Normal Ordering in the Algebra Generated by x x and I \mathrm{I} and a Combinatorial Generalization of Bessel Numbers
Abdelhay Benmoussa · 2025
Read Normal Ordering in the Algebra Generated by x x and I \mathrm{I} and a Combinatorial Generalization of Bessel Numbers by Abdelhay Benmoussa (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-00416.txt.
Introduction
Some operator algebras hide clean combinatorial patterns inside their product rules. Abdelhay Benmoussa investigates normal ordering in the algebra generated by multiplication by x and an integration operator in 2025. The two generators obey a simple commutation relation that drives the whole study.
Normal ordering rewrites products of operators in a standard form, and the numbers that appear often count objects. His work leads to a broader version of the Bessel numbers, one that blends algebra with counting. Such bridges link algebra with counting, and they draw in both operator theorists and fans of integer sequences. Normal ordering is where operator algebra meets the art of counting. The commutation relation is simple, which makes the resulting patterns worth exhibiting. For a reader who likes links, the charm is the bridge. Rules for how marks swap places give rise to counts that no one put in by hand at the start.
Catalog record
| CREATOR | Abdelhay Benmoussa |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 11 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-00416
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Normal Ordering in the Algebra Generated by x x and I \mathrm{I} and a Combinatorial Generalization of Bessel Numbers
Abdelhay Benmoussa
Abstract
We investigate the algebra generated by the operators x x and I = ∫ 0 x \mathrm{I}=\int_{0}^{x} , which satisfy the commutation relation [ I , x ] = I x − x I = − I 2 . [\mathrm{I},x]=\mathrm{I}x-x\mathrm{I}=-\mathrm{I}^{2}. We develop a combinatorial framework for the normal ordering of words in this algebra and show that any word can be written in the form w = ∑ i , j c ( i , j ) x i I j , w=\sum_{i,j}c(i,j)\,x^{i}\mathrm{I}^{j}, where the coefficients c ( i , j ) c(i,j) are signed integers. Focusing on powers of the operator ( x I ) n (x\mathrm{I})^{n} , we demonstrate that the corresponding coefficients coincide with the classical Bessel numbers (OEIS A001498). We further extend this analysis to powers of the generalized operators ( x λ I δ ) n (x^{\lambda}\mathrm{I}^{\delta})^{n} and, finally, provide an explicit normal-ordered expression for an arbitrary word.
1
Introduction
Normal ordering problems in operator algebras have a long history, beginning with the classical Weyl algebra. In the Weyl algebra, the Scherk identity (1823) states
(xD)^{n}=\sum_{k=0}^{n}S(n,k)x^{k}D^{k},
where
S(n,k)
are the Stirling numbers of the second kind.
In this paper, we consider an analogous algebra generated by
x
and
\mathrm{I}=\int_{0}^{x}
with commutation relation
[\mathrm{I},x]=-\mathrm{I}^{2},
and study the normal ordering of powers of
(x\mathrm{I})
and its generalizations.
2
Combinatorial Framework
2.1
Algebra Generated by
x
and
\mathrm{I}
Any word
w
in this algebra can be written uniquely as
w=\sum_{i,j}c(i,j)x^{i}\mathrm{I}^{j},
with
c(i,j)\in\textbf{Z}
.
2.2
Normal Ordering of
(x\mathrm{I})^{n}
We first consider the normal ordering of
(x\mathrm{I})^{n}
.
Theorem 1
For
n\geq 1
,
(x\mathrm{I})^{n}=\sum_{k=0}^{n-1}(-1)^{k}a(n-1,k)x^{n-k}\mathrm{I}^{n+k},
where
a(n,k)
are the
Bessel numbers
(OEIS A001498).
Proof 1
We prove the theorem by induction on
n
. For
n=1
, the statement is immediate.
Assume it holds for some
n\geq 1
, i.e.,
(x\mathrm{I})^{n}=\sum_{k=0}^{n-1}(-1)^{k}a(n-1,k)x^{n-k}\mathrm{I}^{\,n+k}.
