Quantum K-Theory of Grassmannians and Non-Abelian Localization

In the example of complex Grassmannians, we demonstrate various techniques available for computing genus-0 K-theoretic GW-invariants of flag manifolds and more general quiver varieties. In particular, we address explicit reconstruction of all such invariants using finite-difference operators, the ro...

Full description

Saved in:

Bibliographic Details
Published in:	Symmetry, Integrability and Geometry: Methods and Applications
Date:	2021
Main Authors:	Givental, Alexander, Yan, Xiaohan
Format:	Article
Language:	English
Published:	Інститут математики НАН України 2021
Online Access:	https://nasplib.isofts.kiev.ua/handle/123456789/211170
Tags:	Add Tag No Tags, Be the first to tag this record!
Journal Title:	Digital Library of Periodicals of National Academy of Sciences of Ukraine
Cite this:	Quantum K-Theory of Grassmannians and Non-Abelian Localization. Alexander Givental and Xiaohan Yan. SIGMA 17 (2021), 018, 24 pages

Institution

Digital Library of Periodicals of National Academy of Sciences of Ukraine

_version_	1859728478759288832
author	Givental, Alexander Yan, Xiaohan
author_facet	Givental, Alexander Yan, Xiaohan
citation_txt	Quantum K-Theory of Grassmannians and Non-Abelian Localization. Alexander Givental and Xiaohan Yan. SIGMA 17 (2021), 018, 24 pages
collection	DSpace DC
container_title	Symmetry, Integrability and Geometry: Methods and Applications
description	In the example of complex Grassmannians, we demonstrate various techniques available for computing genus-0 K-theoretic GW-invariants of flag manifolds and more general quiver varieties. In particular, we address explicit reconstruction of all such invariants using finite-difference operators, the role of the -hypergeometric series arising in the context of quasimap compactifications of spaces of rational curves in such varieties, the theory of twisted GW-invariants, including level structures, as well as the Jackson-type integrals playing the role of equivariant K-theoretic mirrors.
first_indexed	2026-03-15T11:48:05Z
format	Article
fulltext	Symmetry, Integrability and Geometry: Methods and Applications SIGMA 17 (2021), 018, 24 pages Quantum K-Theory of Grassmannians and Non-Abelian Localization Alexander GIVENTAL and Xiaohan YAN Department of Mathematics, University of California at Berkeley, Berkeley, CA 94720, USA E-mail: givental@math.berkeley.edu, xiaohan yan@berkeley.edu Received August 25, 2020, in final form February 02, 2021; Published online February 26, 2021 https://doi.org/10.3842/SIGMA.2021.018 Abstract. In the example of complex grassmannians, we demonstrate various techniques available for computing genus-0 K-theoretic GW-invariants of flag manifolds and more gene- ral quiver varieties. In particular, we address explicit reconstruction of all such invariants using finite-difference operators, the role of the q-hypergeometric series arising in the context of quasimap compactifications of spaces of rational curves in such varieties, the theory of twisted GW-invariants including level structures, as well as the Jackson-type integrals playing the role of equivariant K-theoretic mirrors. Key words: Gromov–Witten invariants; K-theory; grassmannians; non-abelian localization 2020 Mathematics Subject Classification: 14N35 To Vitaly Tarasov and Alexander Varchenko, on their anniversaries 1 Introduction Just as quantum cohomology theory deals with intersection numbers between interesting cyc- les in moduli spaces of stable maps of holomorphic curves in a given target (say, a Kähler manifold), quantum K-theory studies sheaf cohomology (e.g., in the form of holomorphic Euler characteristics) of interesting vector bundles over these moduli spaces. The beginnings of the subject can be traced back to the 20-year-old note [9] by the first-named author, the foundational work by Y.-P. Lee [25], and their joint paper on complete flag manifolds, q-Toda lattices and quantum groups [20]. In recent years, however, the interest to quantum K-theory expanded due to more discoveries of its diverse relations with representation theory, integrable systems, and q-hypergeometric functions. Apparently the interest was initiated by the 2012 preprint [28] by D. Maulik and A. Okounkov, who connected equivariant quantum cohomology of quiver varieties with R-matrices. In 2014, this led R. Rimányi, V. Tarasov and A. Varchenko [31] to a conjectural description of the quantum K-ring of the cotangent bundle of a partial flag variety. In even more recent literature motivated by representation theory (see, e.g., [24, 29, 30]), certain q-hypergeometric series, interesting from the point of view of the theory of integrable systems, appeared as generating functions for K-theoretic Gromov–Witten (GW) invariants of symplectic quiver varieties. In this literature, K-theoretic computations are based, however, on the quasimap (rather than stable map) compactifications [4] of spaces of rational curves in the GIT quotients of linear spaces. Based on the experience with mirror symmetry and quantum K-theory of toric varieties [14] one anticipates the q-hypergeometric generating functions arising from quasimap spaces to never- theless represent the “genuine” (i.e., based on stable map compactifications) K-theoretic GW This paper is a contribution to the Special Issue on Representation Theory and Integrable Systems in honor of Vitaly Tarasov on the 60th birthday and Alexander Varchenko on the 70th birthday. The full collection is available at https://www.emis.de/journals/SIGMA/Tarasov-Varchenko.html mailto:givental@math.berkeley.edu mailto:xiaohan_yan@berkeley.edu https://doi.org/10.3842/SIGMA.2021.018 https://www.emis.de/journals/SIGMA/Tarasov-Varchenko.html 2 A. Givental and X. Yan invariants, yet such invariants of a different kind, or more complicated ones than naively expec- ted. In any case, this brings up the question of comparison (first attempted by H. Liu [26]) of the two approaches. In this paper, we examine in substantial detail the genus-0 quantum K-theory of grass- mannians Grn,N (C). The grassmannians can be described as the GIT quotients Hom ( Cn,CN ) // GLn(C), and are perhaps the simplest among homogeneous spaces or, more generally, quiver varieties outside the toric class. Most of our methods carry over to other quiver varieties (and all – to any partial flag manifolds), but we prefer to illustrate the available techniques by way of simplest representative examples, trading generality for simplicity of notation. In Sections 2 and 3 we show how the technique of fixed point localization in moduli spaces of stable maps can be used in order to compute the so-called “small J-function” of the grass- mannian – the generating function for simplest genus-0 K-theoretic GW-invariants of it. In Section 4 we combine the same technique with the idea known as “non-abelian localiza- tion” [2] in order to prove the invariance of the genus-0 quantum K-theory of the grassmannian under a suitable infinite dimensional group of pseudo-finite-difference operators. A key point here (inspired by the appendix in the paper [22] by K. Hori and C. Vafa) is to begin with the toric quotient Hom ( Cn,CN ) //Tn = ( CPN−1 )n by the maximal torus Tn ⊂ GLn(C), and use Weyl-group invariant finite-difference operators on n Novikov’s variables of the toric manifold. Just as in the case of toric manifolds [16], this infinite dimensional group of symmetries is large enough in order to reconstruct “all” genus-0 invariants of the grassmannian from the small J-function. In Section 5, we address the aforementioned comparison problem by interpreting (in several somewhat different ways) the q-hypergeometric series arising from quasimap theory of the cotan- gent bundle spaces T ∗Grn,N as certain “genuine” K-theoretic GW-invariants, and in particular show that, contrary to a naive belief articulated in the literature, these series fail to represent “small” J-functions (of anything). In Section 6, we apply the invariance result from Section 5 to illustrate the “non-abelian quantum Lefschetz” principle which characterizes genus-0 quantum K-theory of a complete intersection in (or a vector bundle space over) the grassmannian. In Section 7, we show how our techniques can be used to extend (to the case of grassmannians) the toric results obtained by Y. Ruan and M. Zhang [32] about the level structures in quantum K-theory. As a by-product, we clarify (hopefully) the phenomenon of level correspondence between “dual” grassmannians Grn,N = GrN−n,N discovered recently by H. Dong and Y. Wen [7]. In Section 8, we exhibit a Jackson-type integral formula for the small J-function in the quantum K-theory of the grassmannian, inspired by the “non-abelian localization” framework from Section 4. Our logic is the same as in the aforementioned appendix [22] by K. Hori and C. Vafa, where cohomological mirrors of Grn,N were proposed. However, our mirror formula looks different (and possibly new, see [27]) even in the cohomological GW-theory. Namely, this cohomological mirror of the grassmannian has the form of complex oscillating integral I := ∫ Γ⊂XQ e( ∑ ij xij− ∑ i6=i′ yii′)/z ∧ ij d lnxij ∧ i 6=i′ dyii′∧ i d ln (∏ j xij/ ∏ i′(yii′/yii′) ) . Here XQ is the complex torus in the linear space with coordinates {xij}, i = 1, . . . , n, j = 1, . . . , N and {yii′}, i, i′ = 1, . . . , n, i 6= i′, given by n equations∏ j xij = Q ∏ i′ (yii′/yi′i), i = 1, . . . n, and Γ is a combination of Lefschetz thimbles in XQ, invariant under the Weyl group Sn acting on the coordinates {xij}, {yii′} by simultaneous permutations of the indices i and i′. Quantum K-Theory of Grassmannians and Non-Abelian Localization 3 As a mirror symmetry test, let us examine the critical set of the phase function (“superpo- tential”) using Lagrange multipliers p1, . . . , pn:∑ i,j xij − ∑ i 6=i′ yii′ − ∑ i pi (∑ j lnxij − ∑ i′ 6=i (ln yii′ − ln yi′i)− lnQ ) . The critical points are determined from xij = pi, yii′ = pi − pi′ , pNi + (−1)nQ = 0, where the third set of equations comes from the constraints. The algebra of functions on the critical set (which is a finite lattice ZnN ⊂ Tn) invariant under permutations of (p1, . . . , pn) is indeed isomorphic to the “small” quantum cohomology algebra of the grassmannian (as described by formula (3.39) in [35]). 