betaFeedback

WorksMathematics

Minimal norm of regular vectors in the root lattice of a crystallographic root system

Abstract

For a crystallographic root system ΦΦ with Weyl group WW and root lattice L=ZΦL = \mathbb{Z}Φ, we determine the minimum norm of a vector of LL having trivial stabilizer in WW, and we prove that the vectors attaining it form a single orbit of WW. Although this is a minimization of a quadratic function over the lattice points of a cone, the minimizer in the dominant chamber is always unique: it is ρ+λρ+ λ, where ρρ is the Weyl vector and λλ is 00 or the minuscule weight in the class of ρρ modulo LL. In particular, the optimal vector is the Weyl vector exactly when the Weyl vector lies in LL. We give explicit formulas for the families AnA_n, BnB_n, CnC_n and DnD_n, with elementary proofs, and the values for the exceptional types E6E_6, E7E_7, E8E_8, F4F_4 and G2G_2. Finally, we explain the coincidences between the minimal norms by the folding of Dynkin diagrams.

Introduction

Let Φ\Phi be a crystallographic root system in a Euclidean space VV with Weyl group W=W(Φ)W = W(\Phi), and let L=ZΦL = \mathbb{Z}\Phi be its root lattice. We refer to Humphreys [1 Chapters 1 and 2] for the definitions and basic properties of root systems, Weyl groups and root lattices; crystallographic root systems, root lattices and weight lattices are treated in [1 §2.9]. For the simply laced types, with roots of norm 2, the root lattices are AnA_n, DnD_n, E6E_6, E7E_7 and E8E_8. For the types BnB_n and CnC_n we use the standard realizations in Rn\mathbb{R}^n:

Φ(Bn)={±ei,±ei±ej}andΦ(Cn)={±2ei,±ei±ej},\Phi(B_n) = \{\pm e_i, \pm e_i \pm e_j\} \quad\text{and} \quad\Phi(C_n) = \{\pm2e_i, \pm e_i \pm e_j\},

whose root lattices are Zn\mathbb{Z}^n and DnD_n respectively. In both cases the Weyl group is the group of all signed permutations of the coordinates, which for CnC_n is larger than W(Dn)W(D_n). For G2G_2 we take, in the plane x1+x2+x3=0x_1+x_2+x_3=0 of R3\mathbb{R}^3, the short roots ei−eje_i-e_j and the long roots ±(2ei−ej−ek)\pm(2e_i-e_j-e_k) with {i,j,k}={1,2,3}\{i,j,k\}=\{1,2,3\}; the root lattice is A2A_2 and W(G2)W(G_2) is the dihedral group of order 12. For F4F_4 we take, in R4\mathbb{R}^4, the short roots ±ei±ej\pm e_i \pm e_j and the long roots ±2ei\pm2e_i and (±1,±1,±1,±1)(\pm1,\pm1,\pm1,\pm1); the root lattice is D4D_4 and W(F4)W(F_4) has order 1152. The root lattice alone therefore does not determine the problem considered below: CnC_n and DnD_n share the lattice DnD_n, and F4F_4 and D4D_4 share the lattice D4D_4, but with different Weyl groups.

The reflection hyperplanes Hα={x∈V:⟨x,α⟩=0}H_\alpha= \{x \in V : \langle x,\alpha\rangle= 0\}, for α∈Φ\alpha\in\Phi, partition VV into open Weyl chambers. Since the stabilizer of a vector v∈Vv \in V is generated by the reflections it contains [1 Theorem 1.12], a lattice vector v∈Lv \in L has trivial stabilizer in WW if and only if vv lies in the interior of some Weyl chamber, equivalently, ⟨v,α⟩≠0\langle v,\alpha\rangle\ne0 for all roots α∈Φ\alpha\in\Phi. Such vectors are called regular.

We consider the following optimization problem.

Definition 1. For a root system Φ\Phi, define

Triv⁡(Φ)=min⁡{∥v∥2:v∈ZΦ, Stab⁡W(Φ)(v)={1}}.\operatorname{Triv}(\Phi) = \min\left\{\lVert v\rVert^2 : v \in\mathbb{Z}\Phi,\ \operatorname{Stab}_{W(\Phi)}(v) = \{1\}\right\}.

By symmetry, we may restrict the search to a single closed Weyl chamber. Fixing a system of simple roots α1,…,αn\alpha_1,\ldots,\alpha_n, the dominant chamber is C={x:⟨x,αi⟩≥0C=\{x:\langle x,\alpha_i\rangle\ge0 for all i}i\}, and we seek a lattice point in its interior with minimum norm. In the realizations above the inner product takes integral values on LL, so that this is

Triv⁡(Φ)=min⁡{vGvT:v∈Zn, vGαiT≥1 for all i}.(1)\operatorname{Triv}(\Phi)=\min\left\{vGv^{T}:v\in\mathbb{Z}^{n},\ vG\alpha_i^{T}\ge1\text{ for all }i\right\}. \tag*{(1)}

where GG is the Gram matrix of the simple roots and vv is expressed in the basis of simple roots.

This is an integer quadratic programming problem that can be solved computationally, and a priori nothing guarantees that its optimal solution is unique. Our main result, Theorem 4, shows that it always is: the optimal vector in the dominant chamber is ρ+λ\rho+\lambda, where ρ\rho is the Weyl vector and λ\lambda is the minimal dominant weight in the class of ρ\rho modulo LL, which is 0 or a minuscule weight. For the families AnA_n, BnB_n, CnC_n and DnD_n, we also give elementary proofs of closed-form solutions (Theorems 1 and 2). For the exceptional types, the solutions were first obtained computationally and are reported in § 1.1. In § 3 we compare the optimal vectors with the Weyl vector and prove Theorem 4, and in § 4 we explain the equalities between the minimal norms of different root systems by folding. The results are summarized in Table 1.

