} \sum_{g \in S} \| gv \|_{\ell^1} &\overset{Lemma 2}{=} \tfrac{1}{|S|} \sum_{g \in S} \| v \|_{\ell^1} \\ &= \| v \|_{\ell^1} \end{array} $$ $Pv$ is a strictly positive scalar times a sum of positive function, so it is positive. $\hspace{3cm} \square$

Notation: $P^k$ is $P$ applied $k$ times (e.g. $P^3 v= PPPv$) and $P^0 = I$.

Lemma 4: Let $f_n = \displaystyle \sum_{i=0}^n P^i v$ where $v \geq 0$ and $\|v\|_{\ell^1} =1$. Then $\|f_n\|_{\ell^1}= n+1$.

Proof: $$ \begin{array}{rll} \|f_n\|_{\ell^1} &= \displaystyle \| \sum_{i=0}^n P^i v\|_{\ell^1} & \overset{Lemma 1}{=} \displaystyle \sum_{i=0}^n \| P^i v\|_{\ell^1}\\ & \overset{Lemma 3}{=} \displaystyle \sum_{i=0}^n \|v\|_{\ell^1} & = n+1. \\ \end{array} $$

Lemma 5: Let $f_n = \displaystyle \sum_{i=0}^n P^i v$ where $v \geq 0$ and $\|v\|_{\ell^1} =1$. Then $\displaystyle \frac{\|L f_n\|_{\ell^1}}{\| f_n \|_{\ell^1}} \to 0$.

Proof: Let's compute $Lf_n$ using $L=P-I$: $$ Lf_n = \displaystyle \sum_{i=0}^n LP^i v = \displaystyle \sum_{i=0}^n (P-I)P^i v = \displaystyle \sum_{i=0}^n \big( P^{i+1}v-P^i v \big) = P^{n+1}v -v $$ Now $\|Lf_n\|_{\ell^1} = \|P^{n+1}v -v\|_{\ell^1} \overset{TI}{\leq} \|P^{n+1}v\|_{\ell^1} + \|v\|_{\ell^1} = 2$ where $TI$ stands for the triangle inequality.

Using Lemma 4 $\displaystyle \frac{ \|L f_n\|_{\ell^1}}{\|f_n\|_{\ell^1}} \leq \frac{2}{n+1}$ which obviously tends to 0.

Corollary: 0 is in the spectrum of $L_1$ and the Laplacian cannot be inverted $\ell^1$.

Note the argument is true in any graph. If $v$ is supported on a finite component, then $\tfrac{1}{n+1} f_n$ tends to the constant function, which is indeed in the kernel. Otherwise $\tfrac{1}{n+1} f_n$ tends weak$^*$ to 0 but not in norm.

As a complementary remark the image of the Laplacian is not dense (in a graph with an infinite connected component). By taking $v = \delta_x$, the Dirac mass at some vertex $x$, the sequence $f_n$ above shows its image is weak$^*$ dense. It's easy to check that its image lies in $\ell^1_0X = \lbrace f \in \ell^1X \mid \sum_{x \in X} f(x) =0 \rbrace$. One can further check that $\overline{\mathrm{Im} \Delta} = \ell^1_0 X$ if and only if the graph has no non-constant bounded harmonic functions. Non-amenable groups always have non-constant bounded harmonic functions so that the (norm) closure of the image of $\Delta$ is strict subspace of $\ell^1_0 X$.

tried to rewrite the answer as I'm puzzled by the absence of upvotes although the question got quite a few upvotes since I answered it
Source Link
ARG
  • 4.4k
  • 1
  • 25
  • 46

There is no generating set for which this is possible (actually this is true for any graph with an infinite connected component). Furthermore, your claim that (in non-amenable groups) such $f_n$ cannot exist if $f_n \geq 0$ is false. Here is a detailed proof.

Lemma 1: If $0 \leq v,w \in \ell^1G$ then $\|v+w\|_{\ell^1} = \|v \|_{\ell^1} + \|w\|_{\ell^1}$.

RemarkProof:: There are 2 conventions for the Laplacian with opposite signall $g \in G$, so $\Delta = -L$ has positive spectrum$|v(g) + w(g)| = v(g) + w(g) = |v(g)| + |w(g)|$ this passes to the $\ell^1$-norm (inby summing both sides over all $\ell^2$$g$). $\hspace{10.5cm} \square$

Lemma 2: For any function $v \in \ell^1G$ and for any $L$ has negative spectrum$g \in G$, $\|gv\|_{\ell^1} = \|v\|_{\ell^1}$. If further $v \geq 0$, $gv \geq 0$.