Then,
(x\mathrm{I})^{n+1}f(x)=x\mathrm{I}\Big((t\mathrm{I})^{n}f(t)\Big)(x)=\sum_{k=0}^{n-1}(-1)^{k}a(n-1,k)x\mathrm{I}\big(t^{\,n-k}\mathrm{I}^{\,n+k}f(t)\big)(x).
Applying the integration by parts formula
\mathrm{I}\big(t^{\alpha}\mathrm{I}^{\beta}f(t)\big)(x)=\sum_{j=0}^{\alpha}(-1)^{j}(\alpha)_{j}x^{\alpha+1-j}\mathrm{I}^{\beta+1+j}f(x),
where
(\alpha)_{j}=\alpha(\alpha-1)\cdots(\alpha-j+1)
, we get
(x\mathrm{I})^{n+1}f(x)=\sum_{k=0}^{n-1}\sum_{j=0}^{n-k}(-1)^{k+j}(n-k)_{j}a(n-1,k)x^{n+1-k-j}\mathrm{I}^{\,n+1+k+j}f(x).
Reindexing with
i=k+j
, we obtain a single sum
(x\mathrm{I})^{\,n+1}f(x)=\sum_{i=0}^{n}(-1)^{i}a(n,i)x^{\,n+1-i}\mathrm{I}^{\,n+1+i}f(x),
where the coefficients satisfy the combinatorial identity
\displaystyle a(n,i)=\sum_{k=0}^{\min(n-1,i)}(n-k)_{i-k}a(n-1,k).
(1)
It is straightforward to verify that this recurrence coincides with the Bessel number recurrence
a(n,k)=a(n-1,k)+(n-k+1)a(n,k-1),\quad a(0,0)=1.
Hence, the coefficients
a(n,k)
are exactly the Bessel numbers.
3
Normal Ordering of
(x^{\lambda}\mathrm{I}^{\delta})^{n}
We now generalize the previous result. Let
\lambda,\delta\geq 1
be integers.
Theorem 2
For
n\geq 1
,
(x^{\lambda}\mathrm{I}^{\delta})^{n}=\sum_{k=0}^{\lambda(n-1)}(-1)^{k}a_{n,k}^{(\lambda)(\delta)}x^{\lambda n-k}\mathrm{I}^{\delta n+k},
where the coefficients
a_{n,k}^{(\lambda)(\delta)}
satisfy a recurrence obtained by induction.
Proof 2
We prove the formula by repeated application of the integration-by-parts identity:
\mathrm{I}\big(x^{\alpha}\mathrm{I}^{\beta}\big)=\sum_{j=0}^{\alpha}(-1)^{j}(\alpha)_{j}\,x^{\alpha-j}\mathrm{I}^{\beta+1+j},
where
(\alpha)_{j}=\alpha(\alpha-1)\cdots(\alpha-j+1)
is the falling factorial.
Let us compute the following example
\mathrm{I}^{2}\bigl(x^{5}\mathrm{I}^{3}\bigr)=\mathrm{I}\!\left(\mathrm{I}(x^{5}\mathrm{I}^{3})\right)
Using the previous identity, we obtain
\mathrm{I}(x^{5}\mathrm{I}^{3})=\sum_{k=0}^{5}(-1)^{k}(5)_{k}\,x^{5-k}\mathrm{I}^{\,4+k}.
Applying
\mathrm{I}
again,
\mathrm{I}^{2}(x^{5}\mathrm{I}^{3})=\sum_{k=0}^{5}(-1)^{k}(5)_{k}\,\mathrm{I}\!\bigl(x^{5-k}\mathrm{I}^{\,4+k}\bigr).
Using the same identity,
\mathrm{I}\!\bigl(x^{5-k}\mathrm{I}^{\,4+k}\bigr)=\sum_{i=0}^{5-k}(-1)^{i}(5-k)_{i}\,x^{5-k-i}\mathrm{I}^{\,5+k+i}.