Φ\PhiTriv⁡(Φ)\operatorname{Triv}(\Phi)λ\lambdaReference
AnA_n, nn evenk(k+1)(2k+1)3\frac{k(k+1)(2k+1)}{3}, k=n2k=\frac{n}{2}00Theorem 1
AnA_n, nn oddk(k+1)(2k+1)3\frac{k(k+1)(2k+1)}{3}, k=n+12k=\frac{n+1}{2}ωk\omega_kTheorem 1
BnB_nn(n+1)(2n+1)6\frac{n(n+1)(2n+1)}{6}ωn\omega_nTheorem 2
CnC_n, n≡0,3(mod4)n\equiv0,3\pmod{4}n(n+1)(2n+1)6\frac{n(n+1)(2n+1)}{6}00Theorem 2
CnC_n, n≡1,2(mod4)n\equiv1,2\pmod{4}n(n+1)(2n+1)6+2n+1\frac{n(n+1)(2n+1)}{6}+2n+1ω1\omega_1Theorem 2
DnD_n, n≡0,1(mod4)n\equiv0,1\pmod{4}(n−1)n(2n−1)6\frac{(n-1)n(2n-1)}{6}00Theorem 2
DnD_n, n≡2,3(mod4)n\equiv2,3\pmod{4}(n−1)n(2n−1)6+2n−1\frac{(n-1)n(2n-1)}{6}+2n-1ω1\omega_1Theorem 2
E6E_6787800Figure 1
E7E_7228228ω7\omega_7Figure 1
E8E_862062000Figure 1
F4F_4787800Figure 1
G2G_2141400Figure 1

Table 1. The minimal norm Triv⁡(Φ)\operatorname{Triv}(\Phi) for each irreducible root system Φ\Phi, the weight λ\lambda such that the optimal vector in the dominant chamber is ρ+λ\rho+\lambda (Theorem 4), and where the optimal vector is described. The fundamental weights are numbered as in Bourbaki; λ=0\lambda=0 means that the optimal vector is the Weyl vector ρ\rho.

Remark 1. The problem reduces to irreducible root systems. Let Φ=Φ1∪⋯∪Φr\Phi=\Phi_1\cup\cdots\cup\Phi_r be the decomposition of Φ\Phi into irreducible components, which are pairwise orthogonal. Then L=L1⊕⋯⊕LrL=L_1\oplus\cdots\oplus L_r with Lj=ZΦjL_j=\mathbb{Z}\Phi_j, and W=W(Φ1)×⋯×W(Φr)W=W(\Phi_1)\times\cdots\times W(\Phi_r), each factor acting on the span of its own component and trivially on the others. For v=v= v1+⋯+vrv_1+\cdots+v_r with vj∈Ljv_j \in L_j, we have Stab⁡W(v)=Stab⁡W(Φ1)(v1)×⋯×Stab⁡W(Φr)(vr)\operatorname{Stab}_W(v)=\operatorname{Stab}_{W(\Phi_1)}(v_1)\times\cdots\times\operatorname{Stab}_{W(\Phi_r)}(v_r) and ∥v∥2=∑j∥vj∥2\lVert v\rVert^2=\sum_j\lVert v_j\rVert^2. Hence

Triv⁡(Φ)=Triv⁡(Φ1)+⋯+Triv⁡(Φr),\operatorname{Triv}(\Phi)=\operatorname{Triv}(\Phi_1)+\cdots+\operatorname{Triv}(\Phi_r),

and the optimal vectors of Φ\Phi are the sums of optimal vectors of the components. In particular, the optimal solution is again unique up to the action of WW, and the optimal vector in the dominant chamber is ρ+λ\rho+\lambda, where ρ\rho and λ\lambda are the sums of the corresponding vectors of the components. In the rest of the paper we assume that Φ\Phi is irreducible.

Exceptional root systems

For E6E_6, E7E_7 and E8E_8, we solve the integer quadratic programming problem (1) computationally using the SCIP optimization suite [6]. For F4F_4 and G2G_2, we enumerate all lattice vectors of small norm. The optimal vectors are given in Figure 1, together with the one for D4D_4, which is the smallest case of Theorem 2.

Figure 1. Vectors realizing Triv⁡(Φ)\operatorname{Triv}(\Phi) in the dominant chamber for the types D4D_4, E6E_6, E7E_7, E8E_8, F4F_4, G2G_2. Each node of the Dynkin diagram carries the coefficient of the vector on the corresponding simple root; for F4F_4 and G2G_2 the arrow points from the long to the short simple roots. The inner product of the vector with each simple root α\alpha is ⟨α,α⟩/2\langle\alpha,\alpha\rangle/2, except for the node of E7E_7 with coefficient 1515, where it is 22.

The E8E_8 result is confirmed by Moorhouse [4], who reports: “The 310th shell is the smallest containing a regular W(E8)W(E_8)-orbit; i.e. the shortest lattice vectors having trivial stabilizer in W(E8)W(E_8) have norm 620.”

