Skip to content
GitLab
Explore
Sign in
Primary navigation
Search or go to…
Project
M
Max Destabilizer Rank
Manage
Activity
Members
Labels
Plan
Issues
Issue boards
Milestones
Wiki
Code
Merge requests
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Snippets
Build
Pipelines
Jobs
Pipeline schedules
Artifacts
Deploy
Releases
Package Registry
Container Registry
Model registry
Operate
Environments
Terraform modules
Monitor
Incidents
Analyze
Value stream analytics
Contributor analytics
CI/CD analytics
Repository analytics
Model experiments
Help
Help
Support
GitLab documentation
Compare GitLab plans
Community forum
Contribute to GitLab
Provide feedback
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
luke naylor latex documents
research
Max Destabilizer Rank
Commits
1ea1ba65
Commit
1ea1ba65
authored
1 year ago
by
Luke Naylor
Browse files
Options
Downloads
Patches
Plain Diff
Overhaul expressions for bgmlv1 condition (\Delta(u,v) >= 0)
parent
d439619a
No related branches found
No related tags found
No related merge requests found
Pipeline
#28329
passed
1 year ago
Stage: test
Changes
1
Pipelines
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
main.tex
+117
-127
117 additions, 127 deletions
main.tex
with
117 additions
and
127 deletions
main.tex
+
117
−
127
View file @
1ea1ba65
...
...
@@ -10,6 +10,7 @@
\usepackage
{
sagetex
}
\usepackage
{
minted
}
\usepackage
{
subcaption
}
\usepackage
{
cancel
}
\usepackage
[]
{
breqn
}
\usepackage
[
backend=biber,
...
...
@@ -642,6 +643,7 @@ algorithm will be presented making use of the later theorems in this article,
with the goal of cutting down the run time.
\subsection*
{
Problem statement
}
\label
{
subsect:problem-statement
}
Fix a Chern character
$
v
$
with positive rank,
$
\Delta
(
v
)
\geq
0
$
,
and
$
\beta
_{
-
}
(
v
)
\in
\QQ
$
.
...
...
@@ -732,9 +734,11 @@ As opposed to only eliminating possible values of $\chern_0(E)$ for which all
corresponding
$
\chern
_
1
^{
\beta
}
(
E
)
$
fail one of the inequalities (which is what
was implicitly happening before).
First, let us fix a Chern character for
$
F
$
,
$
\chern
(
F
)
=
(
R,C
\ell
,D
\ell
^
2
)
$
, and consider the possible Chern characters
$
\chern
(
E
)
=
(
r,c
\ell
,d
\ell
^
2
)
$
of some semistabilizer
$
E
$
.
First, let us fix a Chern character for
$
F
$
, and some semistabilizer
$
E
$
:
\begin{align}
v
&
:=
\chern
(F) = (R,C
\ell
,D
\ell
^
2)
\\
u
&
:=
\chern
(E) = (r,c
\ell
,d
\ell
^
2)
\end{align}
\begin{sagesilent}
# Requires extra package:
...
...
@@ -748,8 +752,8 @@ u = Chern_Char(*var("r c d", domain="real"))
Δ = lambda v: v.Q
_
tilt()
\end{sagesilent}
Recall from eqn
\ref
{
eqn-tilt-cat-cond
}
that
$
\chern
_
1
^{
\beta
}$
has fixed
bounds in terms of
$
\chern
(
F
)
$
, and so we can write:
Recall from eqn
\ref
{
eqn-tilt-cat-cond
}
that
$
\chern
_
1
^{
\beta
}
(
u
)
$
has fixed
bounds in terms of
$
\chern
_
1
^{
\beta
}
(
v
)
$
, and so we can write:
\begin{sagesilent}
ts = stability.Tilt
...
...
