Reduction to integration over single Lagrangian submanifold
Here we will explain that integration over a family of Lagrangian submanifolds can be reduced to integration over a single Lagrangian submanifold in a larger phase space.
A family of Lagrangian submanifolds
defines a single
Lagrangian submanifold in
. We will
describe the construction of
in two steps: first locally in a vicinity of some point
of
, and then globally.
1 Local construction
Let us pick some fixed Lagrangian submanifold
from our family
:
Let us define a “first try” Lagrangial submanifold
as follows:
this is a direct product of two Lagrangian submanifolds,
and the zero section
Locally, in the vicinity of
, we can present
as a family of gauge fermions
(for each
, sufficiently close to
, the
is obtained from
as a deformation
corresponding to the gauge fermion
). Let us consider the following function:
We consider it as a gauge fermion deforming
to some new Lagrangian submanifold.
This new Lagrangian submanifold is what we need; we call it
.
2 Global definition of
As a first step, let us consider a submanifold
which is defined as follows:
This can be promoted to a subspace
quite trivially:
Finally, we will construct
as some section of the vector bundle
restricted to
. Which section?
The simplest guess would be the zero section. However
that would not be a Lagrangian submanifold, so zero section is a wrong guess.
There is, however, a natural nonzero section. It is constructed as follows:
where
computes for every tangent vector to
the value of its corresponding generating function on
.
This section defines our big Lagrangian submanifold:
3 Integration
There is a natural BV Hamiltonian
on
. It descirbes
the lift to
of the natural nilpotent vector field
on
.
We have:
|
In the case of Yang-Mills theory, the commonly accepted notations for fields are:
4 Proof that
is Lagrangian
First of all, we have to eliminate the ambiguity in the definition of
. We do this by lifting
to a subset of
. Now to every point
corresponds
, and:
The way we defined the moment map,is a function on
. But since we lifted
to
, it defines for us a function on
, which is the same as a section of
. On the other hand, from any section:
we can naturally construct a sectionof
in the following way: first construct
where
and
and then compose it with the natural embedding. This is what Eq. (50) does.
We have to prove that the restriction of the symplectic form of
to
is zero.
We notice that it is equal to the pullback under the natural projection
(see Eq. (48)) of the following 2-form on
:
Therefore we need to prove that
is zero.
Let
be a vector field on
. By our definition of the moment map,
it generates the vector field
on
.
Let
be a vector in
tangent to
. Then, using general formulas:
Therefore
vanishes on any pair of vectors
when one of them is tangent to
.
Since is a Lagrangian submanifold, this completes the proof that
is zero.