Neutrosophic Knowledge, Volume 8, 2025

University of New Mexico

Generalizing Multidimensional Networks:

A Framework of Multidimensional HyperNetworks and
SuperHyperNetworks
Takaaki Fujita1 ∗ , Arif Mehmood 2 , and Volkan Duran 3

1
Independent Researcher, Tokyo, Japan. Takaaki.fujita060@[Link]
2
Department of Mathematics, Institute of Numerical Sciences, Gomal University, Dera Ismail Khan 29050,
KPK, Pakistan (mehdaniyal@[Link])
3
Department of Computer Engineering Igdır University. Turkey. ([Link]@[Link])
∗
Correspondence: Takaaki.fujita060@[Link]

Received: 11 03, 2025; Accepted: 11 26, 2025

Abstract. Graph theory studies the mathematical structures of vertices and edges to model relationships and
connectivity [1]. Hypergraphs extend this framework by allowing hyperedges to connect arbitrarily many vertices
at once [2], and superhypergraphs further generalize hypergraphs via iterated powerset constructions to capture
hierarchical linkages among edges [3,4]. Weighted multidimensional networks model nodes connected by edges in
multiple layers, assigning each edge a weight to quantify relationship strength in its specific dimension(cf. [5,6]).
In this paper, we extend these ideas using hypergraphs and superhypergraphs to introduce and formalize
multidimensional hypernetworks and multidimensional superhypernetworks.
Keywords: Superhypergraph, Hypergraph, Multidimensional Network, Multidimensional HyperNetwork.
Multidimensional SuperHyperNetwork
—————————————————————————————————————————-

1. Preliminaries

We begin by fixing notation and recalling key concepts that underlie our constructions.
Unless otherwise noted, all graph-based objects in this paper are finite, simple, and undirected.
For more extensive discussions of these topics, the reader is referred to the standard literature.

T. Fujita et al., Generalizing Multidimensional Networks

Neutrosophic Knowledge, Volume 8, 2025 63 of 79

1.1. Hypergraphs and SuperHyperGraphs

A hypergraph allows edges—called hyperedges—to link any number of vertices at once, cap-
turing higher-order relationships beyond pairwise connections [2, 7–10]. A SuperHyperGraph
builds on this by applying the powerset operation repeatedly, so that edges themselves can be
nested in multiple layers, representing hierarchical groupings of interactions [3, 4, 11, 12]. In
what follows, the integer n ≥ 0 always indicates the number of times the powerset is iterated.

Definition 1.1 (Ground Set). Let S be a finite ground set of elements under consideration.
All subsequent constructions—subsets, powersets, and iterated powersets—are formed from
S.

Definition 1.2 (Powerset). (cf. [13–15]) Given a set S, its powerset P(S) is the collection of
all subsets of S, including the empty set and S itself:

P(S) = { A | A ⊆ S}.

Definition 1.3 (Hypergraph). [2, 7] A hypergraph H = (V, E) consists of

• A finite vertex set V .
• A finite set E of nonempty subsets of V , each called a hyperedge.
By design, hypergraphs can model relationships that involve more than two vertices at once.

Definition 1.4 (n-th Iterated Powerset). [16–18] For a set X, define its iterated powersets
by


P0 (X) = X, Pk+1 (X) = P Pk (X) (k ≥ 0).
Thus Pn (X) is the result of applying the powerset operation n times.

Definition 1.5 (n-SuperHyperGraph). [19,20] Fix a finite ground set V0 . Let P k (V0 ) denote
the k-th iterated powerset as above. An n-SuperHyperGraph is an ordered pair

SuHG(n) = (V, E), with V, E ⊆ P n (V0 ),

where elements of V are called n-supervertices and elements of E are n-superedges, each
superedge being a nonempty subset of V . This structure captures interactions at up to n
nested levels.

Example 1.6 (Collaborative Research 2-SuperHyperGraph). Consider a research laboratory

with five investigators:

V0 = {Yuta, Hiroko, Shinya, Taka, Eve}.

They form three project teams (first-level groups):

T1 = {Yuta, Hiroko}, T2 = {Shinya, Taka}, T3 = {Hiroko, Eve}.

T. Fujita et al., Generalizing Multidimensional Networks
Neutrosophic Knowledge, Volume 8, 2025 64 of 79

These teams themselves are organized into consortia (second-level supervertices):

C1 = {T1 , T2 }, C2 = {T2 , T3 }, C3 = {T1 , T3 }.

Taking n = 2, we have
V = { C1 , C2 , C3 } ⊆ P 2 (V0 ).
We record collaborative initiatives as 2-superedges (each a nonempty subset of V ):

E = {C1 }, {C2 }, {C3 }, {C1 , C2 }, {C2 , C3 } .

Here
• {C1 } represents an internal workshop run by Consortium 1.
• {C1 , C2 } denotes a joint symposium between Consortia 1 and 2.
• {C2 , C3 } denotes collaborative grant applications between Consortia 2 and 3.
Thus
SuHG(2) = ( V, E )
is a 2-SuperHyperGraph capturing collaborations at the level of researchers → teams → con-
sortia.