LetProof: $I$ be$gv$ has the identitysame values as $v$ but permuted by the group action, so the norm is the same. Let The same reasoning shows that $Pv$ is positive when $v$ is. $\hspace{7.5cm} \square$

Definition: Let $P$ be the random walk operator, $\Delta = I-P$ so: $P = I+L$,$P = \frac{1}{|S|} \sum_{g \in S} g$ or $P = L+I$ (where $I$ is the identity).

Lemma 3: for any $0 \leq v \in \ell^1G$, more simply $P = \displaystyle \frac{1}{|S|} \sum_{g \in S} g$$\|Pv\|_{\ell^1} = \|v\|_{\ell^1}$ and $Pv \geq 0$.

Denote byProof: Note that for any $P^i$ the$g \in G$, $i^\text{th}$ iterate$gv \geq 0$. $$ \begin{array}{rll} \|Pv\|_{\ell^1} &= \| \tfrac{1}{|S|} \sum_{g \in S} gv \|_{\ell^1} & = \tfrac{1}{|S|} \| \sum_{g \in S} gv \|_{\ell^1} \\ &\overset{Lemma 1}{=} \tfrac{1}{|S|} \sum_{g \in S} \| gv \|_{\ell^1} &\overset{Lemma 2}{=} \tfrac{1}{|S|} \sum_{g \in S} \| v \|_{\ell^1} \\ &= \| v \|_{\ell^1} \end{array} $$ $Pv$ is a strictly positive scalar times a sum of positive function, so it is positive. $\hspace{3cm} \square$

Notation: $P^k$ is $P$ applied $k$ times (for example:e.g. $P^2 = P P$$P^3 v= PPPv$) and $P^0 = I$).

Lemma 4: Let $g_n = \displaystyle \sum_{i=0}^n P^n \delta_x$$f_n = \displaystyle \sum_{i=0}^n P^i v$ where $\delta_x$ is a Dirac mass at some vertex$v \geq 0$ and $x$$\|v\|_{\ell^1} =1$. Since $P^i \delta_x$ is the probability distribution of the random walk at time $i$,Then $\|P^i\delta_x\|_{\ell^1} =1$$\|f_n\|_{\ell^1}= n+1$.

Proof: (In fact,$$ \begin{array}{rll} \|f_n\|_{\ell^1} &= \displaystyle \| \sum_{i=0}^n P^i v\|_{\ell^1} & \overset{Lemma 1}{=} \displaystyle \sum_{i=0}^n \| P^i v\|_{\ell^1}\\ & \overset{Lemma 3}{=} \displaystyle \sum_{i=0}^n \|v\|_{\ell^1} & = n+1. \\ \end{array} $$

Lemma 5: Let $P$ preserves the$f_n = \displaystyle \sum_{i=0}^n P^i v$ where $\ell^1$-norm of positive fucntions$v \geq 0$ and $\|v\|_{\ell^1} =1$.) Next, since all the $P^i\delta_x$ are positive functions, their norm add up in $\ell^1$: henceThen $\| g_n\|_{\ell^1}= n+1$$\displaystyle \frac{\|L f_n\|_{\ell^1}}{\| f_n \|_{\ell^1}} \to 0$.

On the other hand,Proof: Let's compute $L = P-I$ and so$Lf_n$ using $L=P-I$: $$ L g_n = P\bigg( \sum_{i=0}^n P^n \delta_x\bigg) - \bigg( \sum_{i=0}^n P^n \delta_x\bigg) = P^{n+1} \delta_x -\delta_x $$$$ Lf_n = \displaystyle \sum_{i=0}^n LP^i v = \displaystyle \sum_{i=0}^n (P-I)P^i v = \displaystyle \sum_{i=0}^n (P^{i+1}-P^i) v = P^{n+1}v -v $$ SoNow $\| L g_n\|_{\ell^1} \leq \|\delta_x\|_{\ell^1} + \| P^{n+1} \delta_x\|_{\ell^1} = 2$ by$\|Lf_n\|_{\ell^1} = \|P^{n+1}v -v\|_{\ell^1} \overset{TI}{\leq} \|P^{n+1}v\|_{\ell^1} + \|v\|_{\ell^1} = 2$ where $TI$ stands for the triangle inequality.