Therefore
\displaystyle\mathrm{I}^{2}(x^{5}\mathrm{I}^{3})
\displaystyle=\sum_{k=0}^{5}\sum_{i=0}^{5-k}(-1)^{k+i}\,(5)_{k}\,(5-k)_{i}\;x^{5-k-i}\,\mathrm{I}^{\,5+k+i}
\displaystyle=\sum_{k=0}^{5}\sum_{j=k}^{5}(-1)^{j}\,(5)_{k}\,(5-k)_{j-k}\;x^{5-j}\,\mathrm{I}^{\,5+j}
Applying
\mathrm{I}
\delta
times
gives a nested sum
\displaystyle\mathrm{I}^{\delta}(x^{\alpha}\mathrm{I}^{\beta})
\displaystyle=\sum_{k_{1}=0}^{\alpha}\sum_{k_{2}=k_{1}}^{\alpha}\cdots\sum_{k_{\delta}=k_{\delta-1}}^{\alpha}(-1)^{k_{\delta}}(\alpha)_{k_{1}}(\alpha-k_{1})_{k_{2}-k_{1}}\cdots(\alpha-k_{\delta-1})_{k_{\delta}-k_{\delta-1}}x^{\alpha-k_{\delta}}\mathrm{I}^{\beta+\delta+k_{\delta}}
\displaystyle=\sum_{k_{\delta}=0}^{\alpha}(-1)^{k_{\delta}}(\alpha)_{k_{\delta}}\left(\sum_{k_{\delta-1}=0}^{k_{\delta}}\cdots\sum_{k_{1}=0}^{k_{2}}\right)x^{\alpha-k_{\delta}}\mathrm{I}^{\beta+\delta+k_{\delta}}
Now, observe that the nested sums count all
weakly increasing sequences
:
0\leq k_{1}\leq k_{2}\leq\dots\leq k_{\delta}.
For a fixed value of
k_{\delta}
, the number of sequences
(k_{1},\dots,k_{\delta-1})
is exactly the number of weak compositions of
k_{\delta}
into
\delta
parts, which is given by the binomial coefficient
\sum_{0\leq k_{1}\leq k_{2}\leq\dots\leq k_{\delta-1}\leq k_{\delta}}1=\binom{k_{\delta}+\delta-1}{\delta-1}.
Hence, we can evaluate the nested sum and reduce it to a single sum over
k_{\delta}
:
\mathrm{I}^{\delta}(x^{\alpha}\mathrm{I}^{\beta})=\sum_{k_{\delta}=0}^{\alpha}(-1)^{k_{\delta}}(\alpha)_{k_{\delta}}\binom{k_{\delta}+\delta-1}{\delta-1}\,x^{\alpha-k_{\delta}}\mathrm{I}^{\beta+\delta+k_{\delta}}.
This gives the desired formula, where the multiple sum is evaluated as a binomial coefficient.
(x^{\lambda}\mathrm{I}^{\delta})^{n+1}=\sum_{k=0}^{\lambda(n-1)}(-1)^{k}a^{(\lambda,\delta)}_{n,k}\,(x^{\lambda}\mathrm{I}^{\delta})(x^{\lambda n-k}\mathrm{I}^{\delta n+k})
\sum_{k=0}^{\lambda(n-1)}\sum_{i=0}^{\lambda n-k}(-1)^{i+k}\,a^{(\lambda,\delta)}_{n,k}\,(\lambda n-k)_{i}\binom{i+\delta-1}{\delta-1}x^{\lambda(n+1)-k-i}\,\mathrm{I}^{\delta(n+1)+k+i}
Let
j=k+i
. Then
=\sum_{j=0}^{\lambda n}(-1)^{j}\,a^{(\lambda,\delta)}_{n+1,j}\,x^{\lambda(n+1)-j}\,\mathrm{I}^{\delta(n+1)+j}.
Hence the recurrence:
a^{(\lambda,\delta)}_{n+1,j}=\sum_{k=0}^{\min(j,\lambda(n-1))}(\lambda n-k)_{j-k}\binom{j-k+\delta-1}{\delta-1}\,a^{(\lambda,\delta)}_{n,k}.
these numbers generalize the Bessel numbers for
\lambda=\delta=1
.
4
Normal Ordering of any arbitrary word
Theorem 3
Let
w=x^{r_{n}}\mathrm{I}^{s_{n}}\cdots x^{r_{2}}\mathrm{I}^{s_{2}}x^{r_{1}}\mathrm{I}^{s_{1}}
be an arbitrary word.