In the realizations of the introduction, the optimal vector in the dominant chamber is (3,−1,−2)(3,-1,-2) for G2G_2 and (8,3,2,1)(8,3,2,1) for F4F_4, for the simple roots e1−2e2+e3e_1-2e_2+e_3, e2−e3e_2-e_3 of G2G_2 and (1,−1,−1,−1)(1,-1,-1,-1), 2e42e_4, e3−e4e_3-e_4, e2−e3e_2-e_3 of F4F_4, in the order of Figure 1.

For D4D_4, E6E_6, E8E_8, F4F_4 and G2G_2 these vectors are the Weyl vectors, which are optimal and unique by Remark 4, while for E7E_7 the optimal vector is not the Weyl vector. In all cases, the optimal vector is unique and is given by Theorem 4.

Remark 2. We have Triv⁡(G2)=Triv⁡(D4)\operatorname{Triv}(G_2)=\operatorname{Triv}(D_4) and Triv⁡(F4)=Triv⁡(E6)\operatorname{Triv}(F_4)=\operatorname{Triv}(E_6), and in Figure 1 the coefficients for G2G_2 and F4F_4 are those of the nodes of D4D_4 and E6E_6 that are identified by the folding of their Dynkin diagrams. This is explained in § 4.

The AnA_n case

The root lattice AnA_n can be represented as the set of integer vectors (a1,…,an+1)∈Zn+1(a_1,\ldots,a_{n+1})\in\mathbb{Z}^{n+1} with ∑i=1n+1ai=0\sum_{i=1}^{n+1}a_i=0. The Weyl group W(An)≅Sn+1W(A_n)\cong S_{n+1} acts by permuting coordinates. A vector v=(a1,…,an+1)v=(a_1,\ldots,a_{n+1}) has trivial stabilizer if and only if ai≠aja_i\ne a_j for all i≠ji\ne j. The norm is ∥v∥2=∑i=1n+1ai2\lVert v\rVert^2=\sum_{i=1}^{n+1}a_i^2.

Thus Triv⁡(An)\operatorname{Triv}(A_n) reduces to minimizing ∑i=1Nai2\sum_{i=1}^{N}a_i^2 over N=n+1N=n+1 distinct integers summing to zero. Setting k=⌊N/2⌋=⌊(n+1)/2⌋k=\lfloor N/2\rfloor=\lfloor(n+1)/2\rfloor, we prove:

Theorem 1. The problem Triv⁡(An)\operatorname{Triv}(A_n) has a unique optimal solution (up to the action of W(An)W(A_n)). In the ambient representation, the optimal coordinates are:

  1. For nn even (N=n+1N=n+1 odd, k=n/2k=n/2): the set {−k,…,−1,0,1,…,k}\{-k,\ldots,-1,0,1,\ldots,k\}.

  2. For nn odd (N=n+1N=n+1 even, k=(n+1)/2k=(n+1)/2): the set {−k,…,−1,1,…,k}\{-k,\ldots,-1,1,\ldots,k\}.

In both cases, Triv⁡(An)=k(k+1)(2k+1)3\operatorname{Triv}(A_n)=\dfrac{k(k+1)(2k+1)}{3}.

Expressed in the basis of simple roots αi=ei−ei+1\alpha_i=e_i-e_{i+1}, the optimal vector in the dominant Weyl chamber has coefficients:

  1. For nn even: ci=i(n+1−i)/2c_i=i(n+1-i)/2 for 1≤i≤n1\le i\le n.

  2. For nn odd: ci=j(n+2−j)/2c_i=j(n+2-j)/2 where j=min⁡(i,n+1−i)j=\min(i,n+1-i), for 1≤i≤n1\le i\le n.

The proof relies on two exchange lemmas that constrain the structure of optimal solutions.

Lemma 1 (No large gaps). Let S={a1<⋯<aN}S=\{a_1<\cdots<a_N\} be a set of NN distinct integers with ∑ai=0\sum a_i=0 that minimizes ∑ai2\sum a_i^2. Then ai+1−ai≤2a_{i+1}-a_i\leq2 for all 1≤i≤N−11\leq i\leq N-1.

Proof. Suppose ai+1≥ai+3a_{i+1}\geq a_i+3 for some ii. Replace aia_i by ai+1a_i+1 and ai+1a_{i+1} by ai+1−1a_{i+1}-1. Both new values lie strictly between aia_i and ai+1a_{i+1}, so they are distinct from each other and from all other elements of SS. The sum is preserved, and the change in ∑ai2\sum a_i^2 is

(ai+1)2+(ai+1−1)2−ai2−ai+12=2(ai−ai+1+1)≤2(−3+1)=−4<0,(a_i+1)^2+(a_{i+1}-1)^2-a_i^2-a_{i+1}^2=2(a_i-a_{i+1}+1)\leq2(-3+1)=-4<0,

contradicting optimality. ▢

Lemma 2 (At most one gap). Under the same hypotheses as Lemma 1, there is at most one index ii with ai+1−ai=2a_{i+1}-a_i=2.

Proof. Suppose ai+1=ai+2a_{i+1}=a_i+2 and aj+1=aj+2a_{j+1}=a_j+2 for some i<ji<j. The integers ai+1a_i+1 and aj+1a_j+1 are absent from SS. Replace aia_i by ai+1a_i+1 and aj+1a_j+1 by aj+1a_j+1. The two new values are distinct from each other (since ai<aja_i<a_j implies ai+1<aj+1a_i+1<a_j+1) and from all remaining elements of SS. The sum is preserved since (ai+1)+(aj+1)−ai−(aj+2)=0(a_i+1)+(a_j+1)-a_i-(a_j+2)=0. The change in ∑ai2\sum a_i^2 is

(ai+1)2+(aj+1)2−ai2−(aj+2)2=2(ai−aj−1)<0,(a_i+1)^2+(a_j+1)^2-a_i^2-(a_j+2)^2=2(a_i-a_j-1)<0,

contradicting optimality. □

Proof of Theorem 1. Let S={a1<⋯<aN}S=\{a_1<\cdots<a_N\} be an optimal set. By Lemmas 1 and 2, the set SS is either NN consecutive integers or N+1N+1 consecutive integers with one element removed.

