
(2) for every x, y ∈Xand U∈ T such that x◦
y⊆U, there exist Ux, Uy∈ T containing x, y
respectively, such that Ux◦Uy⊆U;
(3) for every x, y ∈Xand U∈ T such that x◦
y⊆U, there exist Ux, Uy∈ T containing x, y
respectively,such that a◦b⊆Ufor any a∈Ux
and b∈Uy.
Let (V, +,◦, K)be hypervector space over a topolog-
ical field Kand Tbe a topology on V. In the follow-
ing we use the topology T∗on P∗(V)and the product
topology on V×V.
3 Topological Hypervector Spaces
In this section we introduce the concept of topological
hypervector spaces and study some their properties.
Definition 6 Let (V, +,◦, K)be a hypervector space
over a topological filed Kand (V, T)be a topological
space. Then (V, +,◦, K, T)is said to be a topological
hypervector space (thvs)
if the operations + : V×V→V, (x, y)7→ x+y,
i:V→V, x 7→ −xand the hyperoperation
◦:K×V→P∗(V),(a, x)7→ a◦xare continuous.
Example 7 Every topological vector space
(V, +,·, K, T)with hyperoperation a◦x={a·x}is
a topological hypervector space over K.
Example 8 Every hypervector space (V, +,◦, K)
with trivial topology Tis a topological hypervector
space. Since, if we have T={∅, V }then
T∗={∅, SV}={∅, P ∗(V)}.
Example 9 Let K=V=Z2={0,1}. Then
(V, +,◦, K)is a hypervector space, where a◦x=
{0,1}for any a∈Kand x∈V. Let T=
{∅,{0},{1}, V }be a topology on V=K. We have
T∗={∅,{{0}},{{1}},{{0},{1}}, P ∗(V)}. It is
clear that Vis a topological hypervector space.
Example 10 By considering the external hyperoper-
ation ◦:R×R2→P∗(R2), a◦(x, y) = a·x×Rthen
(R2,+,◦, R)is a strongly distributive hypervector
space. The family B={(x, y) : a < x < b, y ∈R}
is a base for a topology on R. Then (R2,+,◦, R, T)
is a topological hypervector space.
Example 11 Let:
◦:R×R→P∗(R), a ◦x={a·x, −a·x}
be a external hyperoperation on R.Then (R, +,◦, R)
is a hypervector space, but it is neither the right dis-
tributive nor the left distributive. With standard topol-
ogy on R, (R, +,◦, R, T)is a topological hypervec-
tor space.
Example 12 Let:
◦:R×R→P∗(R), a ◦x={a·x, −a·x, 0}
be a external hyperoperation on R. Then (R, +,◦, R)
is a hypervector space, but it is neither the right dis-
tributive nor the left distributive. With standard topol-
ogy on V=Rand discrete topology on K=R,Vis
a topological hypervector space.
Topological hypervector spaces are a generalization of
topological vector spaces but some characteristics of
topological vector spaces are not valid in topological
hypervector spaces. If Vis a thvs, (V, +) is a topo-
logical group.
Lemma 13 Let Vbe a thvs. Then
(1) for fixed x∈V, the map y7→ x+yis a homeo-
morphism of Vonto V;
(2) if Uis open and x∈V, then x+Uis open; if U
is open and Ais any subset of V, then A+Uis
open;
(3) for fixed a∈K, the map x7→ a◦xis continuous,
but not necessarily open. In Example 11, U=
(2,3) is open and 2◦(2,3) = (−6,−4) ∪(4,6)
is also open, but in the Example 12, U= (2,3)
is open and 2◦(2,3) = (−6,−4) ∪ {0} ∪ (4,6)
is not open in R.
The complete parts were introduced for the first time
by Koskas [24]. Then, this concept was studied by
many authors. Let (V, +,◦, K)be a hypervector
space over Kand Abe a nonempty subset of V. We
say that Ais a complete part of V, if for nonzero nat-
ural number n, for all a1, . . . , anof K, and for all
x1, . . . , xnof V, the following implication holds:
A∩
n
X
i=1
ai◦xi6=∅=⇒
n
X
i=1
ai◦xi⊆A.
Theorem 14 Let Vbe a thvs, A ⊆Vand Ube an
open subset of V, such that Uis a complete part of V.
Then A⊆a−1◦Uif and only if a◦A⊆Ufor all
a∈K.
Suppose that A⊆a−1◦Uand x∈A. So x∈a−1◦U,
and there exists u∈U, such that x∈a−1◦uthus,
a◦x⊆a◦(a−1◦u) = 1◦u. We have u∈1◦u, u ∈U,
which implies that 1◦u⊆Usince Uis complete part.
Therefore a◦x⊆U.
Conversely, suppose that a◦A⊆Uand a∈K.
Then, we have A⊆a−1◦(a◦A)⊆a−1◦U.
Theorem 15 Let Ube an open subset of a thvs, such
that Uis a complete part. Then
PROOF
DOI: 10.37394/232020.2023.3.7
Reza Ameri, M. Hamidi, A. Samadifam