@@ -766,8 +770,8 @@ c_in_terms_of_q = c_lower_bound + q
\begin{equation}
\label
{
eqn-cintermsofm
}
c=
\chern
_
1(
E
) =
\sage
{
c
_
in
_
terms
_
of
_
q
}
\qquad
0
\leq
q
\leq
\chern
_
1
^{
\beta
}
(
F
)
c=
\chern
_
1(
u
) =
\sage
{
c
_
in
_
terms
_
of
_
q
}
\qquad
0
\leq
q
:=
\chern
_
1
^{
\beta
}
(u)
\leq
\chern
_
1
^{
\beta
}
(
v
)
\end{equation}
Furthermore,
$
\chern
_
1
\in
\ZZ
$
so we only need to consider
...
...
@@ -778,128 +782,99 @@ and we shall be varying $\chern_0(E) = r$ to see when certain inequalities fail.
\subsection
{
Numerical Inequalities
}
\subsubsection
{
Size of pseudo-wall:
$
\chern
_
2
^
P
(
u
)
>
0
$
}
\label
{
subsect-d-bound-radiuscond
}
This condition refers to consequence 4 from
lemma
\ref
{
lem:pseudo
_
wall
_
numerical
_
tests
}
.
In the case where
$
P
$
was chosen to be the base of
$
\Theta
_
v
^
-
$
:
$
P
=(
\beta
_{
-
}
,
0
)
$
.
This condition amounts to:
\begin{align}
\label
{
eqn:radius-cond-betamin
}
\chern
_
2
^{
\beta
_{
-
}}
(u)
&
\geq
0
\\
d
&
\geq
\beta
_{
-
}
q +
\frac
{
1
}{
2
}
\beta
_{
-
}^
2r
\end{align}
\subsubsection
{
\texorpdfstring
{
$
\Delta
(
E
)
+
\Delta
(
G
)
\
l
eq
\Delta
(
F
)
$
$
\Delta
(
u,v
-
u
)
\
g
eq
0
$
}{
Δ(
E) + Δ(G) ≤ Δ(F)
Δ(
u,v-u) ≤ 0
}
}
\label
{
subsect-d-bound-bgmlv1
}
This condition expressed in terms of
$
R,C,D,r,c,d
$
looks as follows:
\begin{sagesilent}
# First Bogomolov-Gieseker form expression that must be non-negative:
bgmlv1 = Δ(v) - Δ(u) - Δ(v-u)
\end{sagesilent}
Writing the condition in terms of the twisted chern characters
for
$
u
$
and
$
v
$
at
$
\beta
$
(
$
(
r,
\chern
_
1
^{
\beta
}
(
u
)
,
\chern
_
2
^{
\beta
}
(
u
))
$
and
$
(
R
-
r,
\chern
_
1
^{
\beta
}
(
v
-
u
)
,
\chern
_
2
^{
\beta
}
(
v
-
u
))
$
) yields:
\begin{equation}
\sage
{
0 <= bgmlv1.expand()
}
\label
{
eqn:bgmlv1-pt1
}
(R-r)
\chern
_
2
^{
\beta
}
(u)
\leq
\chern
_
1
^{
\beta
}
(u)
\chern
_
1
^{
\beta
}
(v-u)
- r
\chern
_
2
^{
\beta
}
(v-u)
\end{equation}
\noindent
Expressing
$
c
$
in terms of
$
q
$
as defined in (eqn
\ref
{
eqn-cintermsofm
}
)
we get the following:
\begin{sagesilent}
bgmlv1
_
with
_
q = (
bgmlv1
.expand()
.subs(c == c
_
in
_
terms
_
of
_
q)
)
\end{sagesilent}
Which rearranges to (using additivity of
$
\chern
_
2
^{
\beta
}$
):
\begin{equation}
\sage
{
0 <= bgmlv1
_
with
_
q
}
\label
{
eqn:bgmlv1-pt2
}
(R-2r)
\chern
_
2
^{
\beta
}
(u)
\leq
\chern
_
1
^{
\beta
}
(u)
\chern
_
1
^{
\beta
}
(v-u)
- r
\chern
_
2
^{
\beta
}
(v)
\end{equation}
With
$
u
$
satisfying the condition given by equation
\ref
{
eqn-cintermsofm
}
,
we note that
$
\chern
_
1
^{
\beta
}
(
u
)
,
\chern
_
1
^{
\beta
}
(
v
-
u
)
\geq
0
$
.