1.2. Weighted Multidimensional Network

Weighted multidimensional network models nodes connected by edges in several distinct lay-
ers, assigning each edge a weight to quantify relationship strength along its specific dimension
(cf. [21–23]). Related concepts include Multilayer Networks and Multiplex Networks, which
are also well-studied in the literature(cf. [24–27]).

Definition 1.7 (Unweighted Multidimensional Network). Let V be a finite set of nodes and
let D be a finite set of dimensions (or layers). An unweighted multidimensional network is a
triple
G = (V, E, D),
where
• E ⊆ V × V × D is a set of edges of the form (u, v, d), indicating an (undirected)
connection between u, v ∈ V in dimension d ∈ D;
• for each d ∈ D and each unordered pair {u, v} ⊆ V , at most one edge (u, v, d) appears
in E;
• if the network is directed, we instead allow (u, v, d) and (v, u, d) to be distinct elements
of E.
We say that two nodes u, v are adjacent in dimension d if (u, v, d) ∈ E. The degree of v in
dimension d is
kd (v) = { u ∈ V : (u, v, d) ∈ E} .
T. Fujita et al., Generalizing Multidimensional Networks
Neutrosophic Knowledge, Volume 8, 2025 65 of 79

Example 1.8 (Urban Multimodal Transport Network). Let the set of stops in a city be

V = {S1 , S2 , S3 , S4 , S5 }.

There are three transport modes (dimensions):

D = {bus, metro, tram}.

We represent direct connections by edges tagged with the mode of travel:

E = {(S1 , S2 , bus), (S2 , S3 , bus),

(S1 , S4 , metro), (S4 , S5 , metro),

(S2 , S5 , tram), (S3 , S4 , tram)}.

No more than one connection appears for each pair and mode. Thus

G = (V, E, D)

is an unweighted multidimensional network modeling the city’s multimodal transport system:

• (S1 , S2 , bus) and (S2 , S3 , bus) are direct bus routes.
• (S1 , S4 , metro) and (S4 , S5 , metro) are metro lines.
• (S2 , S5 , tram) and (S3 , S4 , tram) are tram connections.
Passengers can analyze paths that switch modes by traversing edges in different dimensions.

Definition 1.9 (Weighted Multidimensional Network). Let V , D be as in Definition 1.7, and

let
E ⊆ V ×V ×D×R

be a set of weighted edges (u, v, d, w), where w ∈ R is the weight of the connection between u
and v in dimension d. The quadruple

G = (V, E, D, w)

is called a weighted multidimensional network. The strength of node v in dimension d is

X
sd (v) = w.
(u,v,d,w)∈E

Remark 1.10 (Adjacency Tensor). In canonical tensor notation, one may encode any weighted
iα
multidimensional network G = (V, E, D) by a rank-4 adjacency tensor M = (Mjβ )i,j∈V, α,β∈D ,
where 
w, if (i, j, α, w) ∈ E and α = β,
iα
Mjβ =
0, otherwise.

This representation readily extends standard matrix-based methods (e.g. centrality, spectral
analysis) to the multidimensional setting.
T. Fujita et al., Generalizing Multidimensional Networks
Neutrosophic Knowledge, Volume 8, 2025 66 of 79

Example 1.11 (Global Trade Weighted Multidimensional Network). Consider five countries:

V = {US, CN, DE, JP, IN}.

We track trade in three categories (dimensions):

D = {oil, electronics, agriculture}.

Define weighted edges (u, v, d, w) where w is annual export volume (in billions USD):
E = {(US, CN, oil, 50), (CN, US, electronics, 200),

(DE, US, agriculture, 30), (JP, DE, electronics, 120),

(IN, JP, agriculture, 25), (US, IN, oil, 15)}.

Then
G = (V, E, D, w)

is a weighted multidimensional network representing global trade:

• (US, CN, oil, 50) means the US exports $50 billion of oil to China.
• (CN, US, electronics, 200) means China exports $200 billion of electronics to the US.
• (DE, US, agriculture, 30) means Germany exports $30 billion of agricultural products
to the US.
• And so on for the other entries.
One can compute each country’s strength sd (v) in category d by summing outgoing volumes
in that dimension.

1.3. HyperNetwork and SuperhyperNetwork

A hypernetwork is a graph generalization where hyperedges connect any number of

nodes, enabling modeling of multiway relationships beyond pairwise edges [28]. An n-
superhypernetwork uses vertices and hyperedges drawn from the n-th iterated powerset of
a base node set to model nested, hierarchical groupings [29]. The definitions of HyperNetwork
and SuperhyperNetwork are presented below [28].

Definition 1.12 (Hypernetwork). [28] A hypernetwork is an ordered triple

H = V, E, w

where
• V is a nonempty finite set of nodes;
• E ⊆ P(V ) \ {∅} is the set of hyperedges, each hyperedge e ∈ E being a nonempty subset
of nodes (allowing multi-node interactions);
• w : E → R≥0 is a weight or attribute function on hyperedges (omitted if unweighted).
T. Fujita et al., Generalizing Multidimensional Networks
Neutrosophic Knowledge, Volume 8, 2025 67 of 79

A directed hypernetwork may be defined by replacing E ⊆ P(V ) with a set of ordered tuples
of nodes or by equipping each e ∈ E with a head-tail partition. One can further add a node-
labeling ℓV : V → LV and a hyperedge-labeling ℓE : E → LE to record types or properties.

