U.P.B. Sci. Bull., Series A, Vol. 87, Iss.

3, 2025 ISSN 1223-7027

REMARKS ON d-ARY PARTITIONS AND AN APPLICATION TO

ELEMENTARY SYMMETRIC PARTITIONS

Mircea Cimpoeaş1 , Roxana Tănase2

We prove new formulas for pd (n), the number of d-ary partitions of

n, and, also, for Pd (n), its polynomial part.
Given a partition λ = (λ1 , . . . , λℓ ), its associated j-th symmetric elementary
partition, prej (λ), is the partition whose parts are {λi1 · · · λij : 1 ≤ i1 < · · · <
ij ≤ ℓ}. We prove that if λ and µ are two d-ary partitions of length ℓ such that
prej (λ) = prej (µ), then λ = µ.
Keywords: Restricted partitions, d-ary partitions, elementary
symmetric partitions.
MSC2020: 11P81, 11P83.

1. Introduction
Let n be a positive integer. We denote [n] = {1, 2, . . . , n}. A partition of
n is a non-increasing sequence of positive integers λi whose sum equals n. We
define p(n) as the number of partitions of n and we define p(0) = 1. We denote
λ = (λ1 , λ2 , . . . , λℓ ) with λ1 ≥ λ2 ≥ · · · ≥ λℓ ≥ 1 and |λ| := λ1 + · · · + λℓ = n. We
refer to |λ| as the size of λ and the numbers λi as parts of λ. The number ℓ(λ) = ℓ
is the number of parts of λ and it is called the length of λ. For more on the theory
of partitions, we refer the reader to [1].
Let d ≥ 2 be an integer. A partition λ = (λ1 , . . . , λℓ ) is called d-ary, if all
λi ’s are powers of d. A 2-ary partition is called binary. In Proposition 3.3 we
establish a natural bijection between the set of all integer partition and the set of
d-ary partitions, which conserves the length (but not the size).
In Theorem 3.5 we give a new formula for pd (n), the number of d-ary par-
titions of n, using the fact that a d-ary partition is a partition with the parts in
{1, d, d2 , d3 , . . .}. In Theorem 3.6, we give a new formula for Wj (d, n)’s, the Sylvester
waves of pd (n). Also, in Theorem 3.7 and Theorem 3.8 we give new formulas for
Pd (n) = W1 (d, n), the polynomial part of pd (n).

1
Professor, Faculty of Applied Sciences, National University of Science and Technology Po-
litehnica Bucharest, Romania and Simion Stoilow Institute of Mathematics, Romania, e-mail:
[Link]@[Link]

2
Assistant professor, Faculty of Applied Sciences, National University of Science and Technology
Politehnica Bucharest, Romania, e-mail: roxana [Link]@[Link]

3
4 Mircea Cimpoeaş, Roxana Tănase

Now, let K be an arbitrary field and S = K[x1 , . . . , xℓ ] be the ring of poly-

nomials over K in ℓ indeterminates. We recall that the j th elementary symmetric
polynomial of S is
X
ej (x1 , . . . , xℓ ) = xi1 xi2 · · · xij , where 1 ≤ j ≤ ℓ.
1≤i1 <i2 <...<ij ≤ℓ

Also, we define e0 (x1 , . . . , xℓ ) = 1 and ej (x1 , . . . , xℓ ) = 0 for j > ℓ.

Given a partition λ, we have ej (λ) = 0 if ℓ(λ) < j and
X
ej (λ) = λi1 λi2 · · · λij , if 1 ≤ j ≤ ℓ(λ).
1≤i1 <i2 <...<ij ≤ℓ(λ)

For instance, if λ = (3, 2, 1, 1) is a partition of 7 then

e2 (λ) = e2 (3, 2, 1, 1) = 3 · 2 + 3 · 1 + 3 · 1 + 2 · 1 + 2 · 1 + 1 · 1 = 17.
In [2, 3], Ballantine et al introduced the following definition. Given a partition λ,
the partition prej (λ) is the partition whose parts are
{λi1 · · · λij : 1 ≤ i1 ≤ · · · ≤ ij ≤ ℓ(λ)},
and they called it an elementary symmetric partition.
Note that pre1 (λ) = λ, but prej (λ) ̸= λ, for j ≥ 2. For example, if λ =
(3, 2, 1, 1), then pre2 (λ) = (6, 3, 3, 2, 2, 1).
A natural question to ask is the following: If λ and µ are two partitions such
that prej (λ) = prej (µ) then is it true that λ = µ? Only the following cases are
known in literature: (i) j = 2 and m(λ), m(µ) ≤ 3, see [3, Proposition 14] and (ii)
j = 2 and λ and µ are binary partitions; see [3, Proposition 15]. In Theorem 4.2
we extend the later result and we prove that if λ and µ are two d-ary partitions of
length ℓ such that prej (λ) = prej (µ), where 1 ≤ j ≤ ℓ − 1, then λ = µ.

