In the lecture I gave at the AIM workshop on Amoebas and tropical
geometry,
I defined some enumerative invariants of real algebraic convex
3-manifolds. For example, through a generic configuration of real
points
in the complex projective space, there passes only finitely many
irreducible real rational curves of degree
. Their real parts provide a
collection of embedded knots in
.
Equip this real projective space
with a spin structure. Then it is possible to define a spinor orientation
on these knots. Indeed, considering a real subholomorphic line bundle of
maximal degree in the normal bundle of the curves, one first defines a
framing on these knots. From this framing, one can then build a loop in
the
-principal bundle of orthonormal frames of
. Then,
the spinor orientation of the real curve is the obstruction to lift this
loop as a loop of the Spin
-principal bundle given by the spin
structure.
Now the algebraic number of real curves, counted with respect to their
spinor orientation, turns out to be independent of the choice of the
configuration of points and this is my invariant.
Question: Is it possible to compute this invariant with the help of tropical algebraic geometry?
(contributed by Jean-Yves Welschinger)
Back to the
main index
for Amoebas and tropical geometry.