The diagrammatic beauty of Rep(U
q(sl
n)): Part I
Daniel Tubbenhauer
The uncategorified story
March 2014
= =
Daniel Tubbenhauer March 2014 1 / 38
1 The diagrammatic calculus sl2-webs
Thesl2-flow lines Algebra is rigid
2 The representation theory ofUq(sl2) The algebra
Connection to the diagrammatic calculus Invariant tensors
3 Connection to thesln-link-polynomials The Jones polynomial
Reshetikhin-Turaev: Jones is an intertwiner RT polynomials usingsln-webs
Daniel Tubbenhauer March 2014 2 / 38
An old story: Rumer, Teller and Weyl (1932)
Daniel Tubbenhauer sl2-webs March 2014 3 / 38
The sl
2-webs
Definition(Rumar, Teller, Weyl 1932)
Fix two numbersb,t∈Nwithb+t= 2ℓ. Asl2-webw withbbottom points and t top points is an embedding (non-intersecting!) of a finite number of lines and circles in a rectangle withb fixed points at the bottom andt at the top such that the two boundary points of the lines are some of the fixed points. Theset of all sl2-websw betweenbbottom points and t top points in denoted by ˜W2(b,t).
Example(b = 3 and t = 5)
Daniel Tubbenhauer sl2-webs March 2014 4 / 38
The sl
2-web space
Definition
Fix two numbersb,t∈Nwithb+t= 2ℓ. Thesl2-web spaceW2(b,t)is the free Q(q)-vector space¯ generatedby elements of ˜W2(b,t) modulo
Thecircle removal
= [2] = q+q−1 Theisotopy relations
= =
Note thatW2(b,t) is afinite dimensional ¯Q(q)-vector space!
Daniel Tubbenhauer sl2-webs March 2014 5 / 38
The sl
2-web category
Definition(Kuperberg 1997)
Thesl2-web category or web spiderSp(Uq(sl2))is the monoidal, ¯Q(q)-linear 1-category consisting of
Theobjectsare the natural numbersN={0,1,2, . . .}.
The1-cellsw: b→t are the elements ofW2(b,t).
The ¯Q(q)-linear composition isstacking.
The monoidal structure⊗is given byjuxtaposition, i.e.b⊗b′ =b+b′ and
⊗ =
As generatorssufficesthe identities, shifts, cups and caps
Daniel Tubbenhauer sl2-webs March 2014 6 / 38
The sl
2-web category - examples
Example
◦ =
and
◦ = = [2]·id(3,3)
Daniel Tubbenhauer sl2-webs March 2014 7 / 38
An extra information
Definition
Given asl2-webw ∈W2(b,t). Asl2-flow linef forw, denoted bywf, is a choice of orientation for all lines and circles ofw.
If one ignores internal circles
and
then such a flow line iscompletelydetermine by its boundary. There it induces a state stringfor the bottom~Sb= (±, . . . ,±) and top~St = (±, . . . ,±) with a plus for outgoing flow lines and a minus for incoming.
Example
+ + −
− + − − +
Daniel Tubbenhauer Thesl2-flow lines March 2014 8 / 38
The weight of a flow
Definition
Flows on the generators of thesl2-web categorySp(Uq(sl2)) are assigned a certainweightwtby the local rules
wt( ) = 0 wt( ) =−1 wt( ) = 0 wt( ) = +1
and always zero on identities and shifts. Theweightof anysl2-web with flow is the sum over the local weights.
Note that the weight isisotopy invariant, thus,well-definedforsl2-webs without internal circles.
Example
wt= 0
wt=+1
wt=−1
= wt= 0
=
wt= 0
wt=0
wt=0
Daniel Tubbenhauer Thesl2-flow lines March 2014 9 / 38
Rigidity of sl
2-webs
A seemingly very small point turned out to bea crucial stepif we want to consider biggern: Topology is continuous and Algebra is rigid.
Definition, second try - rigid version
Thesl2-web category or web spiderSp(Uq(sl2))is the monoidal, ¯Q(q)-linear 1-category consisting of
Theobjectsare ordered compositions~k of 2ℓ∈Nwith only 0,1,2 as entries.
The1-cellsw:~k →~k′arelabelled ladders(we use the convention and do not draw edges labelled 0 and use a dotted line for those labelled 2) generated by juxtaposition and vertical composition of (plus relations and restas before)
k1 k2 k3 k4
k1 k2 k3 k4
k1 k2 k3 k4
k1 k2
k1±k k2∓k k
k=0,1,2
2 0
1 1
0 2
=
2 0
0 2
What is theupshot? “Easy” to generalize to sln: Take labels 0,1, . . . ,n−1,nand
“directly” connected to the algebra (which I explain in a second!).