2 The small J-function of Grn,N Let X := Grn,N be the grassmannian of n-dimensional subspaces V ⊂ CN . Its K-ring K0(X) is generated by the exterior powers ∧k V of the tautological bundle, k = 1, . . . , n. Using the splitting principle, we will often write them as elementary symmetric functions∑ 1≤i1<···<ik≤n Pi1 · · ·Pik of K-theoretic Chern roots of V = P1 + · · ·+ Pn. Proposition 2.1. The K-theoretic Poincaré pairing on K0(X) is given by residue formula χ(X; Φ(P )) = (−1)n ResP=1 Φ(P ) ∏ i 6=j(1− Pi/Pj) (1− P1)N · · · (1− Pn)N dP1 ∧ · · · ∧ dPn P1 · · ·Pn , where Φ is any symmetric Laurent polynomial of P1, . . . , Pn. The formula is obtained as the non-equivariant limit Λ → 1 from its TN -equivariant coun- terpart, where TN is the torus of diagonal matrices diag(Λ1, . . . ,ΛN ) acting on CN . Proposition 2.2. The TN -equivariant K-theoretic Poincaré pairing on K0 T (X) is given by χT (X; Φ(P,Λ)) = (−1)n n! ResP 6=0,∞ Φ(P,Λ) ∏ i 6=j(1− Pi/Pj)∏n i=1 ∏N j=1(1− Pi/Λj) dP1 ∧ · · · ∧ dPn P1 · · ·Pn . Here Φ is a Laurent polynomial in P and Λ, symmetric in P , χT is the T -equivariant holo- morphic Euler characteristic, taking values in the representation ring Z[Λ±] of the torus, and the residue sum is taken over all poles P1 = Λi1 , . . . , Pn = Λin (with this ordering of the equations, and iα 6= iβ). The formula is proved by the direct application of Lefschetz’ holomorphic fixed point formula. The following q-hypergeometric series has emerged from a study of spaces of rational curves in the grassmannian based on their quasimap compactifications: J = ∑ 0≤d1,...,dn Qd1+···+dn∏n i=1 ∏di m=1(1− qmPi)N n∏ i,j=1 ∏di−dj m=−∞(1− qmPi/Pj)∏0 m=−∞(1− qmPi/Pj) . Remark 2.3. The product on the right can be rearranged as∏ di>dj q( di−dj 2 ) ( −Pi Pj )di−dj Pj − qdi−djPi Pj − Pi , and therefore may contain nilpotent factors Pj − Pi in the denominator. It is not hard to see, however, that transposing Pi and Pj does not change the sum of terms with a fixed d1 + · · ·+dn, implying that after clearing the denominators, the numerator becomes divisible by Pj − Pi (namely, it changes sign under the transposition, and hence vanishes when Pi = Pj). 4 A. Givental and X. Yan Remark 2.4. Another consequence of the above rearrangement is that, with the exception of the term Q0, the series consists of reduced rational functions of q. Namely, the factor at Qd1+···+dn has no pole at q = 0, and the q-degree of the denominator exceeds that of the numerator by N n∑ i=1 ( di + 1 2 ) − ∑ di>dj ( di − dj + 1 2 ) ≥ N − n+ 1 ≥ 2. Theorem 2.5 (cf. [33, 34]). The series (1− q)J is the “small J-function” of the grassmannian Grn,N (C). Recall that the genus-0 quantum K-theory of a target space X, the “big J-function” is defi- ned as t 7→ J (t) := 1− q + t(q) + ∑ d,m,α Qdφα 〈 φα 1− qL0 , t(L1), . . . , t(Lm) 〉Sm 0,m+1,d . Here Q is the Novikov’s variable, {φα} and {φα} are Poincaré-dual bases in K0(X), t = ∑ k tkq k is a Laurent polynomial in q with vector coefficients tk ∈ K0(X) ⊗ Q[[Q]], and the correlator represents the K-theoretic GW-invariant computes a suitable holomorphic Euler characteristic on the moduli spaces of stable maps Xg,m+1,d :=Mg,m+1(X, d) with the input (or “insertion”) at the marked point (with the index i = 0, . . . ,m in the above formula) of the form ∑ k(ev∗i tk)L k i , where Li stands for the universal cotangent line bundle at the ith marked point. From among several flavors of such K-theoretic GW-invariants (ordinary as in [21], or permutation-equivariant as in [17]), we will currently use the permutation-invariant ones (as the superscript Sm indicates), i.e., computing the super-dimension of the part of the sheaf cohomology on the moduli space X0,m+1,d which is invariant under permutations of the m marked points with the indices i = 1, . . . ,m carrying the symmetric inputs t(Li). The “small J-function” is obtained from J by setting the input t = 0. In particular, this eliminates the role of the permutation group, and so J (0) represents the “ordinary” K-theoretic GW-invariants. Thus, according to Theorem 2.5, J (0) := (1− q) + ∑ d,α Qdφα 〈 φα 1− qL0 〉 0,1,d = (1− q)J. Theorem 2.5 is obtained as the non-equivariant limit Λ→ 1 from the following result about the TN -equivariant version J T “big J-function” of the grassmannian. Theorem 2.6 (cf. [33, 34]). J T (0) = (1− q)JT , where JT = ∑ 0≤d1,...,dn Qd1+···+dn∏n i=1 ∏N j=1 ∏di m=1(1− qmPi/Λj) n∏ i,j=1 ∏di−dj m=−∞(1− qmPi/Pj)∏0 m=−∞(1− qmPi/Pj) . Note that by the very definition (the same as for J with the correlators taking values in the representation ring Z[Λ±]), the function J T is a Q-series with coefficients which are rational functions of q with vector values in K := K0 T (X) ⊗ Q[[Q]]. Abusing the language we call such series rational functions of q, denote the space they form by K := K(q±), and call it the loop space. The part (1 − q) + t(q) (“dilaton shift”+”input”) belongs to the subspace K+ consisting of Laurent polynomials (they can have poles only at q = 0,∞), while the sum of the correlators belongs (as it is not hard to see) to the complementary subspace K− = {f ∈ K \| f(0) 6=∞, f(∞) = 0.}. It follows from Remark 2.4 above (which applies to JT as well) that (1− q)JT ≡ 1− q mod K−. Thus, the non-obvious statement of Theorem 2.6 is that (1− q)JT represents a value of J T at all. Quantum K-Theory of Grassmannians and Non-Abelian Localization 5 The technique of fixed point localization we intend to use goes back to paper [3] by J. Brown, and was adapted to the K-theoretic situation in [11]. The technique, applicable whenever the target carries a torus action with isolated fixed points and isolated 1-dimensional orbits, com- pletely characterizes all values of the big J-function J T as the set of those rational functions f ∈ K which pass two tests: criterions (i) and (ii). They are formulated in terms of specializa- tions fα of f to the fixed point of the torus action in X. In the case of the grassmannian, take for example the fixed point V1,...,n = Span(e1, . . . , en) where (we may assume by choosing the ordering) (P1, . . . , Pn) = (Λ1, . . . ,Λn): JT(1,...,n) = ∑ 0≤d1,...,dn Qd1+···+dn∏n i=1 ∏N j=1 ∏di m=1(1− qmΛi/Λj) n∏ i,j=1 ∏di−dj m=−∞(1− qmΛi/Λj)∏0 m=−∞(1− qmΛi/Λj) . Note that the 1st factor contains the product (coming from j = i): 1∏n i=1 ∏di m=1(1− qm) with poles at roots of unity, while all other poles are elsewhere (at q = (Λi/Λj) −1/m). Each term of the series considered as rational functions of q can be split (e.g., using partial fraction decomposition) into the sum of a reduced rational function with poles at the roots of unity and a rational function with poles elsewhere. The result will be interpreted (or rather termed) as a meromorphic function in a neighborhood of the roots of unity. Criterion (i) stipulates that fα, when interpreted as a meromorphic function in a neighborhood of the roots of unity, must represent a value (over a suitable ground ring) of the big J-function of the point target space. We will return to this criterion in the next section and explain how it can be verified. Criterion (ii) controls residues of fα(q)dq/q at the poles originating from T -equivariant covers of 1-dimensional orbits. Namely, the tangent space to the grassmannian at the fixed point V(1,...,n) carries the torus action with the distinct eigenvalues Λj/Λi, i = 1, . . . , n, j = n + 1, . . . , N . Consequently, for each choice of i and j there is a 1-dimensional orbit, which compactifies into CP 1 connecting this fixed point with another one. For instance, taking i = 1 and j = n+1, we find such an orbit connecting V(1,...,n) with V(2,...,n+1). Let φ : CP 1 → CP 1 be the map z 7→ zm0 ramified at z = 0,∞ (representing the two fixed points which we call α and β). Criterion (ii) has the form of the recursion relation: Resq=(Λj/Λi)1/m0 fα(q) dq q = −Q m0 m0 Eu(TαX) Eu(TφX0,2,m0) fβ ∣∣∣ q=(Λj/Λi)1/m0 , where Eu are equivariant K-theoretic Euler classes: of the tangent space to X at α, and to the moduli space of degree-m0 stable maps with 2 marked points at the point represented by the m0-fold cover φ respectively. We compute Resq=(Λn+1/Λ1)1/m0 (1 − q)JT(1,...,n)(q)dq/q, replacing d1 with d1 + m0, assuming di = d1 when i = n+ 1, and using x := (Λn+1/Λ1)1/m0 : Resq=x(1− q)JT(1,...,n)(q) dq q = −(1− x) Qm0 m0 × 1∏N j=1 ∏m0 m=1 (j,m) 6=(n+1,m0) (1− xmΛ1/Λj) n∏ j=2 m0∏ m=1 1− xmΛ1/Λj 1− xmΛj/Λn+1 × ∑ 0≤d1,...,dn Qd1+···+dn∏n+1 i=2 ∏N j=1 ∏di m=1(1− xmΛi/Λj) n+1∏ i,j=2 ∏di−dj m=−∞(1− xmΛi/Λj)∏0 m=−∞(1− xmΛi/Λj) . 6 A. Givental and X. Yan The sum together with the factor 1−x yields (1− q)JT(2,...,n+1)(q)\|q=x, so it remains to interpret the recursion coefficient in terms of the Euler classes. Applying Lefschetz’ fixed point formula on Grn,N (C), we’ve already used that Eu(T(1,...,n)X)=∏n i=1 ∏N j=n+1(1 − Λi/Λj). In order to compute Eu(TφX0,2,m0), we note that the 1-dimensional orbit connecting the fixed points Span(e1, . . . , en) and Span(e2, . . . , en+1) consists of subspaces Vt := Span(te1 + (1− t)en+1, e2, . . . , en−1). Consequently, restricted to CP 1 = {Vt}, the tangent bundle to the grassmannian, which has the form Hom(Vt,CN/Vt), can be described in terms of the Hopf bundle L over CP 1 and its complementary L′ := Span(e1, en+1)/L = Λ1Λn+1L −1 as Hom ( L⊕ Span(e2, . . . , en), L′ ⊕ Span(en+2, . . . , eN ) ) . On the m0-fold cover φ : CP 1 → CP 1, the contributions to the Euler class of the T -modules H0 ( CP 1;φ∗L−1 ⊗ Span(ej) ) and H0 ( CP 1;φ∗L′ ⊗ Span(ei) −1 ) are respectively m0∏ m=0 (1− xmΛ1/Λj) and m0∏ m=0 (1− xmΛi/Λn+1). The contribution of H0 ( CP 1;φ∗L−1⊗L′ ) is (as in the case of X = CP 1) ∏m0 m=−m0,m 6=0(1−xm). Combing all the contributing factors, we find Eu(T(1,...,n)X) Eu(TφX0,2,m0) = ∏n i=1 ∏N j=n+1(1− Λi/Λj)∏n i=2 ∏N j=n+2(1− Λi/Λj) 1∏m0 m=−m0,m6=0(1− xm) × 1∏N j=n+2 ∏m0 m=0(1− xmΛ1/Λj) 1∏n i=2 ∏m0 m=0(1− xmΛi/Λn+1) . Checking that this expression matches exactly the recursion coefficient for JT (the middle line in the formula for the residue) is the matter of a straightforward (though somewhat cumbersome) rearrangement of the factors. Remark 2.7. Note that the structure of the recursion relations (ii) and the values of the recursion coefficients completely characterize the big J-function of a particular theory, since criterion (i) does not involve any additional choices. 