Power Monoids: Structure & Factorization

• Power monoids are monoids composed of finite subsets of a semigroup or monoid where multiplication is induced setwise, forming a bridge between additive combinatorics and factorization theory.
• Their arithmetic is highly non-cancellative, prompting innovative divisibility concepts and minimal-factorization invariants to study irreducible elements and structural rigidity.
• Applications include reconstructing base monoids from reduced power monoids, automorphism classification, and asymptotic analysis in numerical and Puiseux monoids.

Power monoids are monoids whose elements are finite subsets of a given semigroup or monoid, with multiplication induced setwise from the base operation. In multiplicative notation, the basic rule is $XY=\{xy:x\in X,\ y\in Y\}$; in additive notation it becomes the sumset $X+Y=\{x+y:x\in X,\ y\in Y\}$. They occupy a distinctive position at the interface of semigroup theory, factorization theory, and additive combinatorics: the construction is elementary, but the resulting arithmetic is highly non-cancellative and has become a primary testing ground for factorization theory beyond the classical cancellative setting (Fan et al., 2017).

1. Basic construction and standard variants

Let $S$ be a semigroup and $H$ a monoid with identity $1_H$. The large power semigroup $\mathcal P(S)$ consists of all nonempty subsets of $S$, while the finitary power semigroup $\mathcal P_{\mathrm{fin}}(S)$ consists of all nonempty finite subsets. When the base object is a monoid, these become monoids with identity $\{1_H\}$. Two submonoids are especially important: the restricted finitary power monoid

$$\mathcal P_{\mathrm{fin},\times}(H):=\{X\in \mathcal P_{\mathrm{fin}}(H):X\cap H^\times\neq\emptyset\},$$

and the reduced finitary power monoid

$X+Y=\{x+y:x\in X,\ y\in Y\}$0

In additive notation one writes $X+Y=\{x+y:x\in X,\ y\in Y\}$1 instead of $X+Y=\{x+y:x\in X,\ y\in Y\}$2 (Tringali, 17 Feb 2026).

Construction	Elements	Operation
$X+Y=\{x+y:x\in X,\ y\in Y\}$3	nonempty subsets of $X+Y=\{x+y:x\in X,\ y\in Y\}$4	setwise product
$X+Y=\{x+y:x\in X,\ y\in Y\}$5	nonempty finite subsets of $X+Y=\{x+y:x\in X,\ y\in Y\}$6	setwise product
$X+Y=\{x+y:x\in X,\ y\in Y\}$7	finite subsets meeting $X+Y=\{x+y:x\in X,\ y\in Y\}$8	setwise product
$X+Y=\{x+y:x\in X,\ y\in Y\}$9	finite subsets containing $S$0	setwise product

The original monoid embeds by $S$1. If $S$2 is an isomorphism, its augmentation

$S$3

induces isomorphisms on the large, finitary, restricted, and reduced power constructions. Units are completely explicit: the units of $S$4 and $S$5 are exactly the singleton units $S$6 with $S$7, whereas $S$8 has trivial unit group, namely $S$9 (Tringali, 17 Feb 2026).

Historically, power semigroups already appear in Dubreil’s 1953 work and later became relevant in the study of semigroup varieties, automata, and formal languages. Their modern arithmetic significance comes from the observation that setwise multiplication converts product-set and sumset phenomena into intrinsic monoid-theoretic structure (Tringali, 17 Feb 2026).

2. Structural features and non-cancellative factorization

Power monoids are almost never cancellative. In fact,

$H$0

For the reduced finitary power monoid, divisibility becomes unusually rigid: if $H$1 divides $H$2 in $H$3, then $H$4. More generally, $H$5 is Dedekind-finite, has trivial unit group, satisfies ACC on principal two-sided ideals, is factorable, and $H$6 is divisor-closed in $H$7 for every submonoid $H$8 (Tringali, 17 Feb 2026).

Because these monoids are strongly non-cancellative, classical atom-based factorization theory is not sufficient on its own. The modern framework therefore uses divisibility-based notions such as irreducibles, quarks, and minimal factorizations. In reduced finitary power monoids, every irreducible is a quark, and the only irreducibles that fail to be atoms are the two-element sets

$H$9

with $1_H$0 and either $1_H$1 or $1_H$2. This identifies involutions and idempotents in the base monoid as the precise obstruction to the coincidence of irreducibles and atoms (Cossu et al., 11 Mar 2025).