Then
w
has the normal–ordered expansion
\displaystyle w
\displaystyle=\sum_{p_{n-1}=0}^{S_{n-1}}\sum_{p_{n-2}=0}^{\min(p_{n-1},S_{n-2})}\cdots\sum_{p_{1}=0}^{\min(p_{2},S_{1})}(-1)^{p_{n-1}}
\displaystyle\times(S_{1})_{p_{1}}(S_{2}-p_{1})_{p_{2}-p_{1}}\,(S_{3}-p_{2})_{p_{3}-p_{2}}\cdots(S_{n-1}-p_{n-2})_{p_{n-1}-p_{n-2}}
\displaystyle\times\binom{p_{1}+s_{2}-1}{s_{2}-1}\binom{p_{2}-p_{1}+s_{3}-1}{s_{3}-1}\cdots\binom{p_{n-1}-p_{n-2}+s_{n}-1}{s_{n}-1}
\displaystyle\times x^{\,S_{n}-(p_{n-1}-p_{n-2})}\,\mathrm{I}^{\,s+(p_{n-1}-p_{n-2})}.
where
S_{j}:=r_{1}+\cdots+r_{j}
and
s:=s_{1}+\cdots+s_{n}
.
Proof 3
We have
x^{r_{2}}\mathrm{I}^{s_{2}}x^{r_{1}}\mathrm{I}^{s_{1}}=\sum_{i=0}^{r_{1}}(-1)^{i}(r_{1})_{i}\binom{i+s_{2}-1}{s_{2}-1}x^{r_{2}+r_{1}-i}\mathrm{I}^{s_{1}+s_{2}+i}
Then
\displaystyle x^{r_{3}}\mathrm{I}^{s_{3}}x^{r_{2}}\mathrm{I}^{s_{2}}x^{r_{1}}\mathrm{I}^{s_{1}}
\displaystyle=\sum_{i=0}^{r_{1}}(-1)^{i}(r_{1})_{i}\binom{i+s_{2}-1}{s_{2}-1}x^{r_{3}}\mathrm{I}^{s_{3}}\left(x^{r_{2}+r_{1}-i}\mathrm{I}^{s_{1}+s_{2}+i}\right)
\displaystyle=\sum_{i=0}^{r_{1}}\sum_{j=0}^{r_{1}+r_{2}-i}(-1)^{i+j}(r_{1})_{i}(r_{1}+r_{2}-i)_{j}\binom{i+s_{2}-1}{s_{2}-1}\binom{j+s_{3}-1}{s_{3}-1}
\displaystyle x^{r_{1}+r_{2}+r_{3}-j}\,\mathrm{I}^{s_{1}+s_{2}+s_{3}+j}.
Hence
\displaystyle w
\displaystyle=\sum_{i_{1}=0}^{r_{1}}\sum_{i_{2}=0}^{r_{1}+r_{2}-i_{1}}\cdots\sum_{i_{n-1}=0}^{r_{1}+\cdots+r_{n-1}-i_{n-2}}
\displaystyle(-1)^{i_{1}+\cdots+i_{n-1}}(r_{1})_{i_{1}}(r_{1}+r_{2}-i_{1})_{i_{2}}\cdots(r_{1}+\cdots+r_{n-1}-i_{n-2})_{i_{n-1}}
\displaystyle\times\binom{i_{1}+s_{2}-1}{s_{2}-1}\binom{i_{2}+s_{3}-1}{s_{3}-1}\cdots\binom{i_{n-1}+s_{n}-1}{s_{n}-1}\,x^{r_{1}+\cdots+r_{n}-i_{n-1}}\,\mathrm{I}^{s_{1}+\cdots+s_{n}+i_{n-1}}
Now our goal is to invert
\sum_{i_{1}=0}^{r_{1}}\sum_{i_{2}=0}^{r_{1}+r_{2}-i_{1}}\cdots\sum_{i_{n-1}=0}^{r_{1}+\cdots+r_{n-1}-i_{n-2}}\!\!F(i_{1},\dots,i_{n-1})
where
F(i_{1},\dots,i_{n-1})=(-1)^{i_{1}+\cdots+i_{n-1}}(r_{1})_{i_{1}}(r_{1}+r_{2}-i_{1})_{i_{2}}\cdots(r_{1}+\cdots+r_{n-1}-i_{n-2})_{i_{n-1}}
\qquad\times\binom{i_{1}+s_{2}-1}{s_{2}-1}\binom{i_{2}+s_{3}-1}{s_{3}-1}\cdots\binom{i_{n-1}+s_{n}-1}{s_{n}-1}\,x^{r_{1}+\cdots+r_{n}-i_{n-1}}\,\mathrm{I}^{s_{1}+\cdots+s_{n}+i_{n-1}}.