Definition 1.13 (n-SuperHypernetwork). [28, 30, 31] Let V0 be a finite base set of nodes.
Define the n-th iterated powerset recursively by

P 0 (V0 ) = V0 , P k+1 (V0 ) = P P k (V0 )

(k ≥ 0).

An n-superhypernetwork is a tuple

N (n) =


V, E, w

where
• V ⊆ P n (V0 ) is a finite set of n-supernodes;
• E ⊆ P(V ) is a finite set of n-superedges, each superedge e ∈ E being a nonempty subset
of V ;
• w : E → R≥0 is an optional weight function assigning a nonnegative real weight (or
confidence) to each superedge.
In other words, both vertices and hyperedges of the network are drawn from the n-th powerset
of the base node set, capturing up to n levels of hierarchical grouping.

Example 1.14 (Research Consortium 2-SuperHypernetwork). Consider a small research com-

munity of four investigators:

V0 = {Yuta, Hiroko, Shinya, Taka}.

They form three laboratories (first-level teams):

L1 = {Yuta, Hiroko}, L2 = {Shinya, Taka}, L3 = {Hiroko, Shinya}.

These labs themselves collaborate to form two consortia (second-level supernodes):

C1 = {L1 , L2 }, C2 = {L2 , L3 }.

Thus, taking n = 2, we have

V = { C1 , C2 } ⊆ P 2 (V0 ).

We model funded programs as 2-superedges (each a nonempty subset of V ):


E = {C1 }, {C2 }, {C1 , C2 } .

Assigning each program’s annual budget (in million USD) as a weight,

w({C1 }) = 5.0, w({C2 }) = 3.5, w({C1 , C2 }) = 1.2.

T. Fujita et al., Generalizing Multidimensional Networks
Neutrosophic Knowledge, Volume 8, 2025 68 of 79

Therefore
N (2) = V, E, w

is a 2-SuperHypernetwork capturing:
• the hierarchy of individual investigators → labs → consortia,
• the funding programs that each consortium (and joint consortia pair) secures.

2. Result: Weighted Multidimensional Hypernetwork

Weighted multidimensional hypernetwork extends networks by grouping nodes into hyper-

edges across distinct dimensions, assigning weights to each unique hyperedge–dimension pair
to represent multiway relationship intensities.

Definition 2.1 (Weighted Multidimensional Hypernetwork). Let V be a finite set of nodes and
let D be a finite set of dimensions. A weighted multidimensional hypernetwork is a quadruple


H = V, E, D, w ,

where


• E ⊆ P(V ) \ {∅} × D is a set of dimension-tagged hyperedges, each element (e, d) ∈ E
consisting of a nonempty vertex subset e ⊆ V and a dimension d ∈ D;
• w : E → R is a weight function assigning each (e, d) ∈ E a real weight w(e, d).

Example 2.2 (Corporate Team Communication Hypernetwork). Consider a small engineering

team with four members:

V = {Yuta, Hiroko, Shinya, Taka}.

They use three communication channels:

D = {Slack, Email, Zoom}.

We model group communications as dimension-tagged hyperedges, weighted by the average

number of messages (or minutes of meeting) per week:




E = { Yuta, Hiroko, Shinya}, Slack , { Yuta, Taka}, Email , { Hiroko, Shinya, Taka}, Zoom .

Define the weight function w : E → R≥0 by

w {A, B, C}, Slack = 120, w {A, D}, Email = 30, w {B, C, D}, Zoom = 180,

where
w({Yuta, Hiroko, Shinya}, Slack) = 120 (120 Slack messages/week among Yuta, Hiroko, Shinya),

w({Yuta, Taka}, Email) = 30 (30 email exchanges/week between Yuta and Taka),

w({Hiroko, Shinya, Taka}, Zoom) = 180 (180 minutes of Zoom meetings/week among Hiroko, Shinya, Taka).
T. Fujita et al., Generalizing Multidimensional Networks
Neutrosophic Knowledge, Volume 8, 2025 69 of 79

Thus


H = V, E, D, w

is a weighted multidimensional hypernetwork capturing the team’s collaborative interactions

across different channels.

Example 2.3 (Global Supply Chain Hypernetwork). Consider a simplified supply chain with
six locations:
V = {F1 , F2 , W1 , W2 , R1 , R2 },

where Fi are factories, Wj warehouses, and Rk retailers. There are three product categories
(dimensions):
D = {electronics, furniture, food}.

We record weekly shipment groupings as dimension-tagged hyperedges, weighted by units

shipped per week:
E = {({ F1 , W1 , R1 }, electronics),

({ F2 , W2 , R2 }, electronics),

({ F1 , W1 , W2 , R2 }, furniture),

({ F2 , W2 , R1 , R2 }, food)}.
Assign the following weights w : E → R≥0 :


w {F1 , W1 , R1 }, electronics = 1000,


w {F2 , W2 , R2 }, electronics = 800,


w {F1 , W1 , W2 , R2 }, furniture = 500,


w {F2 , W2 , R1 , R2 }, food = 1200.

Thus


H = V, E, D, w

is a weighted multidimensional hypernetwork modeling weekly shipment volumes of electronics,

furniture, and food across factories, warehouses, and retailers in the supply chain.