Several arithmetic classifications are strikingly clean. The reduced finitary power monoid is atomic iff the base monoid has no unit of order $1_H$3 and its only idempotent is $1_H$4. It is BF iff it is FF, and both are equivalent to torsion-freeness of the base monoid. By contrast, UF and HF occur only in the trivial case. Minimal-factorization invariants are much better behaved: $1_H$5 is always FmF and BmF, even though ordinary factorizations can proliferate because of idempotent phenomena (Tringali, 17 Feb 2026). This is one reason power monoids became a central source of examples in the extension of factorization theory to non-cancellative settings.

The classical transfer-Krull strategy is also sharply limited here. The standard examples $1_H$6 and $1_H$7 are reduced BF-monoids but admit no transfer homomorphism to any cancellative monoid, so their arithmetic cannot be reduced to the usual cancellative models (Geroldinger et al., 2019).

3. Ordered additive classes and ascent phenomena

A large part of the modern theory concerns additive monoids that carry order-theoretic control of sumsets. A commutative monoid is linearly orderable if it admits a total order compatible with addition, and this is equivalent to being cancellative and torsion-free. In that setting, minima and maxima behave well under sumset decomposition: $1_H$8 and finite sumsets satisfy

$1_H$9

These estimates force strong growth whenever a non-singleton is added and make the singleton submonoid divisor-closed inside the finitary power monoid (Dani et al., 6 Jan 2025).

The resulting ascent theory is mixed. For linearly orderable monoids, quasi-ACCP and almost ACCP ascend to the finitary power monoid. The maximal common divisor property is preserved exactly: the base monoid is an MCD-monoid iff its finitary power monoid is. Atomicity is more delicate: for a linearly orderable monoid $\mathcal P(S)$0,

$\mathcal P(S)$1

Thus atomicity upstairs is controlled not only by atomicity downstairs but also by the existence of maximal common divisors for finite subsets. The same paper also gives counterexamples showing that atomicity, nearly atomicity, almost atomicity, and quasi-atomicity need not ascend, while Furstenberg, quasi-Furstenberg, almost Furstenberg, and nearly Furstenberg do ascend; FFM and TIDF ascend under positive Archimedean hypotheses but not in arbitrary linearly orderable monoids (Dani et al., 6 Jan 2025).

For Puiseux monoids, the picture is analogous but was established independently in a specifically additive-rational setting. If $\mathcal P(S)$2 is a Puiseux monoid, then $\mathcal P(S)$3 is unit-cancellative. ACCP, BFM, and FFM ascend from $\mathcal P(S)$4 to $\mathcal P(S)$5, but atomicity does not ascend in general. A sufficient condition is the MCD property: if $\mathcal P(S)$6 is atomic and an MCD-monoid, then $\mathcal P(S)$7 is atomic; if $\mathcal P(S)$8 is not $\mathcal P(S)$9-MCD, then $S$0 is not atomic. The paper constructs an atomic Puiseux monoid whose power monoid is not atomic and also shows that LFFM need not ascend (Gonzalez et al., 2024). A plausible implication is that additive order alone is not enough to force good factorization behavior; common-divisor geometry remains decisive.

4. Rigidity, reconstruction, and automorphisms

One of the most active recent questions is whether a reduced finitary power monoid determines the base monoid. For Puiseux monoids the answer is positive in the strongest possible form: $S$1 For numerical monoids, where isomorphism is equality, this becomes

$S$2

The proof reconstructs the base monoid from the abstract reduced power monoid by identifying the two-element subsets $S$3, showing that isomorphisms preserve them, and then recovering addition via

$S$4

The key combinatorial input is the eventual stabilization formula

$S$5

for sufficiently large $S$6, derived from Nathanson’s theorem on iterated sumsets (Tringali et al., 2023).

This rigidity is special. In general monoids, and even among cancellative commutative monoids, reduced finitary power monoids do not determine the base object. For reduced valuation monoids with isomorphic quotient groups,

$S$7

so nonisomorphic valuation monoids may have isomorphic reduced finitary power monoids. The paper gives explicit examples among valuation submonoids of $S$8 (Rago, 28 Sep 2025).