If SS has no gap, then S={a1,a1+1,…,a1+N−1}S=\{a_1,a_1+1,\ldots,a_1+N-1\} and ∑ai=Na1+N(N−1)/2=0\sum a_i=Na_1+N(N-1)/2=0, giving a1=−(N−1)/2a_1=-(N-1)/2. This is an integer if and only if NN is odd, i.e., nn is even. In that case k=n/2k=n/2 and S={−k,…,k}S=\{-k,\ldots,k\}.

If SS has exactly one gap, then S={a1,…,a1+N}∖{g}S=\{a_1,\ldots,a_1+N\}\setminus\{g\} for some integer g∈[a1,a1+N]g\in[a_1,a_1+N]. The sum condition gives

g=(N+1)a1+N(N+1)2.g=(N+1)a_1+\frac{N(N+1)}{2}.

The constraint a1≤g≤a1+Na_1\leq g\leq a_1+N forces −N+12≤a1≤−N−12-\frac{N+1}{2}\leq a_1\leq-\frac{N-1}{2}.

  • If N=2kN=2k (i.e., nn odd): The only integer in that range is a1=−ka_1=-k, giving g=0g=0. Thus S={−k,…,−1,1,…,k}S=\{-k,\ldots,-1,1,\ldots,k\}.

  • If N=2k+1N=2k+1 (i.e., nn even): Then a1∈{−k−1,−k}a_1\in\{-k-1,-k\}. For a1=−ka_1=-k, we get g=k+1=a1+Ng=k+1=a_1+N, so the gap is at the top endpoint. For a1=−k−1a_1=-k-1, we get g=−k−1=a1g=-k-1=a_1, so the gap is at the bottom endpoint. In both cases, removing an endpoint reduces to the case without gap: S={−k,…,k}S=\{-k,\ldots,k\}.

In all cases, the optimal norm is ∑ai2=2∑j=1kj2=k(k+1)(2k+1)3\sum a_i^2=2\sum_{j=1}^k j^2=\frac{k(k+1)(2k+1)}{3}.

The simple root coordinates ci=∑j=1iajc_i=\sum_{j=1}^i a_j (where the ambient coordinates are arranged in decreasing order for the dominant chamber) yield the stated formulas. □

The BnB_n, CnC_n and DnD_n cases

The three families are realized in Rn\mathbb{R}^n with standard basis e1,…,ene_1,\ldots,e_n:

  • BnB_n (n≥2n\geq2): roots ±ei\pm e_i, ±ei±ej\pm e_i\pm e_j; root lattice Zn\mathbb{Z}^n; Weyl group the signed permutations.

  • CnC_n (n≥2n\geq2): roots ±2ei\pm2e_i, ±ei±ej\pm e_i\pm e_j; root lattice Dn={x∈Zn:∑xi≡0(mod2)}D_n=\{x\in\mathbb{Z}^n:\sum x_i\equiv0\pmod2\}; Weyl group the signed permutations.

  • DnD_n (n≥4n\geq4): roots ±ei±ej\pm e_i\pm e_j; root lattice DnD_n; Weyl group the signed permutations with an even number of sign changes.

We take the simple roots αi=ei−ei+1\alpha_i=e_i-e_{i+1} for 1≤i≤n−11\leq i\leq n-1 together with αn=en\alpha_n=e_n for BnB_n, αn=2en\alpha_n=2e_n for CnC_n, and αn=en−1+en\alpha_n=e_{n-1}+e_n for DnD_n. The interior of the dominant chamber is x1>x2>⋯>xn>0x_1>x_2>\cdots>x_n>0 for BnB_n and CnC_n, and x1>x2>⋯>xn−1>∣xn∣x_1>x_2>\cdots>x_{n-1}>|x_n| for DnD_n. Writing x=∑iciαix=\sum_i c_i\alpha_i and si=x1+⋯+xis_i=x_1+\cdots+x_i, the coefficients on the simple roots are

ci=sifor all i (Bn),ci=sifor i≤n−1 and cn=sn2(Cn),(2)c_i=s_i\quad\text{for all }i\ (B_n),\qquad c_i=s_i\quad\text{for }i\leq n-1\text{ and }c_n=\frac{s_n}{2}\quad(C_n), \tag*{(2)}

and, for DnD_n, ci=sic_i=s_i for i≤n−2i\le n-2, cn−1=(sn−1−xn)/2c_{n-1}=(s_{n-1}-x_n)/2 and cn=sn/2c_n=s_n/2.

Lemma 3. A vector v=(x1,…,xn)\mathbf{v}=(x_1,\ldots,x_n) in the root lattice has trivial stabilizer (1) for BnB_n and CnC_n, if and only if ∣x1∣,…,∣xn∣|x_1|,\ldots,|x_n| are nonzero and pairwise distinct; (2) for DnD_n, if and only if ∣x1∣,…,∣xn∣|x_1|,\ldots,|x_n| are pairwise distinct.