Theorem 2.4 (Generalization of Weighted Multidimensional Networks and Hypernetworks). (1)

Every weighted multidimensional network N = (V, E, D, wN ) (Definition 1.9) induces
a weighted multidimensional hypernetwork
 
′
HN = V, { ({u, v}, d) | (u, v, d) ∈ E}, D, wN ,

′ ({u, v}, d) = w (u, v, d). In particular, H

where wN N N is exactly the hypernetwork

whose hyperedges are the size-two edges of N .

T. Fujita et al., Generalizing Multidimensional Networks
Neutrosophic Knowledge, Volume 8, 2025 70 of 79

(2) Every weighted hypernetwork H = (V, E, {d0 }, wH ) with a single dimension D = {d0 }
arises as a special case of a weighted multidimensional hypernetwork by tagging all
hyperedges with d0 .
Hence the class of weighted multidimensional hypernetworks strictly generalizes both
weighted multidimensional networks and weighted hypernetworks.

Proof. (1) Let N = (V, E, D, wN ) be a weighted multidimensional network, so

E ⊆ V × V × D, wN : E → R.

Define

E= ({u, v}, d) (u, v, d) ∈ E ⊆ (P(V ) \ {∅}) × D,

and set
′
wN ({u, v}, d) = wN (u, v, d).
′ ) satisfies the definition of a weighted multidimen-
By construction HN = (V, E, D, wN
sional hypernetwork. The injective correspondence (u, v, d) 7→ ({u, v}, d) shows that
no information is lost.
(2) Let H = (V, E, {d0 }, wH ) be a weighted hypernetwork in the usual sense: E ⊆ P(V )\{∅}
and wH : E → R. Embed it into a weighted multidimensional hypernetwork by

Ee = (e, d0 ) | e ∈ E , w(e,
e d0 ) = wH (e).

Again, this construction is bijective and respects all weights.

Since every weighted multidimensional network and every weighted hypernetwork injects
into the class of weighted multidimensional hypernetworks, and there exist hyperedges of
arbitrary cardinality or multiple dimensions that cannot be realized in the narrower settings,
H indeed strictly generalizes both.

Definition 2.5 (Dimension-Restricted Hypernetwork). For each d ∈ D, define

Ed = { e ⊆ V : (e, d) ∈ E}, wd (e) = w(e, d).

Then Hd = (V, Ed , wd ) is the dimension-d restriction of H.

Theorem 2.6 (Restriction Yields a Weighted Hypernetwork). For each d ∈ D, the restricted
structure Hd is itself a weighted hypernetwork in the usual sense.

Proof. By definition Ed ⊆ P(V ) \ {∅} and wd assigns real weights to each hyperedge e ∈ Ed .
Thus Hd meets the requirements of a weighted hypernetwork.

T. Fujita et al., Generalizing Multidimensional Networks

Neutrosophic Knowledge, Volume 8, 2025 71 of 79

Definition 2.7 (Aggregated Hypernetwork). Define the aggregation of H across all dimensions
by
 X
Eagg = e ⊆ V : ∃ d ∈ D, (e, d) ∈ E , wagg (e) = w(e, d).
d∈D
(e,d)∈E

Then Hagg = (V, Eagg , wagg ) is called the aggregated hypernetwork.

Theorem 2.8 (Aggregation Preserves Hypernetwork Structure). Hagg is a weighted hyper-

network whose edge–weight function wagg is well-defined and nonnegative.

Proof. Clearly Eagg ⊆ P(V )\{∅}. Since each (e, d) contributes a real weight, the sum wagg (e) ≥
0 is well-defined. Hence Hagg satisfies the definition of a weighted hypernetwork.

Definition 2.9 (Primal Graph). The primal graph G = (V, A) of H is the simple graph with

A = {u, v} ⊆ V : ∃ (e, d) ∈ E, {u, v} ⊆ e .

Theorem 2.10 (Primal Connectivity Criterion). The primal graph G is connected if and only
if for every nontrivial partition V = V1 ∪ V2 , there exists some (e, d) ∈ E with e ∩ V1 ̸= ∅ and
e ∩ V2 ̸= ∅.

Proof. (⇒) If G is connected, no partition can isolate two sets without an edge crossing, so
some hyperedge in E must bridge V1 and V2 . (⇐) Conversely, if every partition is bridged by
at least one hyperedge, the primal graph cannot split into disconnected components, hence G
is connected.

Definition 2.11 (Node Strength and Dimension Relevance). For each node v ∈ V define its
dimension-d strength
X
sd (v) = w(e, d),
(e,d)∈E
v∈e

and its total strength

X
s(v) = sd (v).
d∈D

The relevance of dimension d at v is

sd (v)
R(v, d) = .
s(v)

Theorem 2.12 (Relevance Forms a Probability Distribution). For each v ∈ V with s(v) > 0,
P
the values {R(v, d)}d∈D satisfy 0 ≤ R(v, d) ≤ 1 and d∈D R(v, d) = 1.
T. Fujita et al., Generalizing Multidimensional Networks
Neutrosophic Knowledge, Volume 8, 2025 72 of 79

Proof. By construction sd (v) ≥ 0 and s(v) > 0. Thus 0 ≤ R(v, d) ≤ 1. Moreover

X X sd (v) 1 X
R(v, d) = = sd (v) = 1,
s(v) s(v)
d∈D d∈D d∈D

as required.