\noindent
This can be rearranged to express a bound on
$
d
$
as follows:
\begin{sagesilent}
var("r
_
alt",domain="real") # r
_
alt = r - R/2 temporary substitution
bgmlv1
_
with
_
q
_
reparam = (bgmlv1
_
with
_
q.subs(r == r
_
alt + R/2)/r
_
alt).expand()
bgmlv1
_
d
_
ineq = (
((0 >= -bgmlv1
_
with
_
q
_
reparam)/4 + d) # Rearrange for d
.subs(r
_
alt == r - R/2) # Resubstitute r back in
.expand()
)
bgmlv1
_
d
_
lowerbound = bgmlv1
_
d
_
ineq.rhs() # Keep hold of lower bound for d
\end{sagesilent}
\begin{dmath}
\label
{
eqn-bgmlv1
_
d
_
lowerbound
}
\sage
{
bgmlv1
_
d
_
ineq
}
\end{dmath}
\begin{sagesilent}
# Separate out the terms of the lower bound for d
bgmlv1
_
d
_
lowerbound
_
without
_
hyp = bgmlv1
_
d
_
lowerbound.subs(1/(R-2*r) == 0)
In the special case with
$
P
=(
\beta
_{
-
}
,
0
)
$
,
we have
$
\chern
_
2
^{
\beta
_{
-
}}
(
v
)=
0
$
, and we can assume
equation
$
\chern
_
2
^{
\beta
_{
-
}}
(
u
)
>
0
$
(eqn
\ref
{
eqn:radius-cond-betamin
}
)
in the context of our problem.
bgmlv1
_
d
_
lowerbound
_
exp
_
term = (
bgmlv1
_
d
_
lowerbound
- bgmlv1
_
d
_
lowerbound
_
without
_
hyp
).expand()
Finally,
$
r>
0
$
as per the statement of the problem, so the right-hand-side
of equation
\ref
{
eqn:bgmlv1-pt1
}
is always greater than, or equal, to zero.
And so, when
$
P:
=(
\beta
_{
-
}
,
0
)
$
, this condition
$
\Delta
(
u,v
-
u
)
\geq
0
$
is
always satisfied when
$
2
r
\geq
R
$
, provided that the other conditions of the
problem statement (
\ref
{
subsect:problem-statement
}
) hold.
bgmlv1
_
d
_
lowerbound
_
const
_
term = bgmlv1
_
d
_
lowerbound
_
without
_
hyp.subs(r==0)
bgmlv1
_
d
_
lowerbound
_
linear
_
term = (
bgmlv1
_
d
_
lowerbound
_
without
_
hyp
- bgmlv1
_
d
_
lowerbound
_
const
_
term
).expand()
# Verify the simplified forms of the terms that will be mentioned in text
var("chbv",domain="real") # symbol to represent ch
_
1
^
\beta
(v)
assert bgmlv1
_
d
_
lowerbound
_
const
_
term == (
(
# Keep hold of this alternative expression:
bgmlv1
_
d
_
lowerbound
_
const
_
term
_
alt :=
(
chbv/2
+ beta*q
)
)
.subs(chbv == v.twist(beta).ch[2])
.expand()
)
assert bgmlv1
_
d
_
lowerbound
_
exp
_
term == (
(
# Keep hold of this alternative expression:
bgmlv1
_
d
_
lowerbound
_
exp
_
term
_
alt :=
(
- R*chbv/2
- R*beta*q
+ C*q
- q
^
2
)/(R-2*r)
)
.subs(chbv == v.twist(beta).ch[2])
.expand()
)
\end{sagesilent}
However, when
$
2
r<R
$
, this condition does add potentially independent condition
of the others:
\begin{equation}
\label
{
eqn:bgmlv1-pt3
}
\chern
_
2
^{
\beta
}
(u)
\leq
\frac
{
\chern
_
1
^{
\beta
}
(u)
\chern
_
1
^{
\beta
}
(v-u)
- r
\chern
_
2
^{
\beta
}
(v)
}
{
R-2r
}
,
\qquad
2r<R
\end{equation}
\noindent
Viewing equation
\ref
{
eqn-bgmlv1
_
d
_
lowerbound
}
as a lower bound for
$
d
$
given
as a function of
$
r
$
, the terms can be rewritten as follows.