Daniel Tubbenhauer Algebra is rigid March 2014 10 / 38
A rigid example
There is a small number of different ladders, namely theleft and right shifts
and (rigid)cups and caps
These suffice to generate allsl2-webs, e.g.
Daniel Tubbenhauer Algebra is rigid March 2014 11 / 38
The quantum algebra U
q(sl
d)
Definition
Ford∈N>1thequantum special linear algebraUq(sld) is the associative, unital Q(q)-algebra generated by¯ Ki±1 andEi andFi, fori= 1, . . . ,d−1, subject to some relations (that we do not need today).
Definition(Beilinson-Lusztig-MacPherson)
For each~k ∈Zd−1 adjoin anidempotent1~k (think: projection to the~k-weight space!) toUq(sld) and add some relations, e.g.
1~k1~k′ =δ~k,~k′1~k and K±i1~k =q±~ki1~k (noK′s anymore!).
Theidempotented quantum special linear algebrais defined by U˙q(sld) = M
~k,~k′∈Zd−1
1~kUq(sld)1~k′.
Daniel Tubbenhauer The algebra March 2014 12 / 38
The quantum algebra U
q(sl
d) is a Hopf algebra
It is worth noting thatUq(sld) is a Hopf algebrawithcoproduct∆ given by
∆(Ei) =Ei⊗Ki+ 1⊗Ei, ∆(Fi) =Fi⊗1 +Ki−1⊗Fi and ∆(Ki) =Ki⊗Ki. Theantipode S and the counitεare given by
S(Ei) =−EiKi−1, S(Fi) =−KiFi, S(Ki) =Ki−1, ε(Ei) =ε(Fi) = 0, ε(Ki) = 1.
The Hopf algebra structure allows toextendactions to tensor products of representations, to duals of representations and there is a trivial representation.
Example: U
q(sl
2)-vector representation
Consider ¯Q2with basisx−1= (0,1),x+1= (1,0). These are called the weights−1 and +1 andK acts on them byq∓1. The vector representation ofUq(sl2) is:
Think:K = 1 0
0 −1
(0,1)
E
##
(1,0)
F
cc
Think:
E= 0 1
0 0
F= 0 0
1 0
Daniel Tubbenhauer The algebra March 2014 13 / 38
The category Rep(U
q(sl
2))
Definition
Therepresentation categoryRep(Uq(sl2))is the monoidal, ¯Q(q)-linear 1-category consisting of:
Theobjectsare finite tensor products of theUq(sl2)-representations ΛkQ¯2. Denote them by~k = (k1, . . . ,km) withki ∈ {0,1,2}.
The1-cellsw:~k→~k′ areUq(sl2)-intertwiners.
Composition of 1-cells iscomposition of intertwinersand⊗is theordered tensor product.
It is worth noting that Λ0Q¯2= ¯Qis the trivialUq(sl2)-representation, Λ2Q¯2∼= ¯Q its dual and Λ1Q¯2= ¯Q2 is the (self-dual)Uq(sl2)-vector representation.
Example
The 1-cells ofMor(~k, ~k′)“are”(using the Hopf algebra structure!) the invariant tensorsInvUq(sl2)(~k∗⊗~k′) with~k∗= (2−k1, . . . ,2−km).
Daniel Tubbenhauer The algebra March 2014 14 / 38
We have some control over the intertwiner
GivenV =N
iΛkiQ¯2denote thetensor basisofV (recall x−1= (0,1) and x+1= (1,0), setx∅=x{−1,+1}= 1) by{xS |S = (S1, . . . ,Sm),Si ⊂ {−1,+1}}.
Theorem(Kuperberg 1997, n > 3: Cautis-Kamnitzer-Morrison 2012)
Define two ¯Q-linear maps calledsplit and mergeby Msa,b: Λa+bQ¯2→ΛaQ¯2⊗ΛbQ¯2, Msa,b(xS) = X
T⊂S
(−q)ℓ(S,T)xT⊗xS−T
Mma,b: ΛaQ¯2⊗ΛbQ¯2→Λa+bQ¯2, Mma,b(xS⊗xT) =
((−q)−ℓ(T,S)xS∪T, S∩T =∅,
0, else.
for suitablea,b∈ {0,1,2}andℓ(S,T)∈ {−1,0,+1}. These are Uq(sl2)-intertwinerandgenerateRep(Uq(sl2)).