Proof. The stabilizer is trivial if and only if ⟨v,α⟩≠0\langle v,\alpha\rangle\ne0 for every root α\alpha. We have ⟨v,ei±ej⟩=xi±xj\langle v,e_i\pm e_j\rangle=x_i\pm x_j, which is nonzero for both signs if and only if ∣xi∣≠∣xj∣|x_i|\ne|x_j|, and ⟨v,ei⟩,⟨v,2ei⟩\langle v,e_i\rangle,\langle v,2e_i\rangle are nonzero if and only if xi≠0x_i\ne0. □

Lemma 4. Let ϵ∈{0,1}\epsilon\in\{0,1\} and let ϵ≤c1<c2<⋯<cn\epsilon\le c_1<c_2<\cdots<c_n be integers. Then

∑i=1nci2≥∑i=1n(i−1+ϵ)2,\sum_{i=1}^{n}c_i^2\ge\sum_{i=1}^{n}(i-1+\epsilon)^2,

with equality if and only if ci=i−1+ϵc_i=i-1+\epsilon for all ii. If (c1,…,cn)≠(ϵ,…,n−1+ϵ)(c_1,\ldots,c_n)\ne(\epsilon,\ldots,n-1+\epsilon), then

∑i=1nci2≥∑i=1n−1(i−1+ϵ)2+(n+ϵ)2,\sum_{i=1}^{n}c_i^2\ge\sum_{i=1}^{n-1}(i-1+\epsilon)^2+(n+\epsilon)^2,

with equality if and only if ci=i−1+ϵc_i=i-1+\epsilon for i≤n−1i\le n-1 and cn=n+ϵc_n=n+\epsilon.

Proof. Since the cic_i are strictly increasing integers, ci≥c1+(i−1)≥i−1+ϵ≥0c_i\ge c_1+(i-1)\ge i-1+\epsilon\ge0, which gives the first inequality and its equality case. If cn=n−1+ϵc_n=n-1+\epsilon, then ci≤cn−(n−i)=i−1+ϵc_i\le c_n-(n-i)=i-1+\epsilon for all ii, so ci=i−1+ϵc_i=i-1+\epsilon for all ii. Hence if (ci)≠(ϵ,…,n−1+ϵ)(c_i)\ne(\epsilon,\ldots,n-1+\epsilon) we have cn≥n+ϵc_n\ge n+\epsilon, and the second inequality follows from ci≥i−1+ϵc_i\ge i-1+\epsilon for i≤n−1i\le n-1. □

Write S(m)=∑j=1mj2=m(m+1)(2m+1)/6S(m)=\sum_{j=1}^{m}j^2=m(m+1)(2m+1)/6.

Theorem 2. For each of BnB_n, CnC_n, DnD_n, the problem Triv has a unique optimal solution up to the action of the Weyl group. The optimal vector v\mathbf{v} in the dominant chamber, its coefficients cic_i on the simple roots, and the minimum norm are the following.

(1) BnB_n, all n≥2n\ge2: v=(n,n−1,…,2,1)\mathbf{v}=(n,n-1,\ldots,2,1), ci=i(2n+1−i)/2c_i=i(2n+1-i)/2 for 1≤i≤n1\le i\le n, and Triv⁡(Bn)=S(n)\operatorname{Triv}(B_n)=S(n).

(2) CnC_n, n≡0,3(mod4)n\equiv0,3\pmod{4}: v=(n,n−1,…,2,1)\mathbf{v}=(n,n-1,\ldots,2,1), ci=i(2n+1−i)/2c_i=i(2n+1-i)/2 for i≤n−1i\le n-1, cn=n(n+1)/4c_n=n(n+1)/4, and Triv⁡(Cn)=S(n)\operatorname{Triv}(C_n)=S(n).

(3) CnC_n, n≡1,2(mod4)n\equiv1,2\pmod{4}: v=(n+1,n−1,n−2,…,2,1)\mathbf{v}=(n+1,n-1,n-2,\ldots,2,1), ci=i(2n+1−i)/2+1c_i=i(2n+1-i)/2+1 for i≤n−1i\le n-1, cn=(n(n+1)+2)/4c_n=(n(n+1)+2)/4, and Triv⁡(Cn)=S(n)+2n+1\operatorname{Triv}(C_n)=S(n)+2n+1.

(4) DnD_n, n≡0,1(mod4)n\equiv0,1\pmod{4}: v=(n−1,n−2,…,1,0)\mathbf{v}=(n-1,n-2,\ldots,1,0), ci=i(2n−1−i)/2c_i=i(2n-1-i)/2 for i≤n−2i\le n-2, cn−1=cn=n(n−1)/4c_{n-1}=c_n=n(n-1)/4, and Triv⁡(Dn)=S(n−1)\operatorname{Triv}(D_n)=S(n-1).

(5) DnD_n, n≡2,3(mod4)n\equiv2,3\pmod{4}: v=(n,n−2,n−3,…,1,0)\mathbf{v}=(n,n-2,n-3,\ldots,1,0), ci=i(2n−1−i)/2+1c_i=i(2n-1-i)/2+1 for i≤n−2i\le n-2, cn−1=cn=(n(n−1)+2)/4c_{n-1}=c_n=(n(n-1)+2)/4, and Triv⁡(Dn)=S(n−1)+2n−1\operatorname{Triv}(D_n)=S(n-1)+2n-1.

Proof. Let v=(x1,…,xn)\mathbf{v}=(x_1,\ldots,x_n) have trivial stabilizer and let c1<⋯<cnc_1<\cdots<c_n be the absolute values ∣xi∣|x_i| in increasing order; they are pairwise distinct by Lemma 3. The norm of v\mathbf{v} is ∑ci2\sum c_i^2, and since changing the sign of xix_i changes ∑xi\sum x_i by the even number 2xi2x_i, we have ∑xi≡∑ci(mod2)\sum x_i\equiv\sum c_i\pmod{2}.