2. Preliminaries
Let a := (a1 , a2 , . . . , ar ) be a sequence of positive integers, where r ≥ 1. Let λ
be a partition. We say that λ has parts in a if λi ∈ {a1 , . . . , ar } for all 1 ≤ i ≤ ℓ(λ).
The restricted partition function associated to a is pa : N → N, pa (n) := the
number of integer solutions (x1 , . . . , xr ) of ri=1 ai xi = n with xi ≥ 0. In other
P
words, pa (n) counts the number of partitions of n with parts in a.
Note that the generating function of pa (n) is
∞
X 1
pa (n)z n = . (2.1)
(1 − z a1 ) · · · (1 − z ar )
n=0
Let D be a common multiple of a1 , a2 , . . . , ar . Bell [5] proved that pa (n) is a quasi-
polynomial of degree k − 1, with the period D, that is
pa (n) = da,k−1 (n)nk−1 + · · · + da,1 (n)n + da,0 (n), (2.2)
where da,m (n + D) = da,m (n) for 0 ≤ m ≤ k − 1 and n ≥ 0, and da,k−1 (n) is not
identically zero. Sylvester [9],[10] decomposed the restricted partition in a sum of
“waves”: X
pa (n) = Wj (n, a), (2.3)
j≥1
 Remarks on d-ary partitions and an application to elementary symmetric partitions 5

where the sum is taken over all distinct divisors j of the components of a and showed
that for each such j, Wj (n, a) is the coefficient of t−1 in
X ρ−νn
j ent
,
(1 − ρνa
j e
1 −a1 t
) · · · (1 − ρνa
j e
k −ak t
)
0≤ν<j, gcd(ν,j)=1
2πi
where ρj = e j and gcd(0, 0) = 1 by convention. Note that Wj (n, a)’s are quasi-
polynomials of period j. Also, W1 (n, a) is called the polynomial part of pa (n) and
it is denoted by Pa (n).
Theorem 2.1. ([6, Corollary 2.10]) We have
r−1
Y 
1 X n − a1 j1 − · · · − ar jr
pa (n) = +ℓ .
(r − 1)! D
0≤j1 ≤ aD −1,...,0≤jr ≤ aD −1 ℓ=1
1 r
a1 j1 +···+ar jr ≡n( mod D)

The unsigned Stirling numbers are defined by

       
n+r−1 1 (r) 1 r r−1 r r
= n = n + ··· n+ . (2.4)
r−1 n(r − 1)! (r − 1)! r 2 1
Theorem 2.2. ([7, Proposition 4.2]) For any positive integer j with j|ai for some
1 ≤ i ≤ r, we have that
r X j r−1    
1 X X r k
Wj (n, a) = ρℓj (−1)k−m+1 ×
D(r − 1)! k+1 m−1
m=1 ℓ=1 k=m−1
X
× D−k (a1 j1 + · · · + ar jr )k−m+1 nm−1 .
0≤j1 ≤ aD −1,...,0≤jr ≤ aD −1
1 r
a1 j1 +···+ar jr ≡ℓ( mod j)

Theorem 2.3. ([6, Corollary 3.6]) For the polynomial part Pa (n) of the quasi-
polynomial pa (n) we have
r−1
Y  n − a1 j1 − · · · − ar jr 
1 X
Pa (n) = +ℓ .
D(r − 1)! D D
D
0≤j1 ≤ a −1,...,0≤jr ≤ a −1 ℓ=1
1 r

The Bernoulli numbers Bℓ ’s are defined by the identity

X tℓ ∞
t
= Bℓ .
et − 1 ℓ!
ℓ=0

B0 = 1, B1 = − 21 , B2 = 16 , B4 = − 30
1
and Bn = 0 is n is odd and n ≥ 1.
Theorem 2.4. ([6, Corollary 3.11] or [4, page 2])
The polynomial part of pa (n) is
r−1
1 X (−1)u X Bi1 · · · Bir i1
Pa (n) := a · · · airr nr−1−u .
a1 · · · ar (r − 1 − u)! i1 ! · · · ir ! 1
u=0 i1 +···+ir =u
6 Mircea Cimpoeaş, Roxana Tănase