E.g.:Mm1,1(x−1⊗x+1) = (−q)0,Mm1,1(x+1⊗x−1) = (−q)−1,Mm1,1(x±1⊗x±1) = 0.
Daniel Tubbenhauer The algebra March 2014 15 / 38
Intertwiner are pictures
Theorem(Kuperberg 1997, n > 3: Cautis-Kamnitzer-Morrison 2012)
The 1-categoriesRep(Uq(sl2)) andSp(Uq(sl2)) areequivalent. To be more precise, the equivalence Γ : Rep(Uq(sl2))→Sp(Uq(sl2)) is given by:
One objects: SendN
~kΛkiQ¯2to~k.
One 1-cells: We only need to consider the generators splitMsa,b and merge Mma,b. Send them to
Msa,b7→
a+b 0
a b
b and Mma,b7→
a b
0 a+b
a
Check that it iswell-defined!
I amlyinga little bit: One has to be a little more careful with objects and duals, but weignorethis for today.
Daniel Tubbenhauer Connection to the diagrammatic calculus March 2014 16 / 38
Intertwiner are pictures: Some examples
Exempli gratia
What about the“left-plus-ladders”? They are acomposite!
k1 k2
k1+k k2−k k
k=0,1,2
=
k1 k2
k1+k k2−k k
k=0,1,2
e.g.
1 2
2 1
1 =
1 2
2 1
1
Generatethem by composition of merge and split!
Mmk1,k
Msk,k2−k = (Mmk1,k⊗id(k2−k,k2−k))◦(id(k1,k1)⊗Msk,k2−k) e.g.
= (Mm1,1⊗id(1,1))◦(id(1,1)⊗Ms1,1)
Daniel Tubbenhauer Connection to the diagrammatic calculus March 2014 17 / 38
How to prove it? Quantum skew Howe duality!
Theorem
There is anU˙q(sld)-actionon Sp(Uq(sl2))d (objects of lengthd)!
1~k 7→
... ...
k1 ki−1 ki ki+1 ki+2 kd
k1 ki−1 ki ki+1 ki+2 kd
Ei1~k, Fi1~k 7→
... ...
k1 ki−1 ki ki+1 ki+2 kd
k1 ki−1 ki±1 ki+1∓1 ki+2 kd
Thus,Sp(Uq(sl2))d is aU˙q(sld)-moduleandnot justaUq(sl2)-module.
Even better: Since, weonlyneed “left-minus-ladders”, akaF’s, it can be realized as a ˙Uq(sld)-module of a certain highest weight: We can useU˙q(sld)-highest weight theory to prove statements aboutUq(sl2)-intertwiner!
Daniel Tubbenhauer Connection to the diagrammatic calculus March 2014 18 / 38
The invariant tensors suffice - a picture
The Hopf-structure says:Mor(~k, ~k′)∼=InvUq(sl2)(~k∗⊗~k′). The picture says:
~k′
~k
∼=
~k′
~k∗
Remaining question: How toidentifythe invariant tensors?
Daniel Tubbenhauer Invariant tensors March 2014 19 / 38
This is not trivial...
First question: What do wemean by “identify” the invariant tensors?
Recall: InvUq(sl2)(Λk1Q¯2⊗ · · · ⊗ΛkdQ¯2)⊂Λk1Q¯2⊗ · · · ⊗ΛkdQ¯2, and Λk1Q¯2⊗ · · · ⊗ΛkdQ¯2has aeasyto write down, buthorribleto work with basis: The elementary tensorsx~S!
Thus, “identify”v ∈InvUq(sl2)(Λk1Q¯2⊗ · · · ⊗ΛkdQ¯2) is writingv in terms ofx~S. Second question:Howto do it? Recall that the action on tensors is given by
∆(Ei) =Ei⊗Ki+ 1⊗Ei, ∆(Fi) =Fi⊗1 +Ki−1⊗Fi and ∆(Ki) =Ki⊗Ki.
Daniel Tubbenhauer Invariant tensors March 2014 20 / 38
Exempli gratia
Example: ¯ Q
2⊗ Q ¯
2Recall that ¯Q2has basisx+1= (1,0) andx−1= (0,1). Thus, ¯Q2⊗Q¯2has basis {x+1+1 =x+1⊗x+1,x+1−1=x+1⊗x−1,x−1+1=x−1⊗x+1,x−1−1=x−1⊗x−1}.