Define cumulative sums
p_{j}:=i_{1}+i_{2}+\cdots+i_{j}\qquad(1\leq j\leq n-1),\qquad p_{0}:=0.
Then
i_{j}=p_{j}-p_{j-1}
for each
j
.
The original summation region
0\leq i_{1}\leq r_{1},\qquad 0\leq i_{2}\leq r_{1}+r_{2}-i_{1},\qquad\ldots,\qquad 0\leq i_{n-1}\leq r_{1}+\cdots+r_{n-1}-i_{n-2}
is equivalent, after the substitution
i_{j}=p_{j}-p_{j-1}
, to the monotone chain of inequalities
0\leq p_{1}\leq p_{2}\leq\cdots\leq p_{n-1}\leq S_{n-1},\qquad\text{where }S_{j}:=r_{1}+\cdots+r_{j}.
Hence we may invert the summation order and sum first over
p_{n-1}
, then
p_{n-2}
, down to
p_{1}
. The inverted summation domain is
\sum_{p_{n-1}=0}^{S_{n-1}}\sum_{p_{n-2}=0}^{\min(p_{n-1},S_{n-2})}\cdots\sum_{p_{1}=0}^{\min(p_{2},S_{1})}.
We now rewrite every factor of
F
in terms of the
p_{j}
.
The sign becomes
(-1)^{i_{1}+\cdots+i_{n-1}}=(-1)^{p_{n-1}}.
The falling factorials transform as
(r_{1})_{i_{1}}=(S_{1})_{p_{1}},\qquad(r_{1}+r_{2}-i_{1})_{i_{2}}=(S_{2}-p_{1})_{p_{2}-p_{1}},
\ldots,\qquad(S_{n-1}-p_{n-2})_{p_{n-1}-p_{n-2}}.
The binomial factors become
\binom{i_{1}+s_{2}-1}{s_{2}-1}=\binom{p_{1}+s_{2}-1}{s_{2}-1},\qquad\binom{i_{2}+s_{3}-1}{s_{3}-1}=\binom{p_{2}-p_{1}+s_{3}-1}{s_{3}-1},
\ldots,\qquad\binom{i_{n-1}+s_{n}-1}{s_{n}-1}=\binom{p_{n-1}-p_{n-2}+s_{n}-1}{s_{n}-1}.
Finally, since
i_{n-1}=p_{n-1}-p_{n-2}
, the power factors are
x^{r_{1}+\cdots+r_{n}-i_{n-1}}=x^{S_{n}-(p_{n-1}-p_{n-2})},\qquad\mathrm{I}^{s_{1}+\cdots+s_{n}+i_{n-1}}=\mathrm{I}^{\,s+(p_{n-1}-p_{n-2})},
where
S_{n}:=r_{1}+\cdots+r_{n}
and
s:=s_{1}+\cdots+s_{n}
.
Combining these expressions yields the inverted sum:
\displaystyle\sum_{p_{n-1}=0}^{S_{n-1}}\sum_{p_{n-2}=0}^{\min(p_{n-1},S_{n-2})}\cdots\sum_{p_{1}=0}^{\min(p_{2},S_{1})}(-1)^{p_{n-1}}\,(S_{1})_{p_{1}}
\displaystyle\times(S_{2}-p_{1})_{p_{2}-p_{1}}\,(S_{3}-p_{2})_{p_{3}-p_{2}}\cdots(S_{n-1}-p_{n-2})_{p_{n-1}-p_{n-2}}
\displaystyle\times\binom{p_{1}+s_{2}-1}{s_{2}-1}\binom{p_{2}-p_{1}+s_{3}-1}{s_{3}-1}\cdots\binom{p_{n-1}-p_{n-2}+s_{n}-1}{s_{n}-1}