3. New formulas for the number of d-ary partitions

We fix d ≥ 2 an integer. We denote P, the set of integer partitions, and Pd ,
the set of d-ary partitions. Given a positive integer n, we denote pd (n), the number
of d-ary partitions of n.
Definition 3.1. Let λ = (λ1 , . . . , λℓ ) ∈ P be a partition. The d-exponential of λ is
the d-ary partition:
Expd (λ) := (dλ1 −1 , . . . , dλℓ −1 ).
Definition 3.2. Let λ = (λ1 , . . . , λℓ ) ∈ Pd be a d-ary partition. The d-logarithm of
λ is the partition:
Logd (λ) := (logd (λ1 ) + 1, . . . , logd (λℓ ) + 1).
Proposition 3.3. The maps Expd : P → Pd and Logd : Pd → P are bijective and
inverse of each other.
Proof. Let λ = (λ1 , . . . , λℓ ) ∈ P. We have Expd (λ) = (dλ1 −1 , . . . , dλℓ −1 ). Since
logd (dλi −1 ) + 1 = λi − 1 + 1 = λi for all 1 ≤ i ≤ ℓ,
it follows that Logd (Expd (λ)) = λ. Similary, if µ ∈ Pd is a d-ary partition, then it
is easy to see that Expd (Logd (µ)) = µ. Hence, the proof is complete. □
Lemma 3.4. Let n and k be two positive integers such that n < dk+1 . The number
of d-ary partitions of n is
pd (n) = p(1,d,...,dk ) (n).
In particular, the polynomial part of pd (n) is Pd (n) = P(1,d,...,dk ) (n).
Proof. Let λ = (λ1 , . . . , λℓ ) be a d-ary partition of n, that is n = |λ|. It follows that
λi = dci with 0 ≤ ci and dci ≤ n for all 1 ≤ i ≤ ℓ. Since λ1 = dc1 ≤ |λ| < dk+1 and
λ1 ≥ λ2 ≥ · · · ≥ λℓ , it follows that
k ≥ c1 ≥ c2 ≥ · · · ≥ cℓ ≥ 0,
and, therefore, λ is a partition with parts in (1, d, . . . , dk ). On the other hand,
any partition with parts in (1, d, . . . , dk ) is a d-ary partition. Hence, the proof is
complete. □
Theorem 3.5. Let n and k be two positive integers such that n < dk+1 . The number
of d-ary partitions of n is
k 
n − j1 − j2 d − · · · − jk dk−1

1 X Y
pd (n) = +ℓ .
k! dk
0≤j1 ≤dk −1, 0≤j2 ≤dk−1 −1, ...,0≤jk ≤d−1 ℓ=1
j1 +j2 d+···+jk dk−1 ≡n( mod dk )

Proof. According to Lemma 3.4, we have pd (n) = p(1,d,...,dk ) (n), where k = ⌊logd (n)⌋.
Hence, the conclusion follows from Theorem 2.1, taking r = k + 1 and D =
lcm(1, d, . . . , dk ) = dk . □
From Lemma 3.4 and (2.3) we can write
X
pd (n) = Wj (d, n), where Wj (d, n) = Wj (n, (1, d, . . . , dk )),
j≥1
 Remarks on d-ary partitions and an application to elementary symmetric partitions 7

and k = ⌊logd (n)⌋. In particular, the polynomial part of pd (n) is

Pd (n) = W1 (d, n).
Theorem 3.6. Let n and k be two positive integers such that n < dk+1 . We have
that
k+1 j k    
1 XX ℓ X k+1 s−m+1 s
Wj (d, n) = ρj (−1) ×
k!dk s+1 m−1
m=1 ℓ=1 s=m−1
X
−ks
× d (j1 + dj2 + · · · + dk−1 jk−1 )s−m+1 nm−1 .
0≤j1 ≤dk −1,...,0≤jk ≤d−1
j1 +dj2 +···+dk−1 jk−1 ≡ℓ( mod j)

Proof. The conclusion follows from Lemma 3.4 and Theorem 2.2. □
Theorem 3.7. Let n and k be two positive integers such that n < dk+1 . The
polynomial part of pd (n) is
k 
n − j1 − j2 d − · · · − jk dk−1