Test calculation:
F·x+1−1=F·x+1⊗x−1+K−1·x+1⊗F·x−1
=x−1⊗x−1
F·x−1+1=F·x−1⊗x+1+K−1·x−1⊗F·x+1
=q+1x−1⊗x−1
Claim:x+1,−1−q−1x−1+1is invariant and spansInvUq(sl2)( ¯Q2⊗Q¯2).
How to do this ingeneral?
Daniel Tubbenhauer Invariant tensors March 2014 21 / 38
The diagrammatic calculus helps
Theorem(Khovanov-Kuperberg 1997)
The decomposition of asl2-webw ∈Mor(∅, ~k) in terms of the elementary tensors x~S is encoded by the flow linesf onw in the following way:
Each flowf induces a state string ~Sf = (±, . . . ,±) at the boundary and has a weightwt(wf).
Then the coefficient forx~S
f is (−q)wt(wf). Thus,w =P
f(−q)wt(wf)x~S
f.
(Onlyn= 2!) A basis Arc ofMor(∅, ~k) is given by all arc diagrams.
Example: ¯ Q
2⊗ Q ¯
2again
In this case there is exactly one arcuand it has the two flows
wt( ) = 0, ~S = (+1,−1) and wt( ) =−1, ~S= (−1,+1) Conclusion:u=x+1⊗x−1−q−1x−1⊗x+1.
Daniel Tubbenhauer Invariant tensors March 2014 22 / 38
The negative exponent property
Example
= +
+ +
=x+1⊗x+1⊗x−1⊗x−1−q−1x+1⊗x−1⊗x+1⊗x−1
−q−1x−1⊗x+1⊗x−1⊗x+1+q−2x−1⊗x−1⊗x+1⊗x+1
Observation: Oneleadingterm plus a rest with coefficients inq−1Z[q−1]! This is called thenegative exponent property. In fact, the arc basis is thedual canonical basisin the sense of Lusztig.
Daniel Tubbenhauer Invariant tensors March 2014 23 / 38
n > 2? Use q -skew Howe!
Roughly:
Expressasln-web as a string ofF’s acting on a highest weight vectorvnℓ. The action of theF’s isgiven bythe splits and merges. Read of the resulting vector inductively.
There is also apurely combinatorialway to do this!
3 3 3 0 0 0
F3(2)
3 3 1 2 0 0
F4(2)
3 3 1 0 2 0
F5(2)
3 3 1 0 0 2
F2
3 2 2 0 0 2
F3
3 2 1 1 0 2
F4
3 2 1 0 1 2
F1
2 3 1 0 1 2
F2
2 2 2 0 1 2
F3(2)
2 2 0 2 1 2
F2
2 1 1 2 1 2
F1
1 2 1 2 1 2
Daniel Tubbenhauer Invariant tensors March 2014 24 / 38
Dual canonical sl
n-webs? Quantum skew Howe duality!
Counterexample
= + + Rest
=x+1+12 ⊗x+1+11 ⊗x002 ⊗x001 ⊗x−1−12 ⊗x−1−11
+x+1−12 ⊗x+1−11 ⊗x+1−12 ⊗x+1−11 ⊗x+1−12 ⊗x+1−11 + Rest Here Rest has coefficients inq−1Z[q−1].
Note that “most”n>2-webs do not have this property and this makes livevery complicated! But usingq-skew Howe duality one can obtain aniff-conditionfor a web to be dual canonical plus analgorithmto compute the dual canonical basis.
Daniel Tubbenhauer Invariant tensors March 2014 25 / 38
The famous Jones polynomial
LetLD be a diagram of an oriented link. Set [2] =q+q−1 and n+= number of crossings n−= number of crossings
Definition/Theorem(Jones 1984, Kauffman 1987)
Thebracket polynomialof the diagramLD (without orientations) is a polynomial hLDi ∈Z[q,q−1] given by the following rules.
h∅i= 1 (normalization).
h i=h i −qh i (recursion step 1).
h ∐LDi= [2]· hLDi(recursion step 2).
[2]J(LD) = (−1)n−qn+−2n−hLDi(Re-normalization).
The polynomialJ(·)∈Z[q,q−1] is an invariantof oriented links.
