Difference between revisions of "Ptolemy's Inequality"
m (I think that ABCD doesn't have to be cyclic, PM me if I'm incorrect.) |
m (→Theorem) |
||
Line 10: | Line 10: | ||
</center> | </center> | ||
− | with equality for any quadrilateral <math>ABCD</math> with diagonals <math>AC </math> and <math>BD </math>. | + | with equality for any cyclic quadrilateral <math>ABCD</math> with diagonals <math>AC </math> and <math>BD </math>. |
This also holds if <math>A,B,C,D</math> are four points in space not in the same plane, but equality can't be achieved. | This also holds if <math>A,B,C,D</math> are four points in space not in the same plane, but equality can't be achieved. |
Revision as of 10:53, 27 February 2020
Ptolemy's Inequality is a famous inequality attributed to the Greek mathematician Ptolemy.
Contents
Theorem
The inequality states that in for four points in the plane,
,
with equality for any cyclic quadrilateral with diagonals
and
.
This also holds if are four points in space not in the same plane, but equality can't be achieved.
Proof for Coplanar Case
We construct a point such that the triangles
are similar and have the same orientation. In particular, this means that
.
But since this is a spiral similarity, we also know that the triangles are also similar, which implies that
.
Now, by the triangle inequality, we have . Multiplying both sides of the inequality by
and using
and
gives us
,
which is the desired inequality. Equality holds iff. ,
, and
are collinear. But since the triangles
and
are similar, this would imply that the angles
and
are congruent, i.e., that
is a cyclic quadrilateral.
Outline for 3-D Case
Construct a sphere passing through the points and intersecting segments
and
. We can now prove it through similar triangles, since the intersection of a sphere and a plane is always a circle.
Proof for All Dimensions?
Let any four points be denoted by the vectors .
Note that
.
From the Triangle Inequality,
.
Note about Higher Dimensions
Similar to the fact that that there is a line through any two points and a plane through any three points, there is a three-dimensional "solid" or 3-plane through any four points. Thus in an n-dimensional space, one can construct a 3-plane through the four points and the theorem is trivial, assuming the case has already been proven for three dimensions.