The general commutative cancellative classification identifies the precise obstruction. If $S$9, then for commutative cancellative monoids $\mathcal P_{\mathrm{fin}}(S)$0 and $\mathcal P_{\mathrm{fin}}(S)$1,

$\mathcal P_{\mathrm{fin}}(S)$2

iff either $\mathcal P_{\mathrm{fin}}(S)$3, or both are reduced and differ only by replacing the pseudo-unit valuation submonoid $\mathcal P_{\mathrm{fin}}(S)$4 by another reduced valuation monoid with the same quotient group, leaving the complementary semigroup fixed. Thus reduced finitary power monoids are rigid except for a specifically valuation-theoretic ambiguity (Rago, 30 Jan 2026).

Automorphism theory shows a related mixture of rigidity and residual symmetry. For the prototypical reduced additive example,

$\mathcal P_{\mathrm{fin}}(S)$5

the full automorphism group has exactly two elements: $\mathcal P_{\mathrm{fin}}(S)$6 The proof combines additive-combinatorial stabilization with an induction on the “boxing dimension,” the minimal number of intervals needed to cover a set (Tringali et al., 2023). This shows that power monoids can admit genuine non-inner symmetries, but in some foundational cases those symmetries are completely classifiable.

5. Numerical power monoids: prime-like scarcity and atomic abundance

Power monoids of numerical monoids display an especially sharp contrast between prime-like behavior and atomic behavior. For the unrestricted power monoid of $\mathcal P_{\mathrm{fin}}(S)$7, $\mathcal P_{\mathrm{fin}}(S)$8 is a cancellative prime element and

$\mathcal P_{\mathrm{fin}}(S)$9

with no other prime elements. If $\{1_H\}$0 is a numerical monoid, then $\{1_H\}$1 has no prime elements and admits no nontrivial direct-sum decomposition. The full power monoid already determines the underlying numerical monoid $\{1_H\}$2, even though its Grothendieck group does not: for any numerical monoid $\{1_H\}$3,

$\{1_H\}$4

The restricted power monoid of $\{1_H\}$5 also has a complete description of divisor-closed submonoids, each of the form

$\{1_H\}$6

where $\{1_H\}$7 (Bienvenu et al., 2022).

The scarcity extends from prime to primal elements. Restricted numerical power monoids contain no primal elements at all, and among unrestricted numerical power monoids the only primal element is $\{1_H\}$8 in $\{1_H\}$9. This is the precise primality analogue of the earlier nonexistence theorem for absolute irreducibles (Aggarwal et al., 2024).

Atomicity behaves in the opposite direction. For a numerical monoid $$\mathcal P_{\mathrm{fin},\times}(H):=\{X\in \mathcal P_{\mathrm{fin}}(H):X\cap H^\times\neq\emptyset\},$$0, if one counts elements of $$\mathcal P_{\mathrm{fin},\times}(H):=\{X\in \mathcal P_{\mathrm{fin}}(H):X\cap H^\times\neq\emptyset\},$$1 by maximum, then the proportion of atoms tends to $$\mathcal P_{\mathrm{fin},\times}(H):=\{X\in \mathcal P_{\mathrm{fin}}(H):X\cap H^\times\neq\emptyset\},$$2. For the unrestricted power monoid $$\mathcal P_{\mathrm{fin},\times}(H):=\{X\in \mathcal P_{\mathrm{fin}}(H):X\cap H^\times\neq\emptyset\},$$3, the corresponding limit lies in $$\mathcal P_{\mathrm{fin},\times}(H):=\{X\in \mathcal P_{\mathrm{fin}}(H):X\cap H^\times\neq\emptyset\},$$4 and equals $$\mathcal P_{\mathrm{fin},\times}(H):=\{X\in \mathcal P_{\mathrm{fin}}(H):X\cap H^\times\neq\emptyset\},$$5 exactly when $$\mathcal P_{\mathrm{fin},\times}(H):=\{X\in \mathcal P_{\mathrm{fin}}(H):X\cap H^\times\neq\emptyset\},$$6 (Bienvenu et al., 2022). Thus almost all restricted finite sets are indecomposable as proper sumsets.