3 Quantum K-theory of the point Genus-0 permutation-equivariant K-theoretic GW-invariants of the point are represented by the “big J-function” of the form Jpt(t) := (1− q) + t(q) + ∞∑ m=2 χ ( M0,m+1/Sm; 1 1− L0q ⊗mi=1 t(Li) ) . Here Li are the universal cotangent line bundles over the Deligne–Mumford spaces M0,m+1. The holomorphic Euler characteristic χ on the orbifoldM0,m+1/Sm computes the super-dimen- sion of the Sm-invariant part of the sheaf cohomology on M0,m+1. The insertions t(Li) (and hence the input t(q)) can be in fact any rational functions of Li (respectively of q) as long as they don’t have poles at roots of unity. On the contrary, each χ-term is a reduced rational function of q with poles only at the roots of unity (of order ≤ m). Both this and the previous claim easily follow from the general structure of the Lefschetz fixed point formula (applied on M0,m+1/Sm). Quantum K-Theory of Grassmannians and Non-Abelian Localization 7 As it is explained in [10, 17], the action of permutation groups on the sheaf cohomology is captured by the above Sm-invariants taking values in an arbitrary λ-algebra, i.e., a ring R equipped with the Adams operations Ψk : R→ R. E.g., for τ ∈ R χ ( M0,m/Sm;⊗mi=1 ( τLdi )) := 1 m! ∑ h∈Sm ∞∏ k=1 (Ψkτ)lk(h) strhH ∗(M0,m;⊗mi=1L d i ) , where lk(h) denotes the number of cycles of length r in the cycle decomposition of permutation h. For mode detail, we refer to [10] or [17], where this example is extrapolated to general R-valued insertions t(Li). Note that only Ψr with r > 0 are used in this definition. We assume that R is complete in the adic topology defined by a certain ideal R+ ⊂ R, which is respected by the Adams operations Ψk with k > 0 in the sense that Ψk(R+) ⊂ Rk+, and that the input t is “small” in the sense that it takes values in R+. In fact the latter property guarantees that the m-th χ-term of the series Jpt takes values in Rm+ , with assures R+-adic convergence of the series. The genus-0 permutation-equivariant GW-invariants of the point target space are completely described in [12]. Namely, given a ground λ-algebra R, the range of the big J-function t 7→ Jpt(t), which is a semi-infinite cone (that we will denote Lpt) in the (completed) space of R-valued rational functions of q (which we will denote R(q±)) is explicitly parameterized as (1− q)e ∑ k>0 Ψk(τ)/k(1−qk) ( 1 +R+[q±] ) . Here τ ∈ R+, and the notation R+[q±] is reserved for the completion in the R+-adic topology of the space of rational functions of q which have no poles at roots of unity and take values in R+. In our arguments, we will take advantage of the possibility to replace one ground λ-algebra with another related to it by a homomorphism respecting the Adams operations. In simple terms: If some f ∈ R(q±) is known to lie in Lpt (over a given ground ring R), i.e., the part f− with poles at the roots of unity represents K-theoretic GW-invariants with the input defined by the part f+ with poles away from roots of unity, the same will remain true when the values of some parameters (coordinates on SpecR) are specialized in a way commuting with Ψk for all k > 0. Another important property of the cone Lpt that we will rely on is its invariance under a cer- tain group of (pseudo) finite-difference operators. Namely, let R = Q[[Q]], where ΨkQ = Q\|k\|, k = ±1,±2, . . . , and let D(qQ∂Q , Q) be a finite difference operator (which we should assume “small” in R+-adic sense to assure convergence). It is almost obvious that the linear vector field f 7→ Df in R(q±) is tangent to Lpt, and therefore eD preserves Lpt. Moreover, according to a result from [13], Lpt is preserved by the operator f 7→ e ∑ k>0 Ψk(D(q kQ∂Q ,Q))/k(1−qk)f . Our goal in this section is to verify that the series JT(1,...,n) from the previous section satis- fies criterion (i), i.e., that it represents a value of Jpt when interpreted as a meromorphic function of q in a neighborhood of roots of unity. For this, we begin with the ground ring R = Q [ Λ±1 , . . . ,Λ ± N ] [[Q1, . . . , Qn]], with the Adams operations acting by Ψk ( Λ±1 j ) = Λ±kj , Ψk(Qi) = Q \|k\| i , and set R+ = (Q1, . . . , Qn). In particular, taking τ = Q1 + · · · + Qn, we find that Lpt 3 (1− q)Jpt, where Jpt is the following product of q-exponential functions: Jpt := e ∑n i=1 ∑ k>0 Q k i /k(1−qk) = n∏ i=1 ∞∑ di=0 Qdii∏di m=1(1− qm) = ∑ 0≤d1,...,dn Qd1 1 · · ·Qdnn∏n i=1 ∏di m=1(1− qm) . 8 A. Givental and X. Yan Following [13], we are going to use Q-independent finite-difference operators of the form Dl,Λ := −qΛ(1− ql·Q∂Q), where l ·Q∂Q := ∑ i liQi∂Qi , and Λ is a formal variable added to the ground ring, ΨkΛ := Λ\|k\|. The corresponding Lpt-preserving operator is Γl,Λ := e− ∑ k>0 Λk(1−qkl·Q∂Q )qk/k(1−qk). This expression is in fact the asymptotical expansion (near the unit circle on the q-plane) of the ratio: 0∏ m=−∞ ( 1− Λql·Q∂Qqm ) / 0∏ m=−∞ (1− Λqm). The ratio and its asymptotical expansion act on monomials Qd = Qd1 1 · · ·Qdnn the same way: Γl,ΛQ d = Qd ∏0 m=−∞(1− Λqm+(l·d))∏0 m=−∞(1− Λqm) = Qd ∏l·d m=−∞(1− Λqm)∏0 m=−∞(1− Λqm) . Note that the right hand side is a rational function of q with poles away from roots of unity. Considering Γl,Λ(1− q)Jpt as a point of Lpt over the ground ring R[[Λ]], we conclude therefore, that being a Q-series with coefficients which are rational function of q with coefficients polyno- mial in Λ, it is a point of Lpt over R[Λ]. Thus, it will turn into a point of Lpt over R when Λ is replaced by any non-trivial monomial from Q[Λ±1 , . . . ,Λ ± N ], e.g., by Λi/Λj with i 6= j. In order to complete our fixed point localization proof of Theorem 2.6, we apply to Jpt the following operators (where l = 1i contains 1 in the i-th position and 0 everywhere else):( n∏ i=1 N∏ j=1,j 6=i Γ−1 1i,Λi/Λj )( n∏ i,j=1 Γ1i−1j ,Λi/Λj ) Jpt = ∑ 0<d1,...,dn Qd1 1 · · ·Qdnn∏n i=1 ∏N j=1 ∏di m=1(1− qmΛi/Λj) n∏ i,j=1 ∏di−dj m=−∞(1− qmΛi/Λj)∏0 m=−∞(1− qmΛi/Λj) . The terms of the last sum are interpreted as meromorphic functions in the neighborhood of roots of unity, i.e., with poles (which come from the factors with i = j in the left product) at the roots of unity only. When multiplied by 1 − q, the latter series lies in Lpt over the ground λ-algebra R = Q[Λ±][[Q1, . . . , Qn]]. The substitution Q1 = · · · = Qn = Q (which induces a homomorphism of λ-algebras R→ R0 := Q[Λ±][[Q]]) yields therefore a series which lies in the range Jpt over R0. It actually coincides with the localization (1−q)JT(1,...,n) of JT at the indicated fixed point. There- fore (1 − q)JT(1,...,n), when interpreted as a series of meromorphic functions near roots of unity, satisfies criterion (i). Due to the Weyl group symmetry between all fixed points, and between all 1-dimensional orbits connecting them, this finishes the proof of Theorem 2.6. 4 Non-abelian localization and explicit reconstruction The approach to computing GW-invariants of GIT quotients via non-abelian localization (and eventually quasimap compactifications) was proposed by A. Bertram, I. Ciocan-Fontanine and B. Kim [2] following their proof [1] of the Hori–Vafa conjecture. The conjecture (formulated in the appendix to [22]) gave a novel proposal for the mirrors of GIT quotients CM//G. The idea, illustrated by K. Hori and C. Vafa in the example of the grassmannians, was to replace the factorization by a semi-simple G with the (cohomologically equivalent to it) succession of the Quantum K-Theory of Grassmannians and Non-Abelian Localization 9 factorizations by its maximal torus T and then by its Weyl group W = N(T )/T . The first step yields a toric manifold, whose mirror and genus-0 GW-invariants are well-understood. In the case of the grassmannian Grn,N = Hom(Cn,CN )//GLn(C), it is the product X̃ := (CPN−1)n of projective spaces. Its small TN -equivariant (K-theoretic) J-function is (1− q)JT X̃ , where JT X̃ = ∑ d1,...,dn≥0 Qd1 1 · · ·Qdnn∏N j=1 ∏n i=1 ∏di m=1(1− qmPi/Λj) , where Pi are the Hopf bundles over the factors. The second step can be described this way: JT = JTΠg/t\|Q1=···=Qn=Q,Λ0=1, where JTΠg/t := ∏ i 6=j Γ1i−1j ,Λ0Pi/Pj J T X̃ . The Γ-operators here, Γ1i−1j ,Λ0Pi/Pj = e− ∑ k>0(Λ0Pi/Pj) k ( 1−q kQi∂Qi −kQj∂Qj ) qk/k(1−qk), correspond to the roots of g, i.e., in our case of g = gln(C) to the line bundles Pi/Pj for i 6= j. Explicitly JTΠg/t = ∑ d1,...,dn≥0 Qd1 1 · · ·Qdnn∏N j=1 ∏n i=1 ∏di m=1(1− qmPi/Λj) ∏ i 6=j ∏di−dj m=−∞(1− qmΛ0Pi/Pj)∏0 m=−∞(1− qmΛ0Pi/Pj) . According to [14], (1 − q)JTΠg/t represents a value of the big J-function of the super-space ΠE, where E is a vector bundle over X̃, equal to ⊕i 6=jPj/Pi in the case at hands, which is associ- ated with the adjoint action of the maximal torus on g/t, and Π indicates the parity change of the fibers. By definition, the quantum K-theory of such a super-space is obtained by sys- tematically replacing the virtual structure sheaves Og,m,d of the moduli spaces X̃g,m,d with Og,m,d ⊗EuC×(ft∗ ev∗E), where the subscript in the K-theoretic Euler class indicates it is equi- variant with respect to the scalar action of Λ0 ∈ C× on the fibers of E, and ft : X̃g,m+1,d → X̃g,m,d and ev : X̃g,m+1,d → X̃ are respectively the forgetting of and evaluation at the last marked point. On the other hand, the explicit reconstruction results of [16] tell us how to parameterize the entire big J-function of a toric manifold (or super-manifold) from one value of it. Namely, the range of the big J-function, LΠg/t in our example, is invariant under the action of a huge group, P, of pseudo-finite-difference operators in Novikov’s variables Q1, . . . , Qn. It is generated by the exponentials eD of any (R+-adically small) finite-difference operators D(PqQ1∂Q , Q, q), and by operators of the form1 e ∑ k>0 Ψk ( D(Pq kQ∂Q ,Q,q)/k(1−qk ) . The orbit of (1 − q)JT X̃ under this group is the whole of L X̃ ( and moreover, picking suitable operators as described in [16] one obtains an explicit parameterization of L X̃ ) . Theorem 4.1. Elements of the orbit of (1 − q)JTΠg/t under the subgroup PW of the operators invariant with respect to the Weyl group, in the specialization Q1 = · · · = Qn = Q, Λ0 = 1 turn into values of the big J-function of the grassmannian. 