This shows thatUsing Lemma 4 $\displaystyle \frac{ \|L g_n\|_{\ell^1}}{\|g_n\|_{\ell^1}} \leq \frac{2}{n+1}$$\displaystyle \frac{ \|L f_n\|_{\ell^1}}{\|f_n\|_{\ell^1}} \leq \frac{2}{n+1}$ which obviously tends to 0. So

Corollary: 0 is in the spectrum of $L_1$ and the Laplacian cannot be inverted $\ell^1$.

Note the argument is true in any graph. If (without$v$ is supported on a finite connected components)component, then $f_n$ tends to the constant function, which is indeed in the kernel.

AlsoAs a complementary remark the image of the Laplacian is not dense (it is weak$^*$ dense asin a graph with an infinite connected component). By taking $v = \delta_x$, the Dirac mass at some vertex $x$, the sequence $g_n$$f_n$ above shows) its image is weak$^*$ dense. It's easy to check that its image lies in $\ell^1_0X = \lbrace f \in \ell^1X \mid \sum_{x \in X} f(x) =0 \rbrace$. One can further check that $\overline{\mathrm{Im} \Delta} = \ell^1_0 X$ if and only if the graph has no non-constant bounded harmonic functions. Non-amenable groups always have non-constant bounded harmonic functions so that the (norm) closure of the image of $\Delta$ is strict subspace of $\ell^1_0 X$.

There is no generating set for which this is possible (actually this is true for any graph with an infinite connected component).

Remark: There are 2 conventions for the Laplacian with opposite sign, so $\Delta = -L$ has positive spectrum (in $\ell^2$) and $L$ has negative spectrum.

Let $I$ be the identity. Let $P$ be the random walk operator, $\Delta = I-P$ so $P = I+L$, or, more simply $P = \displaystyle \frac{1}{|S|} \sum_{g \in S} g$.

Denote by $P^i$ the $i^\text{th}$ iterate of $P$ (for example: $P^2 = P P$) and $P^0 = I$)

Let $g_n = \displaystyle \sum_{i=0}^n P^n \delta_x$ where $\delta_x$ is a Dirac mass at some vertex $x$. Since $P^i \delta_x$ is the probability distribution of the random walk at time $i$, $\|P^i\delta_x\|_{\ell^1} =1$. (In fact, $P$ preserves the $\ell^1$-norm of positive fucntions.) Next, since all the $P^i\delta_x$ are positive functions, their norm add up in $\ell^1$: hence $\| g_n\|_{\ell^1}= n+1$

On the other hand, $L = P-I$ and so $$ L g_n = P\bigg( \sum_{i=0}^n P^n \delta_x\bigg) - \bigg( \sum_{i=0}^n P^n \delta_x\bigg) = P^{n+1} \delta_x -\delta_x $$ So $\| L g_n\|_{\ell^1} \leq \|\delta_x\|_{\ell^1} + \| P^{n+1} \delta_x\|_{\ell^1} = 2$ by the triangle inequality.

This shows that $\displaystyle \frac{ \|L g_n\|_{\ell^1}}{\|g_n\|_{\ell^1}} \leq \frac{2}{n+1}$ which obviously tends to 0. So 0 is in the spectrum and the Laplacian cannot be inverted $\ell^1$.

Note the argument is true in any graph (without finite connected components).

Also the image of the Laplacian is not dense (it is weak$^*$ dense as the sequence $g_n$ above shows). It's easy to check that its image lies in $\ell^1_0X = \lbrace f \in \ell^1X \mid \sum_{x \in X} f(x) =0 \rbrace$. One can further check that $\overline{\mathrm{Im} \Delta} = \ell^1_0 X$ if and only if the graph has no non-constant bounded harmonic functions. Non-amenable groups always have non-constant bounded harmonic functions so that the closure of the image of $\Delta$ is strict subspace of $\ell^1_0 X$.

There is no generating set for which this is possible (actually this is true for any graph with an infinite connected component). Furthermore, your claim that (in non-amenable groups) such $f_n$ cannot exist if $f_n \geq 0$ is false. Here is a detailed proof.

Lemma 1: If $0 \leq v,w \in \ell^1G$ then $\|v+w\|_{\ell^1} = \|v \|_{\ell^1} + \|w\|_{\ell^1}$.