The fine structure of this density is captured by the blocks

$$\mathcal P_{\mathrm{fin},\times}(H):=\{X\in \mathcal P_{\mathrm{fin}}(H):X\cap H^\times\neq\emptyset\},$$7

The paper establishes exact small-cardinality formulas, general bounds, and the asymptotic law

$$\mathcal P_{\mathrm{fin},\times}(H):=\{X\in \mathcal P_{\mathrm{fin}}(H):X\cap H^\times\neq\emptyset\},$$8

for every fixed $$\mathcal P_{\mathrm{fin},\times}(H):=\{X\in \mathcal P_{\mathrm{fin}}(H):X\cap H^\times\neq\emptyset\},$$9. It further proves that for each $X+Y=\{x+y:x\in X,\ y\in Y\}$00, the sequence $X+Y=\{x+y:x\in X,\ y\in Y\}$01 is “almost unimodal,” and if $X+Y=\{x+y:x\in X,\ y\in Y\}$02 denotes the size of a uniformly random atom in $X+Y=\{x+y:x\in X,\ y\in Y\}$03, then for every fixed moment order $X+Y=\{x+y:x\in X,\ y\in Y\}$04,

$X+Y=\{x+y:x\in X,\ y\in Y\}$05

where $X+Y=\{x+y:x\in X,\ y\in Y\}$06 (Aggarwal et al., 2024). A plausible interpretation is that additive decomposability is so rare in the bulk that the size distribution of random atoms asymptotically matches the size distribution of unrestricted random subsets containing $X+Y=\{x+y:x\in X,\ y\in Y\}$07.

6. Broader arithmetic picture and open directions

Power monoids have become a central laboratory for non-cancellative factorization theory precisely because elementary setwise multiplication produces unexpectedly rich arithmetic. They encode additive combinatorics inside monoid arithmetic, and many of their sharpest results turn on sumset theorems rather than on classical divisor theory alone (Tringali, 17 Feb 2026). This dual character is visible throughout the subject: high non-cancellativity coexists with strong finiteness in reduced variants, and local set-addition geometry often determines global factorization behavior.

The system of sets of lengths remains a major frontier. A broad conjectural picture asserts that if $X+Y=\{x+y:x\in X,\ y\in Y\}$08 is not a torsion monoid, then every nonempty subset of the integers $X+Y=\{x+y:x\in X,\ y\in Y\}$09 should occur as a length set in $X+Y=\{x+y:x\in X,\ y\in Y\}$10. For the basic example $X+Y=\{x+y:x\in X,\ y\in Y\}$11, it is already known that $X+Y=\{x+y:x\in X,\ y\in Y\}$12, $X+Y=\{x+y:x\in X,\ y\in Y\}$13, and $X+Y=\{x+y:x\in X,\ y\in Y\}$14 occur for every $X+Y=\{x+y:x\in X,\ y\in Y\}$15, and Reinhart proved that the monoid is fully elastic. It remains open whether unions of sets of lengths or of minimal length sets are eventually intervals (Tringali, 17 Feb 2026).

Several structural problems remain similarly active. The automorphism theorem for $X+Y=\{x+y:x\in X,\ y\in Y\}$16 led to the conjecture that the reduced power monoid of any proper numerical monoid has trivial automorphism group (Tringali et al., 2023). The isomorphism problem is now completely understood for commutative cancellative monoids, but the classification itself shows that reduced valuation behavior is a genuine obstruction rather than an accident (Rago, 30 Jan 2026). On the quantitative side, the survey highlights the unimodality conjecture for counting $X+Y=\{x+y:x\in X,\ y\in Y\}$17-element atoms with bounded maximum in numerical settings, a question that sits exactly at the boundary of factorization theory and additive combinatorics (Tringali, 17 Feb 2026).

Taken together, these developments show that power monoids are neither a minor variant of classical monoid constructions nor a pathological fringe case. They form a coherent domain in which subset multiplication, order, divisibility, and combinatorial growth interact at full strength, and they continue to drive both new algebraic classification results and new versions of factorization theory adapted to intrinsically non-cancellative arithmetic.