1 X Y
Pd (n) = +ℓ .
k!dk dk
0≤j1 ≤dk −1, 0≤j2 ≤dk−1 −1, ...,0≤jk ≤d−1 ℓ=1

Proof. The conclusion follows from Lemma 3.4 and Theorem 2.3. □
Theorem 3.8. Let n and k be two positive integers such that n < dk+1 . The
polynomial part of pd (n) is
k
1 X (−1)u X Bi1 · · · Bik+1 i2 +2i3 +···+kik+1 k−u
Pd (n) = k(k+1)
d n .
d 2
u=0
(k − u)! i i ! · · · ik+1 !
=u 1
1 +···+ik+1

Proof. The conclusion follows from Lemma 3.4 and Theorem 2.4. □
Example 3.9. Let n = 8 and d = 3. Since n < d1+1 , Theorem 3.5 implies
 
1 X 8 − j1 8−2
p3 (8) = +1 = + 1 = 3.
1! 3 3
0≤j1 ≤2, j1 ≡8( mod 3)

Also, from Theorem 3.7 it follows that the polynomial part of p3 (8) is
2   2
1 X 8 − j1 1X 11 + 10 + 9 10
P3 (8) = 1
+1 = (11 − j1 ) = = .
1! · 3 3 9 9 3
j1 =0 j1 =0

4. An application to elementary symmetric partitions

Given n ≥ 2 an integer, we denote by {e1 , . . . , en }, the standard basis of the
vector space Rn , i.e. ei is the vector with 1 in the i-th position and zeros everywhere
else.
Let 1 ≤ j ≤ n − 1 be an integer. We consider the vectors:

e1 + e2 + · · · + ej ,
 i=1
ci = e1 + e2 + · · · + ej+1 − ei−1 , 2 ≤ i ≤ j + 1

ei−j+1 + ei−j+2 + · · · + ej , j + 2 ≤ i ≤ n


Let C be the n × n matrix whose columns are c1 , c2 , . . . , cn .

8 Mircea Cimpoeaş, Roxana Tănase

To better illustrate the structure of the matrix C, we present the case n = 6

and j = 3:  
1 0 1 1 0 0
1 1 0 1 0 0
 
1 1 1 0 1 0
C= 0 1 1 1
.
 1 1
0 0 0 0 1 1
0 0 0 0 0 1
Lemma 4.1. With the above notations, we have that det(C) = j.
Proof. From the definition of C, we easily note that det(C) = det(A), where
 
1 0 1 ··· 1
1 1 0 · · · 1
 
A =  ... ... ... . . . ... 
 
 
1 1 1 · · · 0
0 1 1 ··· 1
is a (j + 1) × (j + 1) circulant matrix with the associated polynomial
f (x) = 1 + x + x2 + · · · + xj−1 .
For more details on circulant matrices, we refer the reader to [8].
2πi
Let ω = e j+1 be a primitive (j + 1)-th root of unity. Using a basic result on
circulant matrices, we have that
j
Y
det(A) = f (ω k ).
k=0

It is clear that f (ω 0 ) = f (1) = j. On the other hand, for 1 ≤ k ≤ j, we have that

f (ω k ) = 1 + ω k + · · · + ω k(j−1) = −ω kj .
Therefore, it follows that
j 2 (j+1)
det(A) = (−1)j jω 2 .
If j is even, then
j 2 (j+1) j2 j2
ω 2 = (ω j+1 ) 2 = 1 2 = 1.
On the other hand, if j is odd, then
j 2 (j+1) (j+1) 2 2
ω 2 = (ω 2 )j = (−1)j = −1.
Hence, in both cases, we have that det(A) = j. Thus, the proof is complete. □
Theorem 4.2. Let λ and µ be two d-ary partitions with ℓ parts and let 1 ≤ j ≤ ℓ−1
be an integer. If prej (λ) = prej (µ) then λ = µ.
Proof. Since λ is a d-ary partition, it follows that λ = (λ1 , λ2 , . . . , λℓ ) such that
λi = dci , for all 1 ≤ i ≤ ℓ, and c1 ≥ c2 ≥ · · · ≥ cℓ . Similarly, µ = (µ1 , . . . , µℓ ) with
′
µi = dci , for all 1 ≤ i ≤ ℓ, and c1 ≥ c2 ≥ · · · ≥ cℓ .
 Remarks on d-ary partitions and an application to elementary symmetric partitions 9

From the definition, prej (λ) is the partition whose parts are:

{dci1 +ci2 +···+cij : 1 ≤ i1 < i2 < · · · < ij ≤ ℓ}.