The constant term in
$
r
$
is
$
\chern
^{
\beta
}_
2
(
F
)/
2
+
\beta
q
$
.
The linear term in
$
r
$
is
$
\sage
{
bgmlv
1
_
d
_
lowerbound
_
linear
_
term
}$
.
Finally, there is an hyperbolic term in
$
r
$
which tends to 0 as
$
r
\to
\infty
$
,
and can be written:
$
\frac
{
R
\chern
^{
\beta
}_
2
(
F
)/
2
+
R
\beta
q
-
Cq
+
q
^
2
}{
2
r
-
R
}$
.
In the case
$
\beta
=
\beta
_{
-
}$
(or
$
\beta
_{
+
}$
) we have
$
\chern
^{
\beta
}_
2
(
F
)
=
0
$
,
so some of these expressions simplify.
Expressed in terms of
$
d
$
and
$
q
$
:
\begin{equation}
\label
{
eqn:bgmlv1-pt4
}
d
\leq
\beta
_{
-
}
q
+
\frac
{
1
}{
2
}{
\beta
_{
-
}}^
2r
+
\frac
{
q(
\chern
_
1
^{
\beta
}
(v)-q)
- r
\chern
_
2
^{
\beta
}
(v)
}
{
R-2r
}
,
\qquad
2r<R
\end{equation}
\subsubsection
{
\texorpdfstring
{
...
...
@@ -1174,12 +1149,23 @@ vertical wall (TODO as discussed in ref).
% redefine \psi in sage expressions (placeholder for ch_1^\beta(F)
\def\psi
{
\chern
_
1
^{
\beta
}
(F)
}
\begin{align}
d
&
\geq
&
\sage
{
bgmlv1
_
d
_
lowerbound
_
linear
_
term
}
&
+
\sage
{
bgmlv1
_
d
_
lowerbound
_
const
_
term
_
alt.subs(chbv == 0)
}
+
&
\sage
{
bgmlv1
_
d
_
lowerbound
_
exp
_
term
_
alt.subs(chbv == 0)
}
,
&
\qquad\text
{
when
\:
}
r >
\frac
{
R
}{
2
}
\label
{
eqn:bgmlv1
_
d
_
bound
_
betamin
}
d
&
>
&
\frac
{
1
}{
2
}
\beta
^
2 r
&
+
\beta
q,
\phantom
{
+
}&
% to keep terms aligned
&
\qquad\text
{
when
\:
}
r > 0
\label
{
eqn:radiuscond
_
d
_
bound
_
betamin
}
\\
d
&
\leq
&
\frac
{
1
}{
2
}{
\beta
}^
2r
&
+
\beta
q
+
&
\frac
{
q(
\chern
_
1
^{
\beta
}
(v)-q)
}
{
R-2r
}
,
&
\qquad\text
{
when
\:
}
0 < r <
\frac
{
R
}{
2
}
\label
{
eqn:bgmlv1
_
d
_
bound
_
betamin
}
\\
d
&
\leq
&
\sage
{
bgmlv2
_
d
_
upperbound
_
linear
_
term
}
...
...
