The Minkowski sum
of two sets
and
in a vector space is given
by
.
The Minkowski sum of two disks centered at and
with radii
and
, respectively, is given by
the disk centered at
with radius
. The Minkowski sum of two balls
is given similarly.
If and
are polyhedra, then
is a polyhedron and every
extreme point of
is the sum of an extreme point
in
and an extreme
point in
.
For example, taking the Minkowski sums of the following pairs of Platonic solids in dual position but with unit edge lengths give the following polyhedra.
| regular tetrahedron and regular tetrahedron | cuboctahedron |
| cube and regular octahedron | small rhombicuboctahedron |
| regular dodecahedron and regular icosahedron | small rhombicosidodecahedron |
For semialgebraic sets , membership can be written as
Quantifier elimination removes the existential quantifiers to give a semialgebraic set
description of
(Trott 2006). If
is a convex body and
is a closed ball centered at
the origin, then
is a parallel body. For
translated sets,
iff
, which expresses an intersection test as
membership in a Minkowski difference.
The Minkowski sum operation is implemented in the Wolfram Language as RegionDilation.
For compact convex sets, the volume of a scaled Minkowski sum is a polynomial
in the nonnegative parameters
. Its coefficients
define mixed volume (Schneider 2014).