For BnB_n, the cic_i are positive, and Lemma 4 with ϵ=1\epsilon=1 shows that ∥v∥2≥S(n)\lVert\mathbf{v}\rVert^2\ge S(n), with equality if and only if {∣xi∣}={1,…,n}\{|x_i|\}=\{1,\ldots,n\}. All such vectors are signed permutations of (n,…,1)(n,\ldots,1), hence form a single orbit of W(Bn)W(B_n). For CnC_n, the cic_i are positive and ∑ci\sum c_i must be even. If n≡0,3(mod4)n \equiv0,3 \pmod{4}, then ∑i=1ni=n(n+1)/2\sum_{i=1}^{n} i = n(n+1)/2 is even, and we conclude as for BnB_n. If n≡1,2(mod4)n \equiv1,2 \pmod{4}, then n(n+1)/2n(n+1)/2 is odd, so (ci)≠(1,…,n)(c_i) \ne(1,\ldots,n) and the second part of Lemma 4 gives ∥v∥2≥S(n−1)+(n+1)2=S(n)+2n+1\lVert v\rVert^2 \ge S(n-1)+(n+1)^2 = S(n)+2n+1, with equality if and only if {∣xi∣}={1,…,n−1,n+1}\{|x_i|\}=\{1,\ldots,n-1,n+1\}. This set has even sum n(n+1)/2+1n(n+1)/2+1, so all the corresponding vectors lie in DnD_n, and they form a single orbit of W(Cn)W(C_n).

For DnD_n, at most one cic_i is zero, and we apply Lemma 4 with ϵ=0\epsilon=0. The sum ∑i=0n−1i=n(n−1)/2\sum_{i=0}^{n-1}i=n(n-1)/2 is even if and only if n≡0,1(mod4)n\equiv0,1\pmod{4}. In that case the optimum is S(n−1)S(n-1), realized exactly when {∣xi∣}={0,1,…,n−1}\{|x_i|\}=\{0,1,\ldots,n-1\}. Otherwise the second part of Lemma 4 gives the bound S(n−2)+n2=S(n−1)+2n−1S(n-2)+n^2=S(n-1)+2n-1, realized exactly when {∣xi∣}={0,1,…,n−2,n}\{|x_i|\}=\{0,1,\ldots,n-2,n\}, whose sum is even. In both cases one of the xix_i is zero, so every signed permutation of vv can be obtained with an even number of sign changes: the optimal vectors form a single orbit of W(Dn)W(D_n).

The vectors vv listed are the representatives of these orbits in the dominant chamber, and the coefficients cic_i follow from (2) and the partial sums ∑j=1i(n+1−j)=i(2n+1−i)/2\sum_{j=1}^{i}(n+1-j)=i(2n+1-i)/2 and ∑j=1i(n−j)=i(2n−1−i)/2\sum_{j=1}^{i}(n-j)=i(2n-1-i)/2.

Remark 3. The root systems B2B_2 and C2C_2 are the same up to a rotation and a scaling by 2\sqrt{2}, in agreement with Triv⁡(C2)=10=2Triv⁡(B2)\operatorname{Triv}(C_2)=10=2\operatorname{Triv}(B_2). Also, Triv⁡(Cn)=Triv⁡(Dn+1)\operatorname{Triv}(C_n)=\operatorname{Triv}(D_{n+1}) for all n≥3n\ge3, the optimal vector of Dn+1D_{n+1} being the optimal vector of CnC_n with a coordinate 00 appended, and Triv⁡(A2n−1)=Triv⁡(A2n)=2Triv⁡(Bn)\operatorname{Triv}(A_{2n-1})=\operatorname{Triv}(A_{2n})=2\operatorname{Triv}(B_n) for all n≥2n\ge2. These equalities are explained in § 4.

Comparison with the Weyl vector

The Weyl vector of Φ\Phi is

ρ=12∑α>0α\rho=\frac{1}{2}\sum_{\alpha>0}\alpha

It lies in the interior of the dominant chamber and is characterized by ⟨ρ,αi⟩=⟨αi,αi⟩/2\langle\rho,\alpha_i\rangle=\langle\alpha_i,\alpha_i\rangle/2 for every simple root αi\alpha_i, since the simple reflection sαis_{\alpha_i} permutes the positive roots other than αi\alpha_i [1 Proposition 1.4]. We compare it with the optimal vector vv in the dominant chamber given by Theorems 1 and 2 and Figure 1.

Theorem 3. Let Φ\Phi be one of AnA_n (n≥2n\ge2), BnB_n (n≥2n\ge2), CnC_n (n≥2n\ge2), DnD_n (n≥4n\ge4), E6E_6, E7E_7, E8E_8, F4F_4, G2G_2. Then v=ρv=\rho if and only if ρ∈L\rho\in L, and otherwise vv is not a multiple of ρ\rho. We have ρ∈L\rho\in L exactly for AnA_n with nn even, CnC_n with n≡0,3(mod4)n\equiv0,3\pmod{4}, DnD_n with n≡0,1(mod4)n\equiv0,1\pmod{4}, E6E_6, E8E_8, F4F_4 and G2G_2.

Proof. We compare the two vectors case by case.