Proof: for all $g \in G$, $|v(g) + w(g)| = v(g) + w(g) = |v(g)| + |w(g)|$ this passes to the $\ell^1$-norm (by summing both sides over all $g$). $\hspace{10.5cm} \square$

Lemma 2: For any function $v \in \ell^1G$ and for any $g \in G$, $\|gv\|_{\ell^1} = \|v\|_{\ell^1}$. If further $v \geq 0$, $gv \geq 0$.

Proof: $gv$ has the same values as $v$ but permuted by the group action, so the norm is the same. The same reasoning shows that $Pv$ is positive when $v$ is. $\hspace{7.5cm} \square$

Definition: Let $P$ be the random walk operator: $P = \frac{1}{|S|} \sum_{g \in S} g$ or $P = L+I$ (where $I$ is the identity).

Lemma 3: for any $0 \leq v \in \ell^1G$, $\|Pv\|_{\ell^1} = \|v\|_{\ell^1}$ and $Pv \geq 0$.

Proof: Note that for any $g \in G$, $gv \geq 0$. $$ \begin{array}{rll} \|Pv\|_{\ell^1} &= \| \tfrac{1}{|S|} \sum_{g \in S} gv \|_{\ell^1} & = \tfrac{1}{|S|} \| \sum_{g \in S} gv \|_{\ell^1} \\ &\overset{Lemma 1}{=} \tfrac{1}{|S|} \sum_{g \in S} \| gv \|_{\ell^1} &\overset{Lemma 2}{=} \tfrac{1}{|S|} \sum_{g \in S} \| v \|_{\ell^1} \\ &= \| v \|_{\ell^1} \end{array} $$ $Pv$ is a strictly positive scalar times a sum of positive function, so it is positive. $\hspace{3cm} \square$

Notation: $P^k$ is $P$ applied $k$ times (e.g. $P^3 v= PPPv$) and $P^0 = I$.

Lemma 4: Let $f_n = \displaystyle \sum_{i=0}^n P^i v$ where $v \geq 0$ and $\|v\|_{\ell^1} =1$. Then $\|f_n\|_{\ell^1}= n+1$.

Proof: $$ \begin{array}{rll} \|f_n\|_{\ell^1} &= \displaystyle \| \sum_{i=0}^n P^i v\|_{\ell^1} & \overset{Lemma 1}{=} \displaystyle \sum_{i=0}^n \| P^i v\|_{\ell^1}\\ & \overset{Lemma 3}{=} \displaystyle \sum_{i=0}^n \|v\|_{\ell^1} & = n+1. \\ \end{array} $$

Lemma 5: Let $f_n = \displaystyle \sum_{i=0}^n P^i v$ where $v \geq 0$ and $\|v\|_{\ell^1} =1$. Then $\displaystyle \frac{\|L f_n\|_{\ell^1}}{\| f_n \|_{\ell^1}} \to 0$.

Proof: Let's compute $Lf_n$ using $L=P-I$: $$ Lf_n = \displaystyle \sum_{i=0}^n LP^i v = \displaystyle \sum_{i=0}^n (P-I)P^i v = \displaystyle \sum_{i=0}^n (P^{i+1}-P^i) v = P^{n+1}v -v $$ Now $\|Lf_n\|_{\ell^1} = \|P^{n+1}v -v\|_{\ell^1} \overset{TI}{\leq} \|P^{n+1}v\|_{\ell^1} + \|v\|_{\ell^1} = 2$ where $TI$ stands for the triangle inequality.

Using Lemma 4 $\displaystyle \frac{ \|L f_n\|_{\ell^1}}{\|f_n\|_{\ell^1}} \leq \frac{2}{n+1}$ which obviously tends to 0.

Corollary: 0 is in the spectrum of $L_1$ and the Laplacian cannot be inverted $\ell^1$.

Note the argument is true in any graph. If $v$ is supported on a finite component, then $f_n$ tends to the constant function, which is indeed in the kernel.

As a complementary remark the image of the Laplacian is not dense (in a graph with an infinite connected component). By taking $v = \delta_x$, the Dirac mass at some vertex $x$, the sequence $f_n$ above shows its image is weak$^*$ dense. It's easy to check that its image lies in $\ell^1_0X = \lbrace f \in \ell^1X \mid \sum_{x \in X} f(x) =0 \rbrace$. One can further check that $\overline{\mathrm{Im} \Delta} = \ell^1_0 X$ if and only if the graph has no non-constant bounded harmonic functions. Non-amenable groups always have non-constant bounded harmonic functions so that the (norm) closure of the image of $\Delta$ is strict subspace of $\ell^1_0 X$.