@@ -1246,12 +1232,15 @@ def plot_d_bound(
# Equations to plot imminently representing the bounds on d:
eq1 = (
bgmlv1
_
d
_
lowerbound
.subs(R == v
_
example.ch[0])
.subs(C == v
_
example.ch[1])
.subs(D == v
_
example.ch[2])
(
beta
^
2*r/2
+ beta*q
+ q*(chb1v - q)/(R-2*r)
)
.subs(chb1v == v
_
example.twist(beta
_
min(v
_
example)).ch[1])
.subs(beta = beta
_
min(v
_
example))
.subs(q == q
_
example)
.subs(R == v
_
example.ch[0])
)
eq2 = (
...
...
@@ -1284,28 +1273,28 @@ def plot_d_bound(
(r,v
_
example.ch[0],xmax),
color='green',
linestyle = "dashed",
legend
_
label=r"upper bound:
$
\Delta
(
G
)
\geq
0
$
",
legend
_
label=r"upper bound:
$
\Delta
(
v
-
u
)
\geq
0
$
",
)
+ plot(
eq2,
(r,0,xmax),
color='blue',
linestyle = "dashed",
legend
_
label=r"upper bound:
$
\Delta
(
E
)
\geq
0
$
"
legend
_
label=r"upper bound:
$
\Delta
(
u
)
\geq
0
$
"
)
+ plot(
eq4,
(r,0,xmax),
color='orange',
linestyle = "dotted",
legend
_
label=r"lower bound:
$
\mathrm
{
ch
}_
2
^{
\beta
_{
-
}}
(
E
)
>
0
$
"
legend
_
label=r"lower bound:
$
\mathrm
{
ch
}_
2
^{
\beta
_{
-
}}
(
u
)
>
0
$
"
)
+ plot(
eq1,
(r,v
_
example.ch[0]/2
,xmax
),
(r,
0,
v
_
example.ch[0]/2),
color='red',
linestyle = "d
ott
ed",
legend
_
label=r"
low
er bound:
$
\Delta
(
E
)
+
\Delta
(
G
)
\
l
eq
\Delta
(
F
)
$
"
linestyle = "d
ash
ed",
legend
_
label=r"
upp
er bound:
$
\Delta
(
u,v
)
\
g
eq
0
$
"
)
)
example
_
bounds
_
on
_
d
_
plot.ymin(ymin)
...
...
@@ -1328,7 +1317,7 @@ def plot_d_bound(
\hfill
\begin{subfigure}
{
.45
\textwidth
}
\centering
\sageplot
[width=\linewidth]
{
plot
_
d
_
bound(v
_
example, 4)
}
\sageplot
[width=\linewidth]
{
plot
_
d
_
bound(v
_
example, 4
, ymin=-3
)
}
\caption
{$
q
=
\chern
^{
\beta
}
(
F
)
$
(all bounds other than blue coincide on line)
}
\label
{
fig:d
_
bounds
_
xmpl
_
max
_
q
}
\end{subfigure}
...
...
@@ -1376,7 +1365,7 @@ Some of the details around the associated numerics are explored next.
\centering
\sageplot
[
width=
\linewidth
]
{
plot
_
d
_
bound(v
_
example, 2, ymax=6, ymin=-
0.5
, aspect
_
ratio=1)
}
]
{
plot
_
d
_
bound(v
_
example, 2, ymax=6, ymin=-
2
, aspect
_
ratio=1)
}
\caption
{
Bounds on
$
d:
=
\chern
_
2
(
E
)
$
in terms of
$
r:
=
\chern
_
0
(
E
)
$
for a fixed
value
$
\chern
_
1
^{
\beta
}
(
F
)/
2
$
of
$
q:
=
\chern
_
1
^{
\beta
}
(
E
)
$
.
...
...
@@ -1507,6 +1496,7 @@ considering equations
\begin{sagesilent}
var("epsilon")
var("chbv") # symbol to represent
\chern
_
1
^{
\beta
}
(v)
# Tightness conditions:
...
...
This diff is collapsed.
Click to expand it.
Preview
0%
Loading
Try again
or
attach a new file
.
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Save comment
Cancel
Please
register
or
sign in
to comment