3. Result: Weighted Multidimensional SuperHypernetwork

Weighted multidimensional superhypernetwork generalizes hypernetworks by using iterated

powersets to create distinct nested supernodes and superedges across dimensions, with weights
capturing hierarchical multiway relationship strengths.

Definition 3.1 (Weighted Multidimensional SuperHypernetwork). Let V0 be a finite base set

and let D be a finite set of dimensions. Fix a nonnegative integer n. Denote by

P 0 (V0 ) = V0 , P k+1 (V0 ) = P P k (V0 ) (k ≥ 0).

A weighted multidimensional n-superhypernetwork is a quadruple

S (n) = V, E, D, w ,

where
• V ⊆ P n (V0 ) is the set of n-supernodes;


• E ⊆ P(V ) \ {∅} × D is the set of dimension-tagged n-superedges, each (e, d) ∈ E
consisting of a nonempty n-supernode subset e ⊆ V in dimension d ∈ D;
• w : E → R assigns each (e, d) a real weight w(e, d).

Example 3.2 (Regional Multimodal Logistics SuperHypernetwork). Consider a logistics com-

pany serving five cities:
V0 = {A, B, C, D, E}.
There are three transportation modes (dimensions):

D = {road, rail, air}.

We choose n = 1, so

P 1 (V0 ) = P(V0 ) = { ∅, {A}, {B}, {C}, {D}, {E}, {A, B}, . . . , {A, B, C, D, E}}.

Define the set of 1–supernodes as two key regions:


V = {A, B}, {C, D, E} .

Each region groups nearby cities into a single operational cluster.

Next, we model both intra- and inter-region shipping via dimension-tagged superedges:





E = {A, B}, road , {A, B}, rail , {C, D, E}, air , {{A, B}, {C, D, E}}, rail .
T. Fujita et al., Generalizing Multidimensional Networks
Neutrosophic Knowledge, Volume 8, 2025 73 of 79

• ({A, B}, road): shipments by truck within Region {A, B}.

• ({A, B}, rail): rail shipments within Region {A, B}.
• ({C, D, E}, air): air shipments within Region {C, D, E}.
• ({{A, B}, {C, D, E}}, rail): inter-region rail shipments connecting the two regions.
We assign daily shipment volumes (in tons) as weights:


w {A, B}, road = 500,


w {A, B}, rail = 300,


w {C, D, E}, air = 200,


w {{A, B}, {C, D, E}}, rail = 150.
Thus
S (1) = V, E, D, w

is a weighted multidimensional 1-superhypernetwork capturing both hierarchical clustering of

cities into regions and the multimodal shipment volumes within and between those regions.

Example 3.3 (Corporate Communication SuperHypernetwork). Consider a small company

with four employees:
V0 = { Yuta, Hiroko, Shinya, Taka}.

There are three communication modes (dimensions):

D = {email, meeting, codeReview}.

We take n = 2. Then

P 1 (V0 ) = {Yuta, Hiroko}, {Shinya, Taka}, {Hiroko, Shinya}, . . . ,



and
P 2 (V0 ) = P P 1 (V0 ) ,

whose elements are sets of teams. Choose three key teams:

T1 = {Yuta, Hiroko}, T2 = {Shinya, Taka}, T3 = {Hiroko, Shinya}.

Form two departments as 2-supernodes:

⊆ P 2 (V0 ).

V = DA = {T1 , T2 }, DB = {T2 , T3 }

Here DA groups Team 1 and Team 2; DB groups Team 2 and Team 3.

Next, define dimension-tagged 2-superedges (e, d) to capture both intra- and
inter-department communications:

E = ( {DA }, email ), ( {DA }, meeting ), ( {DB }, codeReview ), ( {DA , DB }, meeting ), ( {DA , DB }, email ) .
T. Fujita et al., Generalizing Multidimensional Networks
Neutrosophic Knowledge, Volume 8, 2025 74 of 79

We assign monthly communication volumes as weights w : E → R≥0 :

w {DA }, email = 1200,


w {DA }, meeting = 150,


w {DB }, codeReview = 300,


w {DA , DB }, meeting = 40,


w {DA , DB }, email = 800.

Thus
S (2) = V, E, D, w

is a weighted multidimensional 2-superhypernetwork modeling:

• intra-department email and meetings within DA ,
• code-review activity within DB ,
• cross-department email and joint meetings between DA and DB .

Theorem 3.4 (Generalization of Earlier Models). The class of weighted multidimensional

n-superhypernetworks (Definition 3.1) strictly generalizes each of the following:


(1) Weighted multidimensional networks V, E, D, wN , by taking n = 0 and restricting
all hyperedges to cardinality 2.


(2) Weighted multidimensional hypernetworks V, E, D, wH , by taking n = 0.


(3) Superhypernetworks V, E, wS (with a single trivial dimension), by taking D = {⋆}.

Proof. We exhibit embedding constructions for each case:

(1) From weighted multidimensional networks. Let N = (V, E, D, wN ) be a weighted

multidimensional network, so

E ⊆ V × V × D, wN : E → R.