  • AnA_n: ρ=12(n,n−2,…,−n)\rho=\frac{1}{2}(n,n-2,\ldots,-n). For nn even, ρ=(k,…,−k)=v\rho=(k,\ldots,-k)=v. For nn odd, ρ∉Zn+1\rho\notin\mathbb{Z}^{n+1}, and v=(k,…,1,−1,…,−k)v=(k,\ldots,1,-1,\ldots,-k) is not a multiple of ρ\rho since its consecutive differences are not all equal.

  • BnB_n: ρ=(n−12,…,32,12)∉Zn\rho=(n-\frac{1}{2},\ldots,\frac{3}{2},\frac{1}{2})\notin\mathbb{Z}^n, and v=(n,…,1)v=(n,\ldots,1) is not a multiple of ρ\rho.

  • CnC_n: ρ=(n,…,1)\rho=(n,\ldots,1) lies in LL if and only if n(n+1)/2n(n+1)/2 is even, that is n≡0,3(mod4)n\equiv0,3\pmod{4}, and then ρ=v\rho=v. Otherwise v=(n+1,n−1,…,1)v=(n+1,n-1,\ldots,1) is not a multiple of ρ\rho.

  • DnD_n: ρ=(n−1,…,1,0)\rho=(n-1,\ldots,1,0) lies in DnD_n if and only if n(n−1)/2n(n-1)/2 is even, that is n≡0,1(mod4)n\equiv0,1\pmod{4}, and then ρ=v\rho=v. Otherwise v=(n,n−2,…,1,0)v=(n,n-2,\ldots,1,0) is not a multiple of $\rho.

  • E6_6, E8_8, F4_4, G2_2: by Figure 1, ⟨v,αi⟩=⟨αi,αi⟩/2\langle v,\alpha_i\rangle=\langle\alpha_i,\alpha_i\rangle/2 for every simple root, so v=ρv=\rho, and ρ∈L\rho\in L.

  • E7_7: by Figure 1, ⟨v,αi⟩\langle v,\alpha_i\rangle is 1 for all simple roots but one, where it is 2, so vv is not a multiple of ρ\rho. The coefficients of ρ\rho on the simple roots are 17, 33, 48, 752\frac{75}{2}, 26, 272\frac{27}{2} along the long branch and 492\frac{49}{2} on the short one, so ρ∉\rho\notin E7_7. □

Remark 4. If ρ∈L\rho\in L, then ρ\rho is optimal and unique, without computation. For x∈Lx\in L, ⟨x,αi⟩=⟨ρ,αi⟩⟨x,αi∨⟩\langle x,\alpha_i\rangle=\langle\rho,\alpha_i\rangle\langle x,\alpha_i^\vee\rangle with ⟨x,αi∨⟩∈Z\langle x,\alpha_i^\vee\rangle\in\mathbb{Z}, so if v∈Lv\in L is in the open dominant chamber, then w=v−ρw=v-\rho is dominant. Since ρ\rho is a sum of positive roots, ⟨ρ,w⟩≥0\langle\rho,w\rangle\geq0 and ∥v∥2=∥ρ∥2+2⟨ρ,w⟩+∥w∥2≥∥ρ∥2\|v\|^2=\|\rho\|^2+2\langle\rho,w\rangle+\|w\|^2\geq\|\rho\|^2, with equality if and only if v=ρv=\rho. This proves the results for E6_6, E8_8, F4_4, G2_2.

The argument of Remark 4 extends to the case ρ∉L\rho\notin L, and gives a uniform description of the optimal vector, together with its uniqueness. Let P={x∈V:⟨x,α∨⟩∈Z for all α∈Φ}P=\{x\in V:\langle x,\alpha^\vee\rangle\in\mathbb{Z}\text{ for all }\alpha\in\Phi\} be the weight lattice [1 §2.9], with fundamental weights ω1,…,ωn\omega_1,\ldots,\omega_n, and let P+P^+ be the set of dominant weights. We write μ⪯λ\mu\preceq\lambda if λ−μ\lambda-\mu is a nonnegative integral combination of simple roots. A dominant weight λ\lambda is minuscule if it is nonzero and ⟨λ,α∨⟩∈{0,±1}\langle\lambda,\alpha^\vee\rangle\in\{0,\pm1\} for all α∈Φ\alpha\in\Phi. By Stembridge [5, Corollary 1.4, Proposition 1.12 and Corollary 1.13], each coset μ+L\mu+L of LL in PP contains a unique dominant weight λ\lambda with λ⪯λ′\lambda\preceq\lambda' for every dominant weight λ′∈μ+L\lambda'\in\mu+L, and this weight is 0 for the coset LL and the unique minuscule weight of the coset otherwise.

Theorem 4. Let λΦ\lambda_\Phi be the minimal dominant weight of the coset ρ+L\rho+L. Then v=ρ+λΦv=\rho+\lambda_\Phi is the unique vector of LL of minimum norm in the open dominant chamber, and

Triv⁡(Φ)=∥ρ+λΦ∥2.\operatorname{Triv}(\Phi)=\|\rho+\lambda_\Phi\|^2.

In particular, the optimal solution is unique up to the action of WW.

Proof. Let x∈Lx\in L be in the open dominant chamber. Since L⊆PL\subseteq P, the integers ⟨x,αi∨⟩\langle x,\alpha_i^\vee\rangle are at least 1, and since ⟨ρ,αi∨⟩=1\langle\rho,\alpha_i^\vee\rangle=1, the vector w=x−ρw=x-\rho is a dominant weight. As 2ρ2\rho is the sum of the positive roots, w∈−ρ+L=ρ+Lw\in-\rho+L=\rho+L. Hence w=λΦ+qw=\lambda_\Phi+q with qq a nonnegative integral combination of simple roots,

∥x∥2=∥ρ+λΦ∥2+2⟨ρ+λΦ,q⟩+∥q∥2.\|x\|^2=\|\rho+\lambda_\Phi\|^2+2\langle\rho+\lambda_\Phi,q\rangle+\|q\|^2.