Tried to improve presentation...
Source Link
ARG
  • 4.4k
  • 1
  • 25
  • 46

There is no generating set for which this is possible (actually this is true for any graph with an infinite connected component).

The functions $g_n$ which show 0 are in the spectrum are constructed by trying to invert the Laplacian naively

Remark: My conventionThere are 2 conventions for the Laplacian seem to have thewith opposite sign as yours, so I'm using $\Delta = -L$ has positive spectrum (in $\ell^2$) and $L$ has negative spectrum.

Let $I$ be the identity. Let $P$ be the random walk operator, $P = I- \Delta$$\Delta = I-P$ so $P = I+L$, or, more simply $P = \frac{1}{|S|} \sum_{g \in S} g$$P = \displaystyle \frac{1}{|S|} \sum_{g \in S} g$.

Denote by $P^i$ the $i^\text{th}$ iterate of $P$ (for example: $P^2 = P P$) and $P^0 = I$)

Let $g_n = \sum_{i=0}^n P^n \delta_x$$g_n = \displaystyle \sum_{i=0}^n P^n \delta_x$ where $\delta_x$ is a Dirac mass at some vertex $x$. Since $P^i \delta_x$ is the probability distribution of the random walk at time $i$, $\|P^i\delta_x\|_{\ell^1}$$\|P^i\delta_x\|_{\ell^1} =1$. These (In fact, $P$ preserves the $\ell^1$-norm of positive fucntions.) Next, since all the $P^i\delta_x$ are positive functions, so their norm add up in $\ell^1$: hence $\| g_n\|_{\ell^1}= n+1$

On the other hand, $\Delta = I-P$ so$L = P-I$ and $\Delta g_n = \delta_x - P^{n+1} \delta_x$.so $$ L g_n = P\bigg( \sum_{i=0}^n P^n \delta_x\bigg) - \bigg( \sum_{i=0}^n P^n \delta_x\bigg) = P^{n+1} \delta_x -\delta_x $$ So $\| \Delta g_n\|_{\ell^1} \leq \|\delta_x\|_{\ell^1} + \| P^{n+1} \delta_x\|_{\ell^1} = 2$$\| L g_n\|_{\ell^1} \leq \|\delta_x\|_{\ell^1} + \| P^{n+1} \delta_x\|_{\ell^1} = 2$ by the triangle inequality.

This shows that $\displaystyle \frac{ \|\Delta g_n\|_{\ell^1}}{\|g_n\|_{\ell^1}} \leq \frac{2}{n+1}$$\displaystyle \frac{ \|L g_n\|_{\ell^1}}{\|g_n\|_{\ell^1}} \leq \frac{2}{n+1}$ which obviously tends to 0. So 0 is in the spectrum and the Laplacian cannot be inverted $\ell^1$.

Note the argument is true in any graph (without finite connected components).

Also the image of the Laplacian is not dense (it is weak$^*$ dense as the sequence $g_n$ above shows).

It's It's easy to check that its image lies in $\ell^1_0X = \lbrace f \in \ell^1X \mid \sum_{x \in X} f(x) =0 \rbrace$.

One One can further check that $\overline{\mathrm{Im} \Delta} = \ell^1_0 X$ if and only if the graph has no non-constant bounded harmonic functions.

Non Non-amenable groups always have non-constant bounded harmonic functions so that the closure of the image of $\Delta$ is strict subspace of $\ell^1_0 X$.

There is no generating set for which this is possible (actually this is true for any graph with an infinite connected component).

The functions $g_n$ which show 0 are in the spectrum are constructed by trying to invert the Laplacian naively

Remark: My convention for the Laplacian seem to have the opposite sign as yours, so I'm using $\Delta = -L$

Let $P$ be the random walk operator, $P = I- \Delta$ or $P = \frac{1}{|S|} \sum_{g \in S} g$.