Choose n = 0. Define

V ′ = V ⊆ P 0 (V0 ), E′ = P 0 (V0 ) \ {∅} × D,



({u, v}, d) (u, v, d) ∈ E ⊆

and set w′ ({u, v}, d) = wN (u, v, d). Then S ′ = V ′ , E ′ , D, w′ is a weighted multidimensional

0-superhypernetwork whose superedges are exactly the ordinary edges of N .

(2) From weighted multidimensional hypernetworks. Let H = (V, E, D, wH ) be a

weighted multidimensional hypernetwork with E ⊆ (P(V ) \ {∅}) × D. Again take n = 0 and
set
V ′ = V ⊆ P 0 (V0 ), E ′ = E, w ′ = wH .

Then S ′ = (V ′ , E ′ , D, w′ ) is a weighted multidimensional 0-superhypernetwork identical to H.

T. Fujita et al., Generalizing Multidimensional Networks
Neutrosophic Knowledge, Volume 8, 2025 75 of 79

(3) From superhypernetworks. Let S = (V, E, wS ) be an (unweighted or weighted) n-

superhypernetwork without multiple dimensions. Embed it by setting

D′ = {⋆}, V ′ = V ⊆ P n (V0 ), E ′ = { (e, ⋆) | e ∈ E}, w′ (e, ⋆) = wS (e).

Then V ′ , E ′ , D′ , w′ is a weighted multidimensional n-superhypernetwork reproducing S.

In each case, no information beyond simple relabeling is lost, and the restrictions (to n = 0,
to edge-cardinality 2, or to a singleton D) exhibit that the general definition indeed encom-
passes all earlier models. Moreover, allowing arbitrary n, arbitrary dimension set D, and
arbitrary superedge cardinalities strictly enlarges the modeling power, so the generalization is
proper.

Definition 3.5 (Dimension-Restricted SuperHypernetwork). Let S (n) = (V, E, D, w) be a

weighted multidimensional n-superhypernetwork. For each d ∈ D, define

Ed = { e ⊆ V | (e, d) ∈ E}, wd (e) = w(e, d).

(n)
Then Sd = (V, Ed , wd ) is called the dimension-d restriction of S (n) .

Example 3.6 (Dimension-Restricted Email SuperHypernetwork). Starting from the corporate

communication superhypernetwork of Example 3.3,

S (2) = V, E, D, w ,

with
V = {DA , DB }, D = {email, meeting, codeReview},
we form the email restriction by selecting
 
Eemail = e ⊆ V | (e, email) ∈ E = {DA }, {DA , DB } ,

and setting



wemail {DA } = 1200, wemail {DA , DB } = 800.
Thus the dimension-restricted superhypernetwork
(2)

Semail = V, Eemail , wemail

captures only the email traffic:

• {DA } with weight 1200 represents department DA ’s internal emails;
• {DA , DB } with weight 800 represents cross-department emails between DA and DB .
(n)
Theorem 3.7 (Dimension Restriction). For each d ∈ D, the triple Sd = (V, Ed , wd ) is itself
a weighted n-superhypernetwork (with a single dimension), inheriting all structural and weight
properties from S (n) .
T. Fujita et al., Generalizing Multidimensional Networks
Neutrosophic Knowledge, Volume 8, 2025 76 of 79

Proof. By construction, Ed ⊆ P n (V0 ) and wd : Ed → R. All supernodes remain in V ⊆ P n (V0 ),

(n)
and each superedge e ∈ Ed is nonempty. Therefore Sd satisfies Definition 3.1 with D = {d}.

Definition 3.8 (Primal Graph). The primal graph G(V, A) of S (n) = (V, E, D, w) is the simple
graph whose vertex set is V and whose edge set

A = {u, v} ⊆ V : ∃ (e, d) ∈ E, u ̸= v, {u, v} ⊆ e .

Theorem 3.9 (Primal Connectivity). The primal graph G of S (n) is connected if and only if
for every nontrivial partition V = V1 ∪V2 , there exists (e, d) ∈ E with e∩V1 ̸= ∅ and e∩V2 ̸= ∅.

Proof. ⇒: If G is connected, then no nontrivial partition can isolate vertices without an edge
between them. Any separating partition would contradict connectivity. ⇐: Conversely, if
every bipartition is bridged by some superedge, then the primal graph cannot decompose into
disconnected components, hence must be connected.

Definition 3.10 (Supernode Strength and Relevance). For v ∈ V , define its total strength
X
s(v) = w(e, d),
(e,d)∈E
v∈e

and its dimension-d strength

X
sd (v) = w(e, d).
(e,d)∈E
v∈e
The dimension relevance of d at v is
sd (v)
R(v, d) = , d ∈ D.
s(v)

Theorem 3.11 (Relevance Distribution). For each v ∈ V , the relevance values {R(v, d)}d∈D
form a probability distribution:
X
R(v, d) ∈ [0, 1], R(v, d) = 1.
d∈D
P
Proof. By definition sd (v) ≥ 0 and s(v) = d sd (v) > 0 whenever v lies in at least one
superedge. Hence 0 ≤ R(v, d) ≤ 1. Moreover,
X X sd (v) 1 X
R(v, d) = = sd (v) = 1.
s(v) s(v)
d∈D d∈D d∈D

T. Fujita et al., Generalizing Multidimensional Networks

Neutrosophic Knowledge, Volume 8, 2025 77 of 79

4. Conclusion

In this paper, we extended these ideas using hypergraphs and superhypergraphs to introduce
and formalize multidimensional hypernetworks and multidimensional superhypernetworks. In
future research, it is expected that further extensions will be developed based on Fuzzy Sets [32,
33], Intuitionistic Fuzzy Sets [34], Neutrosophic Sets [35–37], Quadri-Partitioned Neutrosophic
Sets [38], and Plithogenic Sets [39–41].

