The isoperimetric inequality for a plane figure of area
and perimeter
states that
where
is known as the isoperimetric quotient.
The equation becomes an equality
only for a circle.
At fixed area, the isoperimetric inequality gives a lower bound for the perimeter. Reverse isoperimetric inequalities instead give upper bounds under additional geometric restrictions. Convexity alone is insufficient, since rectangles of fixed area can have arbitrarily large perimeter.
For a nondegenerate triangle with side lengths
,
let
be a convex body whose area
is a fraction
of the area of
.
Ait El Manssour et al. (2026) proved
where
is the perimeter. Equality is attained by retaining
the longest side of
and placing the third vertex at
fraction
along the shortest side from their common endpoint, or by a congruent triangle
contained in
.
A curvature constraint leads to a different reverse inequality. The Borisenko conjecture, proved by Drach and Tatarko (2026), states that lambda-convex lenses minimize volume at fixed surface area among lambda-convex bodies.
For a complete simply connected Riemannian manifold with sectional curvature at most , the generalized
Cartan-Hadamard conjecture compares the isoperimetric profile with that of the
simply connected constant-curvature model. Agnoletto et al. (2026) reported
this inequality in all dimensions, together with equality rigidity. The argument
adapts analytic techniques from an earlier OpenAI manuscript and requires independent
assessment.