1Note that above operators Γ1i−1j ,Λ0Pi/Pj are the compositions of the operators of multiplication by e ∑ k>0 Ψk(Λ0Pi/Pj)/k(1−q−k), whose cumulative effect, according to the Adams–Riemann–Roch (see [18]), is to transform LX̃ to LΠg/t, and of the operators of this form with D = qΛ0Piq Qi∂Qi /Pjq Qj∂Qj which preserve LΠg/t. 10 A. Givental and X. Yan We conjecture that a similar result holds universally for non-singular GIT quotients, i.e., that LC//G is obtained from LWΠg/t (where Πg/t is the super-space over the base C//T defined as explained above) by specializing the Novikov ring to its W -invariant part, and passing to the limit Λ0 = 1. Proof. The proof of the theorem is based on TN -fixed point localization. It should be obvi- ous after Section 3 that the criterion (i) of the fixed point method is invariant under P even before the specialization to Qi = Q,Λ0 = 1. To verify (ii), take localizations ( JTΠg/t ) (1,...,n) and( JTΠg/t ) (n+1,2,...,n) of JTΠg/t (thinking of the fixed points Span(e1, . . . , en) and Span(e2, . . . , en+1) in the grassmannian). The ambiguity in the ordering of the values Pi = Λi′ becomes irrelevant in the limit Q1 = · · · = Qn = Q due to the W -symmetry. Before the limit, we take here P1 = Λ1, . . . , Pn = Λn for the first fixed point, and P1 = Λn+1, P2 = Λ2, . . . , Pn = Λn for the second. For x = (Λ1/Λn+1)−1/m0 , we have Resq=x ( JTΠg/t ) (1,...,n) dq q = −Q m0 1 m0 Coeff (2,...,n+1) (1,...,n) (m0) ( JTΠg/t ) (n+1,2,...,n) ∣∣ q=x . This is simply the recursion relation (ii) for the target space X̃ = ( CPN−1 )n corresponding to the 1-dimensional TN -orbit connecting two fixed points, Span(e1) and Span(en+1) in projection to the first factor CPN−1, and constant (and equal to Span(ei), i = 2, . . . , n) in the other projections. The recursion coefficient here turns into the correct one for the grassmannian in the limit Λ0 = 1 and Q1 = Q. Note that operators P ki q kQi∂Qi specialize to Λki q kQi∂Qi at the fixed point Span(e1, . . . , en), and for i > 1 commute with Qm0 1 , while for i = 1 we have Λk1q kQ1∂Q1Qm0 1 = Qm0 1 qkm0Λk1q kQ1∂Q1 ≡ mod 1−qm0 Λ1 Λn+1 Qm0 1 Λkn+1q kQ1∂Q1 . This implies that for any finite difference operator D regular at q = (Λ1/Λn+1)−1/m0 , the localizations ( DJTΠg/t ) (1,...,n) and ( DJTΠg/t ) (n+1,2,...,n) of DJTΠg/t also satisfy the above recursion relation with the same recursion coefficient. In fact this direct verification is not even necessary, since it simply elucidates in terms of fixed point localization the general fact that LΠg/t is P-invariant. We conclude that when D is W -invariant, (1 − q)DJTΠg/t specializes at Q1 = · · · = Qn = Q and Λ0 = 1 into a point in the loop space K (corresponding to the grassmannian) which satisfies the correct recursion relation, and hence belongs to LGrn,N . � Remark 4.2. Of course, the above argument applies more generally than the grassmannian example, and works whenever a torus (TN in this case) acts on C//Tn and C//G with isolated fixed points and isolated one-dimensional orbits. In particular, it applies to twisted quantum K-theories studied in [19] and generalizing the above transition from X̃ = ( CPN−1 )n to Πg/t. Namely, let E = E(P1, . . . , Pn) ∈ K0 TN (Grn,N ) be a virtual vector bundle (for this, E needs to be symmetric in Pi). It can be used to “twist” the virtual structure sheaves of the moduli spaces of stable maps – for both targets, Πg/t and Grn,N : Og,m,d 7→ Og,m,d ⊗ e ∑ k 6=0 Ψk(µk ft∗ ev ∗E)/k, where µk are some prefixed elements of the ground ring R (and, abstractly speaking, should better be taken from R+ as a precaution lest the modifying expression diverges). Then the big J-functions in the twisted quantum K-theories of Πg/t and Grn,N are related the same way as described in the theorem: For any W -invariant value of the twisted J-function of Πg/t (in place of JTΠg/t), the elements of its orbit under PW in the limit Q1 = · · · = Qn = Q specialize into values of the big J-function in the twisted quantum K-theory of the grassmannian. Quantum K-Theory of Grassmannians and Non-Abelian Localization 11 5 Balanced I-functions and T ∗Grn,N In some recent literature motivated by representation theory (see, e.g., [24, 29, 30]), quantum K-theory of symplectic quiver varieties plays a role, and among them, the cotangent bundles of the grassmannians (rather than the grassmannians per se) take the place of the target spaces. K-theoretic computations in the quasimap compactifications of spaces of rational curves in such targets lead A. Okounkov and his followers to q-hypergeometric functions quite interesting from the point of view of the theory of integrable systems. To illustrate one specific property (appa- rently important in their theory for technical reasons) consider the series IT = ∑ 0≤d1,...,dn Qd1+···+dn n∏ i=1 N∏ j=1 di∏ m=1 1− qmY Pi/Λj 1− qmPi/Λj × n∏ i,j=1 ∏di−dj m=−∞(1− qmPi/Pj)∏0 m=−∞(1− qmPi/Pj) ∏0 m=−∞(1− qmY Pi/Pj)∏di−dj m=−∞(1− qmY Pi/Pj) . Here Y ∈ C× (denoted in [29] and elsewhere by ~) represents the circle acting by scalar multi- plication on the fibers of a vector bundle over the compact base (which is meant to be T ∗Grn,N in our example). Note that the series is formed of fractions (1− qmY X)/(1− qmX), which are bounded both as q → 0 and q →∞. In the fixed-point computations on quasimap spaces of sym- plectic targets, this property of generating functions being balanced (in terminology of [26]) is a by-product of tensoring the virtual structure sheaf with the square root of the determinant bundle of the moduli space (i.e., in effect computing indexes of real Dirac operators rather than holomorphic Euler characteristics). The questions we will address here are about the place of the series IT and its close counterparts in the “genuine” (i.e., based on stable map compactifica- tions) quantum K-theory of the grassmannian: Does IT represent a value of the big J-function of any version of quantum K-theory, and if so, then what version and on which space? Is it the small J-function in that theory? We will give several different affirmative answers to the first question, and negative to the second. Theorem 5.1. The series (1 − q)IT represents a value of the torus-equivariant, permutation- invariant big J-function of ΠTGrn,N (the odd tangent bundle of the grassmannian). Initially the interest in GW-theory of ΠE for a bundle E over a compact base is motivated by the fact that in the non-equivariant limit Y → 1, GW-invariants of ΠE, when the limit exist, turn into GW-invariants of the zero locus of a generic section of E (which in the case of E = TX consists of χ(X) isolated points). Proof. We can follow the same route as that of Theorem 2.6. The localization IT(1,...,n) = ∑ 0≤d1,...,dn Qd1+···+dn n∏ i=1 N∏ j=1 di∏ m=1 1− qmY Λi/Λj 1− qmΛi/Λj × n∏ i,j=1 ∏di−dj m=−∞(1− qmΛi/Λj)∏0 m=−∞(1− qmΛi/Λj) ∏0 m=−∞(1− qmY Λi/Λj)∏di−dj m=−∞(1− qmY Λi/Λj) of IT at the fixed point (1, . . . , n) in the grassmannian, together with such localizations at other fixed points, pass the test (i) of the fixed point theory, and the residues at the poles q = (Λj/Λi) 1/m0 satisfy the recursion relation of the familiar form (ii) with suitable recursion coefficients. This should be obvious after our analysis of the series JT in Sections 3 and 2 respec- tively. Moreover, according to Remark 2.7, it only suffices to match the values of these recursion 12 A. Givental and X. Yan coefficients with those in GW-theory of ΠTGrn,N . Using the notation x = (Λn+1/Λ1)1/m0 and our result from Section 2, we find the recursion coefficient corresponding to the pole at q = x in the form −Q m0 m0 Eu(T(1,...,n)X) Eu(TφX0,2,m0) as before, times the modifying factor N∏ j=1 m0∏ m=1 (1− xmY Λ1/Λj) n∏ i=2 m0∏ m=1 1− xmY Λi/Λn+1 1− xmY Λ1/Λi , where the target X = Grn,N . Unsurprisingly, the modifying factor is almost reciprocal to the expression for Eu(T(1,...,n)X)/Eu(TφX0,2,m0). They differ by the presence of Y in each factor, and by the extra factor 1−xm0Y Λ1/Λn+1 (actually equal to 1−Y , and excluded from the expression in Section 2 where Y = 1). In our computation of H0 ( CP 1;φ∗(TX) ) , the latter (zero) factor represents the line spanned by the vector field zm0∂zm0 (infinitesimally rescaling the target CP 1), and falls out of TφX0,2,m0 because of the infinitesimal automorphism z∂z of the source CP 1. Thus, the factor 1−Y remains present in EuC×(ft∗ φ ∗(TX)). Note that Y was introduced as the character of C×-action on T ∗X. The action on TX is given therefore by Y −1, but the definition of the K-theoretic Euler class as the exterior algebra of the dual bundle restores the factors Y everywhere. Thus, the modifying factor coincides with EuC×(ft∗ φ ∗(TX))/EuC×(T(1,...,n)X), and the recursion coefficient altogether has the required form −Q m0 m0 Eu(T(1,...,n)ΠTX) Eu(T(1,...,n)(ΠTX)0,2,m0) . � Corollary 5.2. The series (1−q)IT /(1−Y q) represents a value of the torus-equivariant, permu- tation-invariant big J-function in the quantum Hirzebruch K-theory of the grassmannian Grn,N . Recall that the Hirzebruch χ−Y -genus of a compact complex manifold M is defined by χ−Y (M) := dimM∑ p=0 (−Y )pχ(M ; Ωp(M)) = H∗(M ; EuC×(TM)), where the rightmost interpretation assumes that Y ∈ C× acts fiberwise on the tangent bundle by Y −1. More generally, one can define the (classical) Hirzebruch K-theory by replacing the structure sheaves OM with OM ⊗EuC×(TM). The quantum Hirzebruch K-theory of a target X is defined by similarly modifying the virtual structure sheaves of the moduli spaces Xg,m,d using their virtual tangent bundles: Og,m,d 7→ Og,m,d ⊗ EuC×(TXg,m,d). According to a result from [23], the theory thus obtained can be expressed via the ordinary quantum K-theory, implying in particular Corollary 5.2 (see Remark 5.3 below). However, it also follows from our fixed point approach. Namely, the big J-function of (permutation-invariant) quantum Hirzebruch K-theory has the form 1− q 1− qY + t(q) + ∑ d,m,α Qdφα 〈 φα 1− qY L0 1− qL0 , t(L1), . . . , t(Lm) 〉Sm 0,m+1,d , where the correlators are defined using the virtual structure sheaves of the Hirzebruch K-theory. This is not an ad hoc definition, but is dictated by the general formalism; the dilaton shift and the Quantum K-Theory of Grassmannians and Non-Abelian Localization 13 first input embody respectively: the Euler class (of the universal line bundle q−1) corresponding to the genus, and the reciprocal of the equivariant Euler class of L−1 0 . Consequently, the recursion coefficient of the fixed point theory acquires a new factor 1 − Y : the residue of 1−qY L0 1−qL0 dq q at the pole q = L−1 0 (equal in our computations to (Λ1/Λn+1)−1/m0). But the above explanation why this factor belongs to EuC×(ft∗ φ ∗(TX)) means it does not belong to EuC×(TφX0,2,m0). The latter occurs in Lefschetz’ fixed point formula for the modified virtual structure sheaf. The net result is that the recursion relation (ii) remains the same as in the theory of ΠTX. Note that a scalar factor, such as 1/1−qY in (1−q)IT /(1−qY ) has no effect on the recursion relation (a fact indicating that the range L ⊂ K of the big J-function is an “overruled cone”). The role of this factor is to guarantee that modulo Q, the series equals the dilaton shift, and hence the rest of the series is Q-adically small as required. Remark 5.3. By the way, 1/(1 − qY ) = ∑ m≥0 Y mqm is considered a “Laurent polynomial” in q, i.e., an element of K+ in Hirzebruch K-theory, as it doesn’t have poles relevant in local- ization theory. The correlator part of the big J-function in the quantum Hirzebruch K-theory clearly satisfies J \|q=∞ = Y J \|q=0, and this condition defines the new space K−. The general result of [23], which applies to the all-genera permutation-equivariant quantum K-theory, says that the total descendant potential DYX for the Hirzebruch version of the theory is obtained from the “ordinary” one, D0 X , by three transformations: the Eulerian twisting corresponding to the bundle E = TX − 1 (in genus 0, this has practically the same effect as the twisting by TX, pro- ducing the big J-function of ΠTX), and the above changes in the dilaton shift and polarization K = K+ ⊕K−. The transformations correspond to the three summands in the virtual tangent bundles: TXg,m,d = ft∗ ev∗(TX − 1) + ft∗ ( 1− L−1 ) − (ft∗ j∗OZ)∗, where L is the universal cotangent line over Xg,m+1,d at the m+ 1-st marked point, and j : Z → Xg,m+1,d is the inclusion of the nodal locus. Here ft∗ ev∗ TX represents variations of stable maps from pointed curves with a fixed complex structure, ft∗(L −1) represents variations of the complex structure of the curves, while the last term is supported on the virtual divisor ft(Z) ⊂ Xg,m,d where the combinatorics of the curves changes, and accounts for the difference between the virtual tangent bundle and the sheaf of vector fields tangent to this divisor. Another form of Theorem 5.1 can be derived from Serre’s duality. The cotangent line bun- dle L of a pointed nodal curve and its canonical bundle K are related by L = K(D), where D := ∑m i=1 σi is the divisor of the marked points (i.e., away from the nodes, a section of L is a differential allowed to have 1st order poles at the markings). Given a bundle E over X, on Xg,m,d we have ft∗ ev∗E = − ( ft∗K ev∗E∨ )∨ = − ( ft∗ ev∗E∨ )∨ + ( ft∗(1− L) ev∗E∨ )∨ − m∑ i=1 ev∗i E. Applying the quantum Adams–Riemann–Roch (Theorem 2 in [19]), we find that tensoring of Og,m,d with Eu−1 C× (( ft∗(1−L) ev∗E∨ )∨) in the correlators of permutation-equivariant quan- tum K-theory is equivalent to the change (1 − q) 7→ (1 − q) Eu−1 C×(E) in the dilaton shift. The same change of this inputs: t 7→ Eu−1 C×(E)t, is effected by tensoring with the Euler classes of − ev∗i E. In other words, the dilaton-shifted total descendant potential of the theory twisted by Eu−1 (( ft∗ ev∗E∨ )∨) is obtained from the one twisted by EuC×(ft∗ ev∗E) by the trans- formation DΠE(q) 7→ DΠE ( Eu−1 C×(E)q ) . The potentials are considered as quantum states in suitable Fock spaces, and the transformation is induced by the map f 7→ Eu−1 C×(E)f between two copies of the loop space K (equipped with two different symplectic structures: based on the 14 A. Givental and X. Yan Poincaré pairing χ ( X; Eu−1 C×(E)ab ) on the source space, and χ(X; EuC× ab) on the target. Con- sequently, the big J-functions in the genus-0 theory are related by the inverse transformation: JΠE 7→ EuC×(E)JΠE . Applying all this to E = TGrn,N , we arrive at the following conclusion. Corollary 5.4 (cf. [26]). The series EuC×(TGrn,N )(1 − q)IT represents a value of the big J-function in the torus-equivariant, permutation-invariant quantum K-theory of the grassman- nian Grn,N twisted by O0,m,d 7→ O0,m,d ⊗ Eu−1 C× ( (ft∗ ev∗(T ∗Grn,N ))∨ ) . One more way of modifying virtual structure sheaves, which was recently introduced and explored by Y. Ruan and M. Zhang [32], consists in tensoring Og,m,d with a power of the determinant line bundle (det(ft∗ ev∗E))−l, thereby bringing the level structure (of level l) into the quantum K-theory. Note that in terms of K-theoretic Chern roots L1, . . . , LM of a vector bundle E , Eu(E) Eu(E∗) = ∏M k=1 ( 1− L−1 k )∏M k=1(1− Lk) = (−1)ML−1 1 · · ·L −1 M = (−1)dim E(det E)−1. So, we take E = T ∗Grn,N , E = ft∗ ev∗(T ∗Grn,N ), and describe the modification of Og,m,d used in Corollary 5.4 as tensoring with both Eu−1 C×(E) and (det E)−1. After the first operation we land in the theory of the noncompact bundle space T ∗Grn,N , and after the second in the level 1 version of this theory. The Poincaré pairing changes accordingly into χ ( X; Eu−1 C×(T ∗Grn,N ) (detT ∗Grn,N )−1ab ) . By the Riemann–Roch formula, dim E = (1 − g) dimX + ∫ X c1(T ∗X), which for g = 0 yields (−1)dim E = (−1)dim Grn,N (−1)Nd. The first sign is absorbed by the ratio of the Euler classes (of T and T ∗) in the Poincaré pairings, and the second by the change Q 7→ (−1)NQ, leading to the following conclusion. Corollary 5.5. The series det(T ∗Grn,N )−1 EuC×(T ∗Grn,N ) · (1− q)IT ( (−1)NQ ) represents a value of the big J-function in the level 1, torus-equivariant, permutation-invariant quantum K-theory of the cotangent bundle space T ∗Grn,N . Explicitly, the product of the determinant and the Eulerian pre-factor differs by the sign (−1)dim Grn,N from EuC×(TGrn,N ) = ∏N j=1 ∏n i=1(1− Y Pi/Λj)∏n i,j=1(1− Y Pi/Pj) . Because of this pre-factor, the series even modulo Q is not equal the dilaton shift 1− q (as well as in Corollary 5.4), which already disqualifies it for the role of the “small” J-function. In fact the q-hypergeometric series which arises in the K-theoretic computations on the spaces of quasimaps to the grassmannian is slightly different from the one in Corollaries 5.4 and 5.5. It has the form ĨT = ∑ 0≤d1,...,dn Qd1+···+dn n∏ i=1 N∏ j=1 di−1∏ m=0 1− qmY Pi/Λj 1− qmPi/Λj × n∏ i,j=1 ∏di−dj m=−∞(1− qmPi/Pj)∏0 m=−∞(1− qmPi/Pj) ∏−1 m=−∞(1− qmY Pi/Pj)∏di−dj−1 m=−∞ (1− qmY Pi/Pj) , Quantum K-Theory of Grassmannians and Non-Abelian Localization 15 which differs from IT in that in the products of the factors 1− qmY Pi/Λj , the range of m is not from 1 to di (as for the factors without Y ) but from 0 to di − 1 (and similarly for the factors 1−qmY Pi/Pj). We claim, however, that (1−q)Ĩ and (1−q)(−1)dim Grn,N ĨT ( (−1)NQ ) represent some values of the big J-functions of the same theories as described in Corollary 5.4 and 5.5 respectively. Namely, consider the version of ĨT with Qd1+···+dn is replaced with Qd1 1 · · ·Qdnn , and apply to it the operator∏N j=1 ∏n i=1 ( 1− Y qQi∂QiPi/Λj )∏n i,j=1 ( 1− Y qQi∂Qi−Qj∂QjPi/Pj ) . This results in restoring the “missing” factors with m = di or m = di − dj , and in the limit Q1 = · · · = Qn = Q yields the series of Corollary 5.5 (modulo to the sign (−1) dimGrn,N and the change Q 7→ (−1)NQ). On the other hand, the operator can be written as e − ∑ k>0 Y k [∑N j=1 ∑n i=1 q kQi∂QiPki /Λ k j− ∑n i,j=1 q kQi∂Qi −kQj∂Qj Pki /P k j ] /k , and hence belongs to the group PW , which justifies our claim due to Theorem 4.1 and Re- mark 4.2. The series (1−q)ĨT appears to have better chances to pose for the “small” J-function, because the term with Q0 is 1 − q, and other terms are reduced rational functions of q. And indeed, H. Liu [26], looking for a stable-map K-theory interpretation of the q-hypergeometric series arising in the quasimap K-theory of quiver varieties, shows that in the case n = 1 of projective spaces, the series (1 − q)ĨT is the small J-function in the theory described by Corollary 5.4. However, he falls short of sticking to this interpretation, because he finds an example (namely the manifold of flags in C3) where the similarly twisted small J-function is unbalanced. In fact none of (1 − q)ĨT with n > 1 (and none of other I-series featuring in this section) represent “small” J-functions, and not because some rational functions are not reduced, but for much more dramatic reasons. Namely, in our fixed point characterization of the big J-function, the poles participating in the recursion relations come from the characters of the torus action on the grassmannian per se: q = (Λi/Λj) −1/m. The terms of a balanced I-series containing the factors 1/(1− qmY Pi/Pj) lead to the poles at q = (Y Λi/Λj) −1/m, which cannot come from fixed point localization. Such fractions should therefore be interpreted as elements of K+, i.e., as geometric series ∑ k≥0 q mk(Y Pi/Pj) k converging in the Y -adic topology. Thus, representing (1−q)ĨT as (1−q)+t(q) mod K− results in a very complicated value of t(q), meaning that the series represents the value of the big J-function with the inputs t(Li) which are rather far from 0. What makes the effect even more dramatic is that it is the input in the permutation-invariant quantum K-theory, no counterpart of which has been discussed so far in the context of quasimap spaces. Apart from this, the interpretation of the series given in Corollary 5.5 is quite parallel to its definition [24, 29, 30] in the quasimap theory as a generating function capturing some K-theoretic GW-invariants of the cotangent bundle of the grassmannian based on the virtual structure sheaves “symmetrized” by the determinant factors. Finally, we would like to stress that, although we have formulated Theorem 5.1 and its corollaries as statements about the particular I-function, modified slightly in one way or another, in fact these modifications affect the recursion coefficients in a simple and controllable way, implying that the whole big J-functions of the respective theories coincide up to these minor modifications. In particular, Corollary 5.5 is connected to Theorem 5.1 by a general phenomenon called the “non-linear (or quantum) Serre duality” [5]. It relates GW-invariants of the super- space ΠE and bundle space E∨, and was first observed in [8] (for cohomological GW-invariants) 16 A. Givental and X. Yan via fixed point localization. For the full treatment (including higher genus) of the K-theoretic reincarnation of the quantum Serre duality we refer to [36]. 