Funding

This study did not receive any financial or external support from organizations or individ-
uals.

Acknowledgments

We extend our sincere gratitude to everyone who provided insights, inspiration, and as-
sistance throughout this research. We particularly thank our readers for their interest and
acknowledge the authors of the cited works for laying the foundation that made our study
possible. We also appreciate the support from individuals and institutions that provided the
resources and infrastructure needed to produce and share this paper. Finally, we are grateful
to all those who supported us in various ways during this project.

Data Availability

This research is purely theoretical, involving no data collection or analysis. We encour-

age future researchers to pursue empirical investigations to further develop and validate the
concepts introduced here.

Ethical Approval

As this research is entirely theoretical in nature and does not involve human participants
or animal subjects, no ethical approval is required.

Conflicts of Interest

The authors confirm that there are no conflicts of interest related to the research or its
publication.

References

[1] Reinhard Diestel. Graph theory. Springer (print edition); Reinhard Diestel (eBooks), 2024.
[2] Claude Berge. Hypergraphs: combinatorics of finite sets, volume 45. Elsevier, 1984.
[3] Florentin Smarandache. Introduction to the n-SuperHyperGraph-the most general form of graph today.
Infinite Study, 2022.

T. Fujita et al., Generalizing Multidimensional Networks

Neutrosophic Knowledge, Volume 8, 2025 78 of 79

[4] Mohammad Hamidi, Florentin Smarandache, and Mohadeseh Taghinezhad. Decision Making Based on
Valued Fuzzy Superhypergraphs. Infinite Study, 2023.
[5] Michele Berlingerio, Michele Coscia, and Fosca Giannotti. Finding and characterizing communities in
multidimensional networks. In 2011 international conference on advances in social networks analysis and
mining, pages 490–494. IEEE, 2011.
[6] Gergely Honti and János Abonyi. Frequent itemset mining and multi-layer network-based analysis of rdf
databases. Mathematics, 9(4):450, 2021.
[7] Alain Bretto. Hypergraph theory. An introduction. Mathematical Engineering. Cham: Springer, 1, 2013.
[8] Yue Gao, Zizhao Zhang, Haojie Lin, Xibin Zhao, Shaoyi Du, and Changqing Zou. Hypergraph learning:
Methods and practices. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(5):2548–2566,
2020.
[9] Yue Gao, Yifan Feng, Shuyi Ji, and Rongrong Ji. Hgnn+: General hypergraph neural networks. IEEE
Transactions on Pattern Analysis and Machine Intelligence, 45(3):3181–3199, 2022.
[10] Yifan Feng, Haoxuan You, Zizhao Zhang, Rongrong Ji, and Yue Gao. Hypergraph neural networks. In
Proceedings of the AAAI conference on artificial intelligence, volume 33, pages 3558–3565, 2019.
[11] Takaaki Fujita. Advancing Uncertain Combinatorics through Graphization, Hyperization, and Uncertainiza-
tion: Fuzzy, Neutrosophic, Soft, Rough, and Beyond. Biblio Publishing, 2025.
[12] Masoud Ghods, Zahra Rostami, and Florentin Smarandache. Introduction to neutrosophic restricted su-
perhypergraphs and neutrosophic restricted superhypertrees and several of their properties. Neutrosophic
Sets and Systems, 50:480–487, 2022.
[13] Yasar Nacaroglu, Nihat Akgunes, Sedat Pak, and I Naci Cangul. Some graph parameters of power set
graphs. Advances & Applications in Discrete Mathematics, 26(2), 2021.
[14] Adel Al-Odhari. Neutrosophic power-set and neutrosophic hyper-structure of neutrosophic set of three
types. Annals of Pure and Applied Mathematics, 31(2):125–146, 2025.
[15] Thomas Jech. Set theory: The third millennium edition, revised and expanded. Springer, 2003.
[16] Ajoy Kanti Das, Rajat Das, Suman Das, Bijoy Krishna Debnath, Carlos Granados, Bimal Shil, and Rakhal
Das. A comprehensive study of neutrosophic superhyper bci-semigroups and their algebraic significance.
Transactions on Fuzzy Sets and Systems, 8(2):80, 2025.
[17] Florentin Smarandache. Foundation of superhyperstructure & neutrosophic superhyperstructure. Neutro-
sophic Sets and Systems, 63(1):21, 2024.
[18] Florentin Smarandache. The SuperHyperFunction and the Neutrosophic SuperHyperFunction (revisited
again), volume 3. Infinite Study, 2022.
[19] Mohammad Hamidi, Florentin Smarandache, and Elham Davneshvar. Spectrum of superhypergraphs via
flows. Journal of Mathematics, 2022(1):9158912, 2022.
[20] Florentin Smarandache. Extension of HyperGraph to n-SuperHyperGraph and to Plithogenic n-
SuperHyperGraph, and Extension of HyperAlgebra to n-ary (Classical-/Neutro-/Anti-) HyperAlgebra. Infi-
nite Study, 2020.
[21] Michele Coscia, Giulio Rossetti, Diego Pennacchioli, Damiano Ceccarelli, and Fosca Giannotti. ” you know
because i know” a multidimensional network approach to human resources problem. In Proceedings of the
2013 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, pages
434–441, 2013.
[22] Manlio De Domenico. More is different in real-world multilayer networks. Nature Physics, 19(9):1247–1262,
2023.
[23] Manlio De Domenico, Albert Solé-Ribalta, Emanuele Cozzo, Mikko Kivelä, Yamir Moreno, Mason A Porter,
Sergio Gómez, and Alex Arenas. Mathematical formulation of multilayer networks. Physical Review X,
3(4):041022, 2013.