Since ρ+λΦ\rho+\lambda_\Phi is in the open dominant chamber, ⟨ρ+λΦ,q⟩>0\langle\rho+\lambda_\Phi,q\rangle>0 for q≠0q\ne0. So ∥x∥2≥∥ρ+λΦ∥2\|x\|^2\geq\|\rho+\lambda_\Phi\|^2, with equality if and only if x=ρ+λΦx=\rho+\lambda_\Phi, and ρ+λΦ\rho+\lambda_\Phi is in LL and in the open dominant chamber. □

Remark 4 is the case λΦ=0\lambda_\Phi=0, that is ρ∈L\rho\in L. In the cases where ρ∉L\rho\notin L, the weight λΦ\lambda_\Phi and the optimal vector of Theorems 1 and 2 are recovered as follows:

  • A2k−1_{2k-1}: λΦ=ωk=12(1,…,1,−1,…,−1)\lambda_\Phi=\omega_k=\frac{1}{2}(1,\ldots,1,-1,\ldots,-1) and v=(k,…,1,−1,…,−k)v=(k,\ldots,1,-1,\ldots,-k);

  • Bn_n: λΦ=ωn=12(1,…,1)\lambda_\Phi=\omega_n=\frac{1}{2}(1,\ldots,1) and v=(n,…,1)v=(n,\ldots,1);

  • Cn_n, n=1,2(mod4)n=1,2\pmod4: λΦ=ω1=e1\lambda_\Phi=\omega_1=e_1 and v=(n+1,n−1,…,1)v=(n+1,n-1,\ldots,1);

  • Dn_n, n=2,3(mod4)n=2,3\pmod4: λΦ=ω1=e1\lambda_\Phi=\omega_1=e_1 and v=(n,n−2,…,1,0)v=(n,n-2,\ldots,1,0);

  • E7_7: λΦ\lambda_\Phi is the minuscule weight ω7\omega_7, dual to the simple root at the end of the long branch, which is the node with coefficient 15 in Figure 1. With ∥ρ∥2=3992\|\rho\|^2=\frac{399}{2}, ⟨ρ,ω7⟩=272\langle\rho,\omega_7\rangle=\frac{27}{2} and ∥ω7∥2=32\|\omega_7\|^2=\frac{3}{2}, we get Triv⁡(E7)=3992+27+32=228\operatorname{Triv}(\mathrm{E}_7)=\frac{399}{2}+27+\frac{3}{2}=228.

In particular, the results for the exceptional types do not depend on the computations of Section 2, which serve as a confirmation.

Folding

The equalities between minimal norms observed in § 1.1 and § 2, namely

Triv⁡(G2)=Triv⁡(D4),Triv⁡(F4)=Triv⁡(E6),Triv⁡(Cn)=Triv⁡(Dn+1),\operatorname{Triv}(G_2)=\operatorname{Triv}(D_4),\qquad\operatorname{Triv}(F_4)=\operatorname{Triv}(E_6),\qquad\operatorname{Triv}(C_n)=\operatorname{Triv}(D_{n+1}),
2Triv⁡(Bn)=Triv⁡(A2n−1)=Triv⁡(A2n),2\operatorname{Triv}(B_n)=\operatorname{Triv}(A_{2n-1})=\operatorname{Triv}(A_{2n}),

come from the automorphisms of the Dynkin diagrams of the simply laced root systems. These diagram automorphisms and the corresponding folding of Dynkin diagrams are at the basis of the construction of the twisted affine Lie algebras [2 §8.3].

Let Φ\Phi be a simply laced root system with simple roots Δ={α1,…,αn}\Delta=\{\alpha_1,\ldots,\alpha_n\} and let σ\sigma be a nontrivial automorphism of its Dynkin diagram. Since the Gram matrix of Δ\Delta is determined by the Dynkin diagram, σ\sigma extends to an isometry of VV, still denoted σ\sigma, which permutes Δ\Delta. It therefore preserves Φ\Phi, the root lattice LL and the dominant chamber CC. Let VσV^\sigma be the subspace of vectors fixed by σ\sigma and Lσ=L∩VσL^\sigma=L\cap V^\sigma. For a σ\sigma-orbit O⊆ΔO\subseteq\Delta, set αO=∑β∈Oβ\alpha_O=\sum_{\beta\in O}\beta. Since σ\sigma permutes the basis Δ\Delta of LL, the vectors αO\alpha_O form a basis of LσL^\sigma. Two roots in the same orbit are orthogonal, except for the middle orbit {αn,αn+1}\{\alpha_n,\alpha_{n+1}\} of A2nA_{2n}, where they are adjacent; so ⟨αO,αO⟩=2∣O∣\langle\alpha_O,\alpha_O\rangle=2|O|, except for that orbit, where it is 2. A direct computation of the Gram matrix of the αO\alpha_O shows that they form a system of simple roots of a root system Φσ\Phi^\sigma in VσV^\sigma, whose type and root lattice Lσ=ZΦσL^\sigma=\mathbb{Z}\Phi^\sigma are given in the following table.

Φ\Phiorder of σ\sigmaΦσ\Phi^\sigma∥αO∥