6 Non-abelian quantum Lefschetz A somewhat different proof of Theorem 5.1 could be derived from Theorem 4.1 together with Remark 4.2, applied to GW-invariants of the grassmannian Euler-twisted by the tangent bundle E = N∑ j=1 n∑ i=1 Λj/Pi − n∑ i,j=1 Pj/Pi. Here we illustrate this approach using as an example Eulerian twistings applied to the dual tautological bundle E = P−1 1 + · · ·+ P−1 n . The Euler-twisted theory (of both Πg/t and Grn,N ) is defined by Og,m,d 7→ Og,m,d ⊗ EuC×(ft∗ ev∗E) = Og,m,d ⊗ e− ∑ k>0 Y kΨ−k(ft∗ ev∗ E)/k, where Y ∈ C× acts by multiplication on the fibers of E. According to the quantum Adams– Riemann–Roch theorem [19], the twisted theory is obtained from the untwisted one by the multiplication: LΠE = 2−1LGrn,N , where 2 := e ∑ k>0 Y kΨ−k(E)qk/k(1−qk) = EuC×(E) ∑ d1,...,dn≥0 Y d1+···+dnP d1 1 · · ·P dnn∏n i=1 ∏di m=1(1− qmPi) . This is a convenient moment to address one general technical issue. Values of big J-functions are supposed to lie in R+-neighborhood of the dilaton shift 1−q. The terms containing Novikov’s variables are R+-small, and remain such after multiplication by anything like 2. Moreover, for Laurent polynomials t(q) with R+-small coefficients, 2 t contains only finitely many non- reduced terms, and so modulo K− it remains a Laurent polynomial (with R+-small coefficients). However, the product 2(1− q) ≡ (1− q) +Y E∨ mod K− seems to present a problem. One way to resolve it is to postulate that R+ 3 Y . Here is a better way to deal with this issue, which is especially useful if one also needs to use Y −1 in the same context. Consider the operator D = e ∑ k>0 Y kqk ∑n i=1 P k i (1−qkQi∂Qi )/k(1−qk), which is 2 times the pseudo-finite-difference operator from the group PW corresponding to the finite-difference operator −qY ∑n i=1 Piq Qi∂Qi . Therefore 2−1LΠg/t = D−1LΠg/t, which by Theorem 4.1 (or rather its generalization explained in Remark 4.2) turns into LΠE in the limit Q1 = · · · = Qn = Q, Λ0 = 1. The advantage of using D instead of 2 is that D does not change terms constant in Q1, . . . , Qn: D(1− q) = 1− q. Theorem 6.1 (non-abelian quantum Lefschetz). Suppose∑ d1,...,dn≥0 Id1,...,dn(P1, . . . , Pn,Λ0)Qd1 1 · · ·Q dn n is a W -invariant point in LΠg/t, then ∑ d1,...,dn≥0 Id1,...,dn \|Λ0=1Q d1+···+dn n∏ i=1 d∏ m=1 (1− qmY Pi) is a point in LΠE. Quantum K-Theory of Grassmannians and Non-Abelian Localization 17 Proof. Apply D−1 and use e− ∑ k>0 q kY kPki (1−qQi∂Qi )/k(1−qk)Qdii = e− ∑ k>0 q kY kPki (1−qdik)/k(1−qk)Qdii = e− ∑ k>0 ∑di m=1 q mkY kPk/kQdii =Qd di∏ m=1 (1− qmY Pi). � Corollary 6.2. The small J-function of ΠE equals (1− q)I, where I = ∑ d1,...,dn≥0 Q ∑ di ∏n i=1 ∏di m=1(1− qmY Pi)∏N j=1 ∏n i=1 ∏di m=1(1− qmPi/Λj) n∏ i,j=1 ∏di−dj m=−∞(1− qmPi/Pj)∏0 m=−∞(1− qmPi/Pj) . 7 Level structures and dual grassmannians Of course, the approach illustrated by Theorem 6.1 applies to any bundle E over the grassman- nian, since E can always be written as a symmetric combination of monomials ∏ i P li i . We are going to use this together with the observation (see Section 5) that det−1(−E) = Eu(E)/Eu(E∨) in order to describe the effect of the level structure on the genus-0 quantum K-theory of the grass- mannian. For the sake of illustration, we take E to be the tautological bundle V = P1 + · · ·+Pn. With E := ft∗ ev∗ V , we have level-l twisted structure sheaves Og,m,d ⊗ det−l(−E) = Og,m,d ⊗ Eu(lE) Eu(lE∨) = Og,m,d ⊗ e−l ∑ k 6=0 Ψk(E)/k. The Adams–Riemann–Roch theorem from [19] yields the multiplication operator 2 = e−l ∑ k 6=0 Ψk(V )/k(1−qk), and the respective pseudo-finite-difference operator D = e−l ∑ k 6=0 ∑n i=1 P k i (1−qkQi∂Qi )/k(1−qk). Applying it to Qdii , we find DQdii = Qdii e−l ∑ k>0 ∑di−1 m=0 (Pki q mk−P−ki q−mk)/k = Qdii di−1∏ m=0 ( 1− Piqm 1− P−1 i q−m )l = Qdii (−Pi)ldiql( di 2 ). The above calculation is somewhat formal. The initial determinantal twisting of Og,m,d and the finial modifying factors are well-defined, but in order to justify intermediate steps, one needs add to R+ two variables Y , Y ′, and replace E and E∨ with Y E and Y ′E∨. This will lead to the product of fractions (1−Y Piqm)/(1−Y ′P−1 i q−m), where one can pass to the limit Y = Y ′ = 1, thus obtaining the following result. Theorem 7.1 (cf. [32]). Suppose∑ d1,...,dn≥0 Id1,...,dn(P1, . . . , Pn,Λ0)Qd1 1 · · ·Q dn n is a W -invariant point in LΠg/t. Then ∑ d1,...,dn≥0 Id1,...,dn \|Λ0=1Q d1+···+dn n∏ i=1 ( P dii q (di2 ) i )l is a point in L(V,l) Grn,N . 18 A. Givental and X. Yan Here L(V,l) Grn,N is the range of the big J-function in the level-l permutation-invariant genus-0 quantum K-theory of the grassmannian Grn,N , where V is its tautological bundle, and the level-l twisted structure sheaves are defined as in [32]: Og,m,d⊗det−l(ft∗ ev∗ V ). Note, that the spurious signs (−1)l dim E = (−1)l ∑ di+l dimV initially introduced in the determinantal twisting disappears from our ultimate formulation. The first part of it can be absorbed by the change Q 7→ (−1)lQ of the Novikov variable, which is offset by the signs of (−Pi)ldi in our computation. The second part, (−1)l dimV , is the discrepancy in Poincaré pairings (it affects the notion of dual bases {φα}, {φα} in the definition of J-functions) which correspond to the two twistings of OGrn,N : by Eu(lV )/Eu(lV ∨) = ∏ (−Pi)−l and det−l(V ) = ∏ P−li . The theorem is a non-abelian counterpart of the result of Ruan–Zhang for toric manifolds, obtained in [32] on the basis of adelic characterization. Both can also be derived by fixed point localization. Corollary 7.2. The series (1− q)IT(V,l), where IT(V,l) = ∑ 0≤d1,...,dn Qd1+···+dn∏n i=1 P ldi i ql( di 2 )∏n i=1 ∏N j=1 ∏di m=1(1− qmPi/Λj) n∏ i,j=1 ∏di−dj m=−∞(1− qmPi/Pj)∏0 m=−∞(1− qmPi/Pj) represents a point in L(l) Grn,N , and for −n < l ≤ N − n + 1 is the small J-function of the level-l theory. Proof. The formula itself is obtained, of course, from the non-abelian representation of the small J-function (1 − q)JT = (1 − q)IT(V,0) by the recipe described in the theorem. For l ≥ 0, terms of (1− q)IT(V,l) have no pole at q = 0, and for l ≤ N − n+ 1 can be shown to be reduced rational functions of q (except the Q0-term 1− q). Indeed, the difference between the q-degrees of the denominator and numerator of the coefficient of IT(V,l) indexed by (d1, . . . , dn) is N ∑ i ( di + 1 2 ) − ∑ di>dj ( di − dj + 1 2 ) − l ∑ i ( di + 1 2 ) + l ∑ i di. When l ≤ N − n + 1, the binomial sum is non-negative, since for each i the number of j with dj < di does not exceed n − 1. The linear term is > 1 unless all di 6= 1. Note that in this case the whole expression doesn’t depend on l, and is still ≥ N − n + 1 > 1. Thus, even after multiplication by 1− q the rational function remains reduced. For l < 0, the terms of the series are therefore also reduced, but can have a pole at q = 0. However, even when this happens, the pole disappears after summing the terms with the same degree d1 + · · · + dn – at least when l > −n. This follows from a non-trivial combinatorial result of H. Dong and Y. Wen [7], according to which IT(V,l) = ĨT(V ∨,−l) for −n < l < N − n, where ĨT (Ṽ ∨,−l) is the similar series corresponding to the dual grassmannian GrN,N−n := Hom ( CN−n,CN∨ ) //GLN−n(C), and the bundle Ṽ ∨ dual to the tautological one: ∑ 0≤d1,...,dn Qd1+···+dN−n ∏N−n i=1 P̃−ldii q−l( di+1 2 )∏N−n i=1 ∏N j=1 ∏di m=1(1− qmP̃i/Λ−1 j ) N−n∏ i,j=1 ∏di−dj m=−∞(1− qmP̃i/P̃j)∏0 m=−∞(1− qmP̃i/P̃j) . Here P̃i are K-theoretic Chern roots of the tautological N − n-dimensional bundle Ṽ . Note the characters Λ−1 j of the torus TN action on CN∨. Also note the binomial coefficient ( di+1 2 ) ( instead of ( di 2 )) : this is the effect of using Ṽ ∨ rather than Ṽ in the construction of the determinantal Quantum K-Theory of Grassmannians and Non-Abelian Localization 19 twistings. In the previous section we already had the experience of using E = V ∨, from which it is easy to infer the origin of the modifying factors: di∏ m=1 1− qmP̃i 1− q−mP̃i = ( −P̃ )diq(di+1 2 ). The dual grassmannians are canonically identified by ( V ⊂ CN ) 7→ ( V ⊥ ⊂ CN∨ ) , and the result of [7] identifies the two expressions as Q-series with coefficients in K0 T (Grn,N ) = K0 T (GrN,N−n)- valued rational functions of q when −n < l < N −n. By the previous estimates of the q-degrees( where this time the linear term l ∑ di isn’t present ) , (1− q)ĨT (Ṽ ∨,−l) passes the requirements to be a small J-function when 0 ≤ −l < N − (N − n) + 1, i.e., 0 ≥ l ≥ −n. Therefore (though this is not apparent) so does (1− q)ITV,l at least for 0 > l > −n. � Example. For Gr1,N = CPN−1, we have (1− q)IT(l) = (1− q) ∑ d≥0 QdP lql( d 2)∏N j=1 ∏d m=1(1− qmP/Λj) . Obviously the series is the small J-function only when −1 < l ≤ N , i.e., the boundaries given by the corollary are sharp. Proposition 7.3. L(V,l) Grn,N = L(Ṽ ∨,−l) GrN,N−n . Proof. From Ṽ ∨ = CN/V we find det(ft∗ ev∗ V )⊗ det ( ft∗ ev∗ Ṽ ∨ ) = det ( ft∗ ev∗CN ) which over moduli spaces of rational curves equals detCN = ∏N j=1 Λj = detV ⊗ det Ṽ ∨, the factor absorbed by the discrepancy in Poincaré pairings between the two theories. � 8 Mirrors Consider the improper Jackson integral (or q-integral), defined as∫ ∞ 0 f(X) X−1dqX := ∑ d∈Z f ( q−d ) , in the example f(X) = X ln Λ/ ln q ∞∏ m=1 (1−X/qm). For \|q\| > 1, the infinite product converges to an entire function of X which coincides with the q-exponential function eX/(1−q)q = ∑ d≥0 Xd (1− q) ( 1− q2 ) · · · ( 1− qd ) . Since the integrand vanishes at X = q−d with d < 0, the q-integral can be computed as∑ d≥0 Λ−d ∞∏ m=d+1 ( 1− q−m ) = ∑ d≥0 Λ−d ∏∞ m=1 ( 1− q−m )( 1− q−1 )( 1− q−2 ) · · · ( 1− q−d ) = ∏∞ m=1 ( 1− q−m )∏∞ m=0 ( 1− q−m/Λ ) . 