The Riesz capacity of a nonempty compact set
measures how widely a probability measure
can spread over
, using pairwise distances in
its energy. In a normalization homogeneous of degree one, for
it is
|
(1)
|
where
is the set of Borel probability measures on
,
is the Euclidean norm,
and infinite minimum energy gives capacity zero. For
, replace the infimum
by a supremum; at
, use the exponential of the negative minimum logarithmic
energy, giving logarithmic capacity (Clark
and Laugesen 2025). Translation and orthogonal
transformations preserve capacity, while multiplying
by
multiplies its capacity by
. The kernel parameter
differs from the exponent in variational
-capacity.
A capacitary Brunn-Minkowski inequality takes the form
|
(2)
|
where addition is the Minkowski sum. Clark and Laugesen (2026a) proved this for all nonempty compact and every
. For
, capacity is
times the diameter, and
equality always holds on the line. For
, with at least one
capacity positive when
, equality holds exactly
when a summand is a singleton or both summands are closed intervals.
In , Clark and Laugesen (2026b) proved the inequality
for
and arbitrary nonempty compact summands. When both capacities are positive, equality
holds exactly when the summands are convex and positively
homothetic, meaning
for some
and
. The range
remains conjectural for
. An extension to all negative
fails in higher dimensions: in the plane,
the perpendicular two-point sets
and
violate it for every
.
For every , a weaker bound holds for arbitrary nonempty compact
summands:
|
(3)
|
where
(Clark and Laugesen 2026b). This includes parameter ranges where the stronger inequality
is open or false. The authors report ChatGPT assistance with developing the mathematical
arguments and checking calculations (Clark and Laugesen 2026ab).