Let $g_n = \sum_{i=0}^n P^n \delta_x$ where $\delta_x$ is a Dirac mass at some vertex $x$. Since $P^i \delta_x$ is the probability distribution of the random walk at time $i$, $\|P^i\delta_x\|_{\ell^1}$. These are positive functions, so their norm add up in $\ell^1$: hence $\| g_n\|_{\ell^1}= n+1$

On the other hand, $\Delta = I-P$ so and $\Delta g_n = \delta_x - P^{n+1} \delta_x$. So $\| \Delta g_n\|_{\ell^1} \leq \|\delta_x\|_{\ell^1} + \| P^{n+1} \delta_x\|_{\ell^1} = 2$ by the triangle inequality.

This shows that $\displaystyle \frac{ \|\Delta g_n\|_{\ell^1}}{\|g_n\|_{\ell^1}} \leq \frac{2}{n+1}$ which obviously tends to 0. So 0 is in the spectrum and the Laplacian cannot be inverted $\ell^1$.

Note the argument is true in any graph (without finite connected components).

Also the image of the Laplacian is not dense (it is weak$^*$ dense as the sequence $g_n$ above shows).

It's easy to check that its image lies in $\ell^1_0X = \lbrace f \in \ell^1X \mid \sum_{x \in X} f(x) =0 \rbrace$.

One can further check that $\overline{\mathrm{Im} \Delta} = \ell^1_0 X$ if and only if the graph has no non-constant bounded harmonic functions.

Non-amenable groups always have non-constant bounded harmonic functions so that the closure of the image of $\Delta$ is strict subspace of $\ell^1_0 X$.

There is no generating set for which this is possible (actually this is true for any graph with an infinite connected component).

Remark: There are 2 conventions for the Laplacian with opposite sign, so $\Delta = -L$ has positive spectrum (in $\ell^2$) and $L$ has negative spectrum.

Let $I$ be the identity. Let $P$ be the random walk operator, $\Delta = I-P$ so $P = I+L$, or, more simply $P = \displaystyle \frac{1}{|S|} \sum_{g \in S} g$.

Denote by $P^i$ the $i^\text{th}$ iterate of $P$ (for example: $P^2 = P P$) and $P^0 = I$)

Let $g_n = \displaystyle \sum_{i=0}^n P^n \delta_x$ where $\delta_x$ is a Dirac mass at some vertex $x$. Since $P^i \delta_x$ is the probability distribution of the random walk at time $i$, $\|P^i\delta_x\|_{\ell^1} =1$. (In fact, $P$ preserves the $\ell^1$-norm of positive fucntions.) Next, since all the $P^i\delta_x$ are positive functions, their norm add up in $\ell^1$: hence $\| g_n\|_{\ell^1}= n+1$

On the other hand, $L = P-I$ and so $$ L g_n = P\bigg( \sum_{i=0}^n P^n \delta_x\bigg) - \bigg( \sum_{i=0}^n P^n \delta_x\bigg) = P^{n+1} \delta_x -\delta_x $$ So $\| L g_n\|_{\ell^1} \leq \|\delta_x\|_{\ell^1} + \| P^{n+1} \delta_x\|_{\ell^1} = 2$ by the triangle inequality.

This shows that $\displaystyle \frac{ \|L g_n\|_{\ell^1}}{\|g_n\|_{\ell^1}} \leq \frac{2}{n+1}$ which obviously tends to 0. So 0 is in the spectrum and the Laplacian cannot be inverted $\ell^1$.

Note the argument is true in any graph (without finite connected components).

Also the image of the Laplacian is not dense (it is weak$^*$ dense as the sequence $g_n$ above shows). It's easy to check that its image lies in $\ell^1_0X = \lbrace f \in \ell^1X \mid \sum_{x \in X} f(x) =0 \rbrace$. One can further check that $\overline{\mathrm{Im} \Delta} = \ell^1_0 X$ if and only if the graph has no non-constant bounded harmonic functions. Non-amenable groups always have non-constant bounded harmonic functions so that the closure of the image of $\Delta$ is strict subspace of $\ell^1_0 X$.

Changed convention on the Laplacian, added details, removed useless part of the answer
Source Link
ARG
  • 4.4k
  • 1
  • 25
  • 46
Loading
corrected the answer!
Source Link
ARG
  • 4.4k
  • 1
  • 25
  • 46
Loading
deleted 893 characters in body
Source Link
ARG
  • 4.4k
  • 1
  • 25
  • 46
Loading
Source Link
ARG
  • 4.4k
  • 1
  • 25
  • 46
Loading