T. Fujita et al., Generalizing Multidimensional Networks

Neutrosophic Knowledge, Volume 8, 2025 79 of 79

[24] Federico Battiston, Vincenzo Nicosia, and Vito Latora. Structural measures for multiplex networks. Phys-
ical Review E, 89(3):032804, 2014.
[25] Federico Battiston, Vincenzo Nicosia, and Vito Latora. The new challenges of multiplex networks: Measures
and models. The European Physical Journal Special Topics, 226(3):401–416, 2017.
[26] Alberto Aleta and Yamir Moreno. Multilayer networks in a nutshell. Annual Review of Condensed Matter
Physics, 10(1):45–62, 2019.
[27] Stefano Boccaletti, Ginestra Bianconi, Regino Criado, Charo I Del Genio, Jesús Gómez-Gardenes, Miguel
Romance, Irene Sendina-Nadal, Zhen Wang, and Massimiliano Zanin. The structure and dynamics of
multilayer networks. Physics reports, 544(1):1–122, 2014.
[28] Takaaki Fujita. Exploration of graph classes and concepts for superhypergraphs and n-th power mathe-
matical structures. Advancing Uncertain Combinatorics through Graphization, Hyperization, and Uncer-
tainization: Fuzzy, Neutrosophic, Soft, Rough, and Beyond, 3(4):512.
[29] Takaaki Fujita. Telecommunications hypernetwork and telecommunications superhypernetwork. Intelli-
gence Modeling in Electromechanical Systems, 2(1):16–31, 2025.
[30] Takaaki Fujita. Hypergraph and superhypergraph approaches in electronics: A hierarchical framework for
modeling power-grid hypernetworks and superhypernetworks. Journal of Energy Research and Reviews,
17(6):102–136, 2025.
[31] Shouxian Zhu. Neutrosophic n-superhypernetwork: A new approach for evaluating short video communi-
cation effectiveness in media convergence. Neutrosophic Sets and Systems, 85:1004–1017, 2025.
[32] Lotfi A Zadeh. Fuzzy sets. Information and control, 8(3):338–353, 1965.
[33] John N Mordeson and Premchand S Nair. Fuzzy graphs and fuzzy hypergraphs, volume 46. Physica, 2012.
[34] Krassimir T Atanassov. On intuitionistic fuzzy sets theory, volume 283. Springer, 2012.
[35] Florentin Smarandache. A unifying field in logics: Neutrosophic logic. In Philosophy, pages 1–141. American
Research Press, 1999.
[36] Said Broumi, Mohamed Talea, Assia Bakali, and Florentin Smarandache. Single valued neutrosophic
graphs. Journal of New theory, (10):86–101, 2016.
[37] Haibin Wang, Florentin Smarandache, Yanqing Zhang, and Rajshekhar Sunderraman. Single valued neu-
trosophic sets. Infinite study, 2010.
[38] R Radha, A Stanis Arul Mary, and Florentin Smarandache. Quadripartitioned neutrosophic pythagorean
soft set. International Journal of Neutrosophic Science (IJNS) Volume 14, 2021, page 11, 2021.
[39] Florentin Smarandache. Plithogenic set, an extension of crisp, fuzzy, intuitionistic fuzzy, and neutrosophic
sets-revisited. Infinite study, 2018.
[40] Takaaki Fujita and Florentin Smarandache. A Dynamic Survey of Fuzzy, Intuitionistic Fuzzy, Neutrosophic,
Plithogenic, and Extensional Sets. Neutrosophic Science International Association (NSIA), 2025.
[41] Fazeelat Sultana, Muhammad Gulistan, Mumtaz Ali, Naveed Yaqoob, Muhammad Khan, Tabasam Rashid,
and Tauseef Ahmed. A study of plithogenic graphs: applications in spreading coronavirus disease (covid-19)
globally. Journal of ambient intelligence and humanized computing, 14(10):13139–13159, 2023.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publi-
cations are solely those of the individual author(s) and contributor(s) and not of the publisher
and/or the editor(s). This publisher and/or the editor(s) disclaim responsibility for any injury
to people or property resulting from any ideas, methods, instructions or products referred to
in the content.

T. Fujita et al., Generalizing Multidimensional Networks