20 A. Givental and X. Yan The last equality holds for \|Λ\| > 1, but analytically extends the value of the q-integral to all Λ 6= q−m, m = 0, 1, 2, . . . . The ratio is closely related to the q-gamma function (see [6] for a modern treatment of it, including the above q-integral representation). In particular, the application of the translation operator qΛ∂Λ : Λ→ qΛ results in the multiplication of the whole expression by 1− Λ−1. This also follows from the properties of the integrand:( 1− qX∂X/Λ ) f(X) = Xf(X) = qΛ∂Λf(X), since the q-integral is obviously preserved by the translation qX∂X of the integrand. The latter property of q-integrals will be more useful to us than the previous explicit calculation of their values in terms of q-gamma functions. More generally, one can define improper q-integrals using shifted multiplicative q-lattices{ qd/A \| d ∈ Z } :∫ ∞/A 0 g(Y ) Y −1dqY := ∫ ∞ 0 g(Y/A)Y −1dqY = ∑ d∈Z g ( q−d/A ) . We will need such q-integrals (still assuming \|q\| > 1) for g(Y ) = Y ln Λ/ ln q∏∞ m=0(1− Y/qm) . The integrand satisfies( 1− Λq−Y ∂Y ) g(Y ) = Y g(Y ) = qΛ∂Λg(Y ), implying that the q-integral, if defined, is multiplied by (1 − Λ) when Λ is replaced by qΛ. In fact at A = 1 the q-integral is not defined (because of the factor 1− Y in the denominator). However, as it is shown in [6] (see formulas (1.12), (1.13), and (1.15) keeping in mind that q there corresponds to our q−1), it is well-defined for A 6= qd. The value of this q-integral does depend on A, but the properties remain the same. For the sake of certainty we may use A = −1, and indicate this by the notation ∫ −∞ 0 g(Y )Y −1dqY . Our goal will be to represent the small J-function JTX of the grassmannian X = Gn,N by suitable Jackson-like integrals in a fashion similar to representing cohomological J-functions by complex oscillating integrals in the mirror theory of, say, toric or flag manifolds. To maintain visual resemblance with complex oscillating integrals, we will denote X−1dqX as d lnqX, and often replace the infinite products in the integrands with their asymptotical expressions: ∞∏ m=1 (1−X/qm) ∼ e ∑ k>0 X k/k(1−qk), 1∏∞ m=0(1− Y/qm) ∼ e− ∑ k>0 q kY k/k(1−qk). Let us recall from Section 4 that JTX is obtained from JTΠg/t by passing to the limit Λ0 = 1, Q1 = · · · = Qn = Q, and takes values in K0 T (X) consisting of symmetric functions of P1, . . . , Pn. Before the limit, JTΠg/t is the J-function of a toric superspace. We begin with setting up the toric mirror (cf. [15]) to this toric superspace, and studying its properties. In the complex torus X with multiplicative coordinates Xij , i = 1, . . . , n, j = 1, . . . , N , Yii′ , i, i′ = 1, . . . , n, i 6= i′, consider the n-parametric family of tori XQ1,...,Qn := { (X,Y ) ∈ X ∣∣ ∏ j Xij = Qi ∏ i′ 6=i (Yii′/Yi′i), i = 1, . . . , n } , Quantum K-Theory of Grassmannians and Non-Abelian Localization 21 and introduce the q-integral I := ∫ Γ⊂XQ1,...,Qn e ∑ k>0( ∑ i,j X k ij−qk ∑ i6=i′ Y k ii′)/k(1−qk) × ∏ i 6=i′ Yii′ ∧ i (∧ j dq lnXij ∧ i′ 6=i dq lnYii′ )∧ i (∑ j dq lnXij − ∑ i′ 6=i dq ln(Yii′/Yi′i) ) . To clarify the wedge-product expression: if the subscript in all dq is removed, the expression becomes the standard translation-invariant holomorphic volume on the complex torus XQ1,...,Qn (and coincides with the one found in Introduction). By the “cycle” Γ we understand a “multiplicative” q-lattice in XQ1,...,Qn , i.e., a ln q-lattice on the universal covering of the torus of rank nN + n2 − 2n, suitable for multi-dimensional q-integration, or a formal linear combination of such q-lattices; we’ll meet some examples later. This should be considered as the K-theoretic mirror to what we denoted in Section 4 by Πg/t: the toric super-bundle over ( CPN−1 )n = CNn//Tn with the fiber g/t associated with the adjoint action of the maximal torus Tn in GLN (C) on Lie GLN (C)/LieTn. Namely, in the torus-non- equivariant limit JΠg/t, the “small J-function” JTΠg/t introduced in Section 4 satisfies (as it is not hard to check) the system of finite difference equations∏ i′ 6=i ( 1− qPi′P−1 i q Qi′∂Qi′ −Qi∂Qi )( 1− PiqQi∂Qi )N JΠg/t = Qi ∏ i′ 6=i ( 1− qPiP−1 i′ q Qi∂Qi−Qi′∂Qi′ ) JΠg/t, i = 1, . . . , n. So, the claim is that our mirror q-integral satisfies the same system (for scalar-valued rather than K0(X)-valued functions):∏ i′ 6=i ( 1− q qQi′∂Qi′−Qi∂Qi )( 1− qQi∂Qi )NI = Qi ∏ i′ 6=i ( 1− q qQi∂Qi−Qi′∂Qi′ ) I, i = 1, . . . , n. To check this, we note that translation operators q Xij∂Xij and q −Yii′∂Yii′ project to the Q-space into respectively qQi∂Qi and q Qi∂Qi−Qi′∂Qi′. Applying q Xij∂Xij to the factor4ij :=e ∑ k>0X k ij/k(1−qk) in the integrand of I containing Xij , we obtain e ∑ k>0 q kXk ij/k(1−qk) = e ∑ k>0X k ij/k(1−qk)e− ∑ k>0X k ij/k = (1−Xij)4ij . Therefore, applying ∏ j(1 − q Xij∂Xij ), we find the integrand multiplied by ∏ j Xij . Similarly, applying q −Yii′∂Yii′ to ∇ii′ := e− ∑ k>0 q kY k ii′/k(1−qk)Yii′ , we obtain (1 − Yii′)q −1∇ii′ , and hence applying 1 − q q−Yii′∂Yii′ we find the integrand multiplied by (1 − q(1 − Yii′)q−1) = Yii′ . Since∏ i′ 6=i Yi′i ∏ j Xij = Qi ∏ i′ 6=i Yii′ for i = 1, . . . , n, the promised finite difference equations follow. In this argument it was assumed that the family of q-integration lattices Γ ⊂ XQ1,...,Qn depending on Q1, . . . , Qn was invariant under all coordinate multiplicative q-translations in the ambient torus X . Also note that the same argument applies to the ordinary (as opposed to Jackson’s) integrals, provided that (the families of) the cycles of integration are homologous to their q-translates. The catch is that it is not entirely clear how to produce such lattices and/or cycles. Below we will resolve this catch in the T -equivariant case. The torus-equivariant counterpart IT of I is obtained by inserting into the integrand the factor∏ i,j X ln Λj/ ln q ij ∏ i 6=i′ Y ln Λ0/ ln q ii′ . 22 A. Givental and X. Yan By repeating the above computations, we find that IT satisfies finite difference equations∏ i′ 6=i ( 1− qΛ0q Qi′∂Qi′ −Qi∂Qi )∏ j ( 1− qQi∂Qi/Λj ) IT = Qi ∏ i′ 6=i ( 1− qΛ0q Qi∂Qi−Qi′∂Qi′ ) IT , i = 1, . . . , n. Replacing the natural action of finite difference operators on scalar-valued functions with the representation on K0 T (X)-valued functions by qQi∂Qi 7→ Piq Qi∂Qi , we obtain the equations satis- fied by the series JTΠg/t from Section 4. Let us now examine IT for a cycle Γ fitting coordinate charts on XQ1,...,Qn . Picking an injec- tive function J : {1, . . . , n} → {1, . . . , N}, we express XiJ(i), i = 1, . . . , n, in terms of Qi and the remaining variables, using the equations of XQ1,...,Qn , and then rewrite the integral IT in this chart. For instance, taking J(i) = i we find Xii = Qi ∏ j 6=iX −1 ij ∏ i′ 6=i(Yii′/Yi′i). Consequently IT(1,...,n) = ∑ 0≤d1,...,dn ∏ i Q di+ln Λi/ ln q i (1− q)(1− q2) · · · (1− qdi) I(d1,...,dn) (1,...,n) , where I(d1,...,dn) (1,...,n) = ± ∫ Γ(1,...,n) e ∑ k>0( ∑ j 6=iX k ij−qk ∑ i 6=i′ Y k ii′)/k(1−qk) × ∏ i 6=j X −di+ln(Λj/Λi)/ ln q ij ∏ i 6=i′ Y di−d′i+1+ln(Λ0Λi/Λi′ )/ ln q ii′ ∧ i 6=j dq lnXij ∧ i 6=i′ dq lnYii′ , and the sign is determined by the order of the q-differentials and an orientation of the multi- plicative q-lattice Γ(1,...,n). This is the product of model 1-dimensional q-integrals I (d) + := ∫ ∞ 0 e ∑ k>0 X k/k(1−qk)X−d+ln(Λ′/Λ)/ ln qdq lnX, I (d) − := ∫ −∞ 0 e− ∑ k>0 q kY k/k(1−qk)Y 1−d+ln(Λ0Λ′/Λ)/ ln qdq lnX, considered at the beginning of this section. They can be expressed via I (0) ± by the recursive property of the q-gamma-like function. Explicitly, applying to the integrand of I (d) + the operator 1− qdqX∂XΛ/Λ′, we find (after a short computation) that ( 1− qdΛ/Λ′ ) I (d) + = I (d−1) + . Applying this inductively we conclude that I (d) + = I (0) + / ∏d m=1(1 − qmΛ/Λ′). Using this and a similar recursion for I (d) − , we can reduce I(d1,...,dn) (1,...,n) to I(0,...,0) (1,...,n): I(d1,...,dn) (1,...,n) = I(0,...,0) (1,...,n)∏ i 6=j ∏di m=1(1− qmΛi/Λj) ∏ i 6=i′ ∏di−d′i m=−∞(1− qmΛ0Λi/Λi′)∏0 m=−∞(1− qmΛ0Λi/Λi′) . Comparing this to the terms of the series JTΠg/t from Section 4 localized at the fixed point (P1, . . . , Pn) = (Λ1, . . . ,Λn), we arrive at IT(1,...,n) = ( JTΠg/t ) (1,...,n) ∏ i Q ln Λi/ ln q i I(0,...,0) (1,...,n). Note that I(0,...,0) (1,...,n) doesn’t depend on Q, while the role of the factor Q ln Λi/ ln q i is to conju- gate qQi∂Qi into Piq Qi∂Qi \|Pi=Λi . Thus, one can say that (J TΠg/t)(1,...,n) is given by the q-integral Quantum K-Theory of Grassmannians and Non-Abelian Localization 23 over the “cycle” which is the formal multiple of the q-lattice Γ(1,...,n) (lifted from our chart to the torus XQ1,...,Qn) with the coefficient inverse to ∏ iQ ln Λi/ ln q i I(0,...,0) (1,...,n). When the indexing function J : {1, . . . , n} → {1, . . . , N} is changed to a permutation σ of {1, . . . , n}, the value of the q-integral does not change ITσ(1),...,σ(n) = IT1,...,n, assuming that the order the orientation of the q-lattice Γ(σ(1),...,σ(n)) is consistent with the order of the q-differentials in the wedge-product. Thus, the above computation extended to arbitrary injective indexing functions J shows that all relevant components of the vector-function JTΠg/t can be represented by our q-integrals using “cycles” fitting appropriate charts. Setting in the q-integral Q1 = · · · = Qn = Q and Λ0 = 1, we obtain a torus-equivariant K-theoretic mirror of the grassmannian. Theorem 8.1. The multi-dimensional q-integral ITX := ∫ Γ⊂XQ e ∑ k>0( ∑ i,j X k ij−qk ∑ i 6=i′ Y k ii′)/k(1−qk) × ∏ i,j X ln Λj/ ln q ij ∏ i 6=i′ Yii′ ∧ i (∧ j dq lnXij ∧ i′ 6=i dq lnYii′ )∧ i (∑ j dq lnXij − ∑ i′ 6=i dq ln(Yii′/Yi′i) ) , with suitable choices of (linear combinations of) q-lattices Γ in XQ := { (X,Y ) ∈ X ∣∣ ∏ j Xij = Q ∏ i′ 6=i (Yii′/Yi′i), i = 1, . . . , n } represents components of the K0 T (X)-valued small J-function JTX of the grassmannian X=Grn,N . Acknowledgments This material is based upon work supported by the National Science Foundation under Grant DMS-1906326. We are thankful to P. Koroteev and A. Smirnov for their effort in educating us about their work on quantum K-theory of symplectic quiver varieties, and to H. Liu and Y. Wen for sharing and discussing their preprints. References [1] Bertram A., Ciocan-Fontanine I., Kim B., Two proofs of a conjecture of Hori and Vafa, Duke Math. J. 126 (2005), 101–136, arXiv:math.AG/0304403. [2] Bertram A., Ciocan-Fontanine I., Kim B., Gromov–Witten invariants for abelian and nonabelian quotients, J. Algebraic Geom. 17 (2008), 275–294, arXiv:math.AG/0407254. [3] Brown J., Gromov–Witten invariants of toric fibrations, Int. Math. Res. Not. 2014 (2014), 5437–5482, arXiv:0901.1290. [4] Ciocan-Fontanine I., Kim B., Maulik D., Stable quasimaps to GIT quotients, J. Geom. Phys. 75 (2014), 17–47, arXiv:1106.3724. [5] Coates T., Givental A., Quantum Riemann–Roch, Lefschetz and Serre, Ann. of Math. 165 (2007), 15–53, arXiv:math.AG/0110142. [6] De Sole A., Kac V.G., On integral representations of q-gamma and q-beta functions, Atti Accad. Naz. Lincei Cl. Sci. Fis. Mat. Natur. Rend. Lincei (9) Mat. Appl. 16 (2005), 11–29, arXiv:math.QA/0302032. [7] Dong H., Wen Y., Level correspondence of K-theoretic I-functions in Grassmannian duality, arXiv:2004.10661. [8] Givental A., Equivariant Gromov–Witten invariants, Int. Math. Res. Not. 1996 (1996), 613–663, arXiv:alg- geom/9603021. https://doi.org/10.1215/S0012-7094-04-12613-2 https://arxiv.org/abs/math.AG/0304403 https://doi.org/10.1090/S1056-3911-07-00456-0 https://arxiv.org/abs/math.AG/0407254 https://doi.org/10.1093/imrn/rnt030 https://arxiv.org/abs/0901.1290 https://doi.org/10.1016/j.geomphys.2013.08.019 https://arxiv.org/abs/1106.3724 https://doi.org/10.4007/annals.2007.165.15 https://arxiv.org/abs/math.AG/0110142 https://arxiv.org/abs/math.QA/0302032 https://arxiv.org/abs/2004.10661 https://doi.org/10.1155/S1073792896000414 https://arxiv.org/abs/alg-geom/9603021 https://arxiv.org/abs/alg-geom/9603021 24 A. Givental and X. Yan [9] Givental A., On the WDVV equation in quantum K-theory, 2000, 295–304, Michigan Math. J. 48 (2000), 295–304, arXiv:math.AG/0003158. [10] Givental A., Permutation-equivariant quantum K-theory I. Definitions. Elementary K-theory of M0,n/Sn, Mosc. Math. J. 17 (2017), 691–698, arXiv:1508.02690. [11] Givental A., Permutation-equivariant quantum K-theory II. Fixed point localization, arXiv:1508.04374. [12] Givental A., Permutation-equivariant quantum K-theory III. Lefschetz’ fixed point formula on M0,n/Sn, arXiv:1508.06697. [13] Givental A., Permutation-equivariant quantum K-theory IV. Dq-modules, arXiv:1509.00830. [14] Givental A., Permutation-equivariant quantum K-theory V. Toric q-hypergeometric functions, arXiv:1509.03903. [15] Givental A., Permutation-equivariant quantum K-theory VI. Mirrors, arXiv:1509.07852. [16] Givental A., Permutation-equivariant quantum K-theory VIII. Explicit reconstruction, arXiv:1510.06116. [17] Givental A., Permutation-equivariant quantum K-theory IX. Quantum-Riemann–Roch in all genera, arXiv:1709.03180. [18] Givental A., Permutation-equivariant quantum K-theory X. Quantum Hirzebruch–Riemann–Roch in genus 0, SIGMA 16 (2020), 031, 16 pages, arXiv:1710.02376. [19] Givental A., Permutation-equivariant quantum K-theory XI. Quantum Adams–Riemann–Roch, arXiv:1711.04201. [20] Givental A., Lee Y.-P., Quantum K-theory on flag manifolds, finite-difference Toda lattices and quantum groups, Invent. Math. 151 (2003), 193–219, arXiv:math.AG/0108105. [21] Givental A., Tonita V., The Hirzebruch–Riemann–Roch theorem in true genus-0 quantum K-theory, in Sym- plectic, Poisson, and noncommutative geometry, Math. Sci. Res. Inst. Publ., Vol. 62, Cambridge University Press, New York, 2014, 43–91, arXiv:1106.3136. [22] Hori K., Vafa C., Mirror symmetry, arXiv:hep-th/0002222. [23] Huq-Kuruvilla I., Multiplicative quantum cobordism theory, arXiv:2101.09305. [24] Koroteev P., Pushkar P.P., Smirnov A., Zeitlin A.M., Quantum K-theory of quiver varieties and many-body systems, arXiv:1705.10419. [25] Lee Y.-P., Quantum K-theory. I. Foundations, Duke Math. J. 121 (2004), 389–424, arXiv:math.AG/0105014. [26] Liu H., Self-duality in quantum K-theory, arXiv:1906.10824. [27] Marsh R.J., Rietsch K., The B-model connection and mirror symmetry for Grassmannians, Adv. Math. 366 (2020), 107027, 131 pages, arXiv:1307.1085. [28] Maulik D., Okounkov A., Quantum groups and quantum cohomology, Astérisque 408 (2019), ix+209 pages, arXiv:1211.1287. [29] Okounkov A., Lectures on K-theoretic computations in enumerative geometry, in Geometry of Moduli Spaces and Representation Theory, IAS/Park City Math. Ser., Vol. 24, Amer. Math. Soc., Providence, RI, 2017, 251–380, arXiv:1512.07363. [30] Pushkar P.P., Smirnov A.V., Zeitlin A.M., Baxter Q-operator from quantum K-theory, Adv. Math. 360 (2020), 106919, 63 pages, arXiv:1612.08723. [31] Rimányi R., Tarasov V., Varchenko A., Trigonometric weight functions as K-theoretic stable envelope maps for the cotangent bundle of a flag variety, J. Geom. Phys. 94 (2015), 81–119, arXiv:1411.0478. [32] Ruan Y., Zhang M., The level structure in quantum K-theory and mock theta functions, arXiv:1804.06552. [33] Taipale K., K-theoretic J-functions of type A flag varieties, Int. Math. Res. Not. 2013 (2013), 3647–3677, arXiv:1110.3117. [34] Wen Y., K-theoretic I-functions of V//θG and applications, arXiv:1906.00775. [35] Witten E., The Verlinde algebra and the cohomology of the Grassmannian, in Geometry, Topology, & Physics, Conf. Proc. Lecture Notes Geom. Topology, Vol. 4, Int. Press, Cambridge, MA, 1995, 357–422, arXiv:hep-th/9312104. [36] Yan X., Serre’s duality in quantum K-theory, and level structures, in preparation. https://doi.org/10.1307/mmj/1030132720 https://arxiv.org/abs/math.AG/0003158 https://doi.org/10.17323/1609-4514-2016-16-4-691-698 https://arxiv.org/abs/1508.02690 https://arxiv.org/abs/1508.04374 https://arxiv.org/abs/1508.06697 https://arxiv.org/abs/1509.00830 https://arxiv.org/abs/1509.03903 https://arxiv.org/abs/1509.07852 https://arxiv.org/abs/1510.06116 https://arxiv.org/abs/1709.03180 https://doi.org/10.3842/SIGMA.2020.031 https://arxiv.org/abs/1710.02376 https://arxiv.org/abs/1711.04201 https://doi.org/10.1007/s00222-002-0250-y https://arxiv.org/abs/math.AG/0108105 https://arxiv.org/abs/1106.3136 https://arxiv.org/abs/hep-th/0002222 https://arxiv.org/abs/2101.09305 https://arxiv.org/abs/1705.10419 https://doi.org/10.1215/S0012-7094-04-12131-1 https://arxiv.org/abs/math.AG/0105014 https://arxiv.org/abs/1906.10824 https://doi.org/10.1016/j.aim.2020.107027 https://arxiv.org/abs/1307.1085 https://doi.org/10.24033/ast https://arxiv.org/abs/1211.1287 https://doi.org/10.1090/pcms/024 https://arxiv.org/abs/1512.07363 https://doi.org/10.1016/j.aim.2019.106919 https://arxiv.org/abs/1612.08723 https://doi.org/10.1016/j.geomphys.2015.04.002 https://arxiv.org/abs/1411.0478 https://arxiv.org/abs/1804.06552 https://doi.org/10.1093/imrn/rns156 https://arxiv.org/abs/1110.3117 https://arxiv.org/abs/1906.00775 https://arxiv.org/abs/hep-th/9312104 1 Introduction 2 The small J-function of Gr n,N 3 Quantum K-theory of the point 4 Non-abelian localization and explicit reconstruction 5 Balanced I-functions and T* Gr n,N 6 Non-abelian quantum Lefschetz 7 Level structures and dual grassmannians 8 Mirrors References
id	nasplib_isofts_kiev_ua-123456789-211170
institution	Digital Library of Periodicals of National Academy of Sciences of Ukraine
issn	1815-0659
language	English
last_indexed	2026-03-15T11:48:05Z
publishDate	2021
publisher	Інститут математики НАН України
record_format	dspace
spelling	Givental, Alexander Yan, Xiaohan 2025-12-25T13:21:10Z 2021 Quantum K-Theory of Grassmannians and Non-Abelian Localization. Alexander Givental and Xiaohan Yan. SIGMA 17 (2021), 018, 24 pages 1815-0659 2020 Mathematics Subject Classification: 14N35 arXiv:2008.08182 https://nasplib.isofts.kiev.ua/handle/123456789/211170 https://doi.org/10.3842/SIGMA.2021.018 In the example of complex Grassmannians, we demonstrate various techniques available for computing genus-0 K-theoretic GW-invariants of flag manifolds and more general quiver varieties. In particular, we address explicit reconstruction of all such invariants using finite-difference operators, the role of the -hypergeometric series arising in the context of quasimap compactifications of spaces of rational curves in such varieties, the theory of twisted GW-invariants, including level structures, as well as the Jackson-type integrals playing the role of equivariant K-theoretic mirrors. This material is based upon work supported by the National Science Foundation under Grant DMS-1906326. We are thankful to P. Koroteev and A. Smirnov for their effort in educating us about their work on quantum K-theory of symplectic quiver varieties, and to H. Liu and Y. Wen for sharing and discussing their preprints. en Інститут математики НАН України Symmetry, Integrability and Geometry: Methods and Applications Quantum K-Theory of Grassmannians and Non-Abelian Localization Article published earlier
spellingShingle	Quantum K-Theory of Grassmannians and Non-Abelian Localization Givental, Alexander Yan, Xiaohan
title	Quantum K-Theory of Grassmannians and Non-Abelian Localization
title_full	Quantum K-Theory of Grassmannians and Non-Abelian Localization
title_fullStr	Quantum K-Theory of Grassmannians and Non-Abelian Localization
title_full_unstemmed	Quantum K-Theory of Grassmannians and Non-Abelian Localization
title_short	Quantum K-Theory of Grassmannians and Non-Abelian Localization
title_sort	quantum k-theory of grassmannians and non-abelian localization
url	https://nasplib.isofts.kiev.ua/handle/123456789/211170
work_keys_str_mv	AT giventalalexander quantumktheoryofgrassmanniansandnonabelianlocalization AT yanxiaohan quantumktheoryofgrassmanniansandnonabelianlocalization

Quantum K-Theory of Grassmannians and Non-Abelian Localization

Institution

Similar Items