Daniel Tubbenhauer The Jones polynomial March 2014 26 / 38
Exempli gratia
* +
:
■■
■■
▲▲
▲▲
▲ rr rr r
✉✉
✉✉
[2]2 −2q·[2] +q2·[2]2
Thus,J(Hopf) =q5+q, i.e the Hopf link isnot trivial!
Daniel Tubbenhauer The Jones polynomial March 2014 27 / 38
“The Jones revolution”
Definition/Theorem(HOMFLY 1985, PT 1987)
Define a polynomialPn(LD)∈Z[q,q−1] uniquely determined by the property Pn() = 1 and the so-calledslnskein relations
q2n·Pn( )−q−2n·Pn( ) = (q+q−1)·Pn( ). Thesln-HOMFLY-PT polynomialis alink invariantandP2(LD) =J(LD).
Shortlyafter Jones several authors independently found new knot polynomials.
One example is the HOMFLY-PT polynomial. Moreover, researches discovered connectionsto different parts of mathematics and physics. Before the “Jones revolution” there was alackof knot polynomials and after there wheretoo many.
The questions shifted to:
“Why do they exist? How can we order them?”
Daniel Tubbenhauer The Jones polynomial March 2014 28 / 38
A tangle is an intertwiner
Letgbeanyclassical Lie algebra. Denote byλi, µj theUq(g)-representation of highest weightVλi,Vµj. LetTD be a diagram of a,λi, µj-colored, oriented tangle.
λ1, λ2, λ3 λ4 λ5
µ1 µ2 µ3 µ4 µ5
Vλ1⊗Vλ2⊗Vλ3⊗Vλ4⊗Vλ5 Vµ1⊗Vµ2⊗Vµ3⊗Vµ4⊗Vµ5
f(TD)
Daniel Tubbenhauer Reshetikhin-Turaev: Jones is an intertwiner March 2014 29 / 38
Representation theory does the trick!
Definition(Reshetikhin-Turaev 1990)
Given the set-up from before we define a certainUq(g)-intertwiner f(TD) :Vλ1⊗ · · · ⊗Vλk →Vλk+1⊗ · · · ⊗Vλl.
Theorem(Reshetikhin-Turaev 1990)
TheUq(g)-intertwinerf(TD) is aninvariantofTD.
Corollary(Reshetikhin-Turaev 1990)
In the case of colored, orientedlinksLD we have
f(LD) : ¯Q(q)→Q(q),¯ 17→PRT(LD)∈Z[q,q−1],
that is each configuration as above gives apolynomial invariant of oriented links!
Daniel Tubbenhauer Reshetikhin-Turaev: Jones is an intertwiner March 2014 30 / 38
This is powerful!
Example
We have the followinglistof examples!
Letg=sl2. If we restrictto theUq(sl2)-vector representation ¯Q2, then the Reshetikhin-Turaev polynomialPRT(·) is the Jones orsl2-polynomial.
Letg=sl2. If we allowanykind of coloring withUq(sl2)-representations, thenPRT(·) is the so-calledcoloredJones polynomial.
Letg=sln. If werestrictto theUq(sln)-vector representation ¯Qn, then the Reshetikhin-Turaev polynomialPRT(·) is thesln-polynomial.
But the Reshetikhin-Turaev polynomial is much moregeneralizethan all of them and“explains”them using one concept.
Moral: A lot of link polynomials arespecial instancesofsymmetries of the quantum groupsUq(g)!
Question: Can we do this more explicit forg=sln?
Daniel Tubbenhauer Reshetikhin-Turaev: Jones is an intertwiner March 2014 31 / 38
“Straightening” again
Consider a diagram of an oriented tangle. Its components can becoloredwith colorsk ∈ {0, . . . ,n}. These colorscorrespond to the fundamental
Uq(sln)-representations ΛkQ¯n.Straightening it into a Morse position.
2
1
1 2
4 5
7→
4 4 4 0 2 2 2 2 5 5 5 5
1 1 4 n 2 1 1 5 2 2 n 2
5 4 1 1 1 2 5 1 1 2 0 2
4 5 5 5 5 5 2 2 2 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1
Daniel Tubbenhauer RT polynomials usingsln-webs March 2014 32 / 38
Tangles to sl
n-webs
We can define as before the categorySp(Uq(sln)) consisting of 1-cells as
k1 k2 k3 k4
k1 k2 k3 k4
k1 k2 k3 k4
k1 k2
k1+k k k2−k
k=0,1,2,...,n
k1 k2
k1−k k2+k k
k=0,1,2,...,n
Letb≤a. Define anUq(sln)-intertwinerΛaQ¯n⊗ΛbQ¯n→ΛbQ¯n⊗ΛaQ¯nas follows.
b a
=
b
X
k=0
(−1)k+(a+1)bq−b+k
a b
a+k−b
k
a+k b−k
b a
,
“Morally” (up to some signs, shifts, re-orientations) the same fora<b and .
Daniel Tubbenhauer RT polynomials usingsln-webs March 2014 33 / 38
Exempli gratia: Hopf link for sl
22 1 1 1 1 1 1 1 1 0 0
0 1 2 2 1 1 1 0 0 1 0
2 2 1 0 1 1 1 1 1 2 2 1 2
0 0 0 1 1 1 1 1 1 1 2 2 2
7→
2 1 1 1 1 2 1 1 1 1 1 0 0
0 1 2 2 1 0 1 1 1 0 0 1 0
2 2 1 0 1 1 1 1 1 2 2 1 2
0 0 0 1 1 1 1 1 1 1 2 2 2
f10(Hopf) : Λ2Q¯2⊗Q¯ ⊗Λ2Q¯2⊗Q¯ →Q¯ ⊗Q¯ ⊗Λ2Q¯2⊗Λ2Q¯2is anintertwinerin Sp(Uq(sln)). In the end we get thesame polynomialas before (up to a shift).
Conclusion: Worksfine forn= 2. What aboutn>2?
Daniel Tubbenhauer RT polynomials usingsln-webs March 2014 34 / 38
Quantum skew Howe duality helps
↑ U˙q(sld)−action
↓
Uq(sln)−web →
←2
1 1 1 1 2 1 1 1 1 1 0 0
0 1 2 2 1 0 1 1 1 0 0 1 0
2 2 1 0 1 1 1 1 1 2 2 1 2
0 0 0 1 1 1 1 1 1 1 2 2 2
F1
F2
F3
F2
E1
F1
1 1
F2
F3
F1
F2
Recall that we have anU˙q(sld)-actiononSp(Uq(sln))d. In the example above f10(Hopf) =F2F1F3F2F1E1F2F3F2F1F2(2)v2200.
Daniel Tubbenhauer RT polynomials usingsln-webs March 2014 35 / 38
The lower part ˙ U
−q(sl
d) suffices!
↑
U˙−q(sld)−action
↓
Uq(sln)−web →
2 ←
1 1 1 1 0 0 0 0 0 0 0 0
0 1 2 2 1 2 1 1 0 0 0 0 0
0 0 0 0 0 0 1 0 1 1 1 0 0
0 0 0 0 0 0 0 1 1 0 0 1 2
2 2 1 0 1 1 1 1 1 2 1 1 0
0 0 0 1 1 1 1 1 1 1 2 2 2
F1
F2 F3 F4
F5
F2 F3 F4
F1
F2
F3
F2
F4
F5
F3
F4
A crucial observation: We needonlyF’s!
f10(Hopf) =F4(2)F4F3F5F4F2F3F2F1F4F3F2F5F4F3F2F1F4(2)F3(2)F2(2)v220000.
Daniel Tubbenhauer RT polynomials usingsln-webs March 2014 36 / 38
The sl
n-polynomials using sl
d-symmetries
Let ussummarizethe connection between (colored)sln-polynomials and the U˙q(sld)-Uq(sln)-skew Howe duality.
Reshetikhin-Turaev: Thesln-polynomialsPn(·)areUq(sln)-intertwiner.
Uq(sln)-intertwinerarevectors in hom’s between ˙Uq(sld)-weight spaces.
OnlyF’s: The space of invariant Uq(sln)-tensors is aU˙q(sld)-representation of somehighest weightvh and ˙U−q(sld)suffices.
Conclusion: The (colored)sln-polynomialsPn(·) are instances of highest U˙q(sld)-weight representation theory!
IfLD is a link diagram, thenPn(LD) is obtained byjumping viaF’sfrom a highest ˙Uq(sld)-weightvhto a lowest ˙Uq(sld)-weightvl!
Daniel Tubbenhauer RT polynomials usingsln-webs March 2014 37 / 38
There is stillmuchto do...
Daniel Tubbenhauer RT polynomials usingsln-webs March 2014 38 / 38
Thanks for your attention!
Daniel Tubbenhauer RT polynomials usingsln-webs March 2014 38 / 38