Similarly, prej (µ) is the partition whose parts are:
c′i +c′i +···+c′i
{d 1 2 j : 1 ≤ i1 < i2 < · · · < ij ≤ ℓ}.
Since prej (λ) = prej (µ) it follows that
c′i1 + c′i2 + · · · + c′ij = ci1 + ci2 + · · · + cij , for all 1 ≤ i1 < i2 < · · · < ij ≤ ℓ.
For convenience, we denote
ci1 ,...,ij := ci1 + ci2 + · · · + cij , for all 1 ≤ i1 < i2 < · · · < ij ≤ ℓ.
From Proposition 3.3, in order to prove that λ = µ, it suffices to show that
(c1 , . . . , cℓ ) = (c′1 , . . . , c′ℓ ). In order to do that, it is enough to prove that the linear
system
n
xi1 + xi2 + · · · + xij = ci1 ,...,ij , where 1 ≤ i1 < i2 < · · · < ij ≤ ℓ, (4.1)

has a unique solution. Since (c1 , . . . , cn ) is already a solution of (4.1), it is enough to

prove that the matrix associated to (4.1) has the rank n. We consider the following
subsystem of (4.1):

x1 + x2 + · · · + xj = c1,2,...,j





 x2 + x3 + · · · + xj+1 = c2,...,j+1
x1 + x3 + · · · + xj+1 = c1,3,...,j+1




.

.


.


x1 + · · · + xj−1 + xj+1 = c1,...,j−1,j+1 . (4.2)

x3 + x4 + · · · + xj+2 = c3,...,j+2







 x4 + x5 + · · · + xj+3 = c4,...,j+3
...






xℓ−j+1 + · · · + xℓ = cℓ−j+1,...,ℓ


Note that the matrix associated to (4.2) is C T , where C was defined at the beginning
of this section.
According to Lemma 4.1 we have det(C T ) = det(C) = j ̸= 0. Hence, (4.2)
has a unique solution. Thus (4.1) has also a unique solution, as required. □

5. Conclusions
Let n ≥ 1 and d ≥ 2 be two integers. We proved new formulas for pd (n), the
number of d-ary partitions of n, and, also, for Pd (n), its polynomial part.
Given λ a partition of length ℓ and 1 ≤ j ≤ ℓ − 1, we denote prej (λ), its
associated j-th elementary symmetric partition; see [2, 3]. Given λ and µ two d-ary
partitions of length ℓ and 1 ≤ j ≤ ℓ − 1, we proved that if prej (λ) = prej (µ) then
λ = µ, thus giving a partial positive answer to a problem raised in [2].
10 Mircea Cimpoeaş, Roxana Tănase

REFERENCES
[1] G. E. Andrews, The Theory of Partitions, Cambridge Mathematical Library, Cambridge Uni-
versity Press, Cambridge 1998.
[2] C. Ballantine, G. Beck, M. Merca et al., Elementary Symmetric Partitions, Ann. Comb. (2024),
[Link]
[3] C. Ballantine, G. Beck, M. Merca, Partitions and elementary symmetric polynomials: an exper-
imental approach, Ramanujan J. 66, 34 (2025), [Link]
[4] M. Beck, I. M. Gessel, T. Komatsu, The polynomial part of a restricted partition function
related to the Frobenius problem, Electron. J. Comb. 8(1) (2001), N 7 (5 pages).
[5] E. T. Bell, Interpolated denumerants and Lambert series, Am. J. Math. 65 (1943), 382–386.
[6] M. Cimpoeaş, F. Nicolae, On the restricted partition function, Ramanujan J. 47(3) (2018),
565–588.
[7] M. Cimpoeaş, Remarks on the restricted partition function, Math. Reports 23(73)(4) (2021),
425–436.
[8] A. W. Ingleton, The Rank of Circulant Matrices, J. London Math. Soc. 31(4) (1956), 445–460.
[9] J. J. Sylvester, On the partition of numbers, Quart. J. Pure Appl. Math. 1 (1857), 81–85.
[10] J. J. Sylvester, On subinvariants, i.e. semi-invariants to binary quantics of an unlimited order
with an excursus on rational fractions and partitions, Am. J. Math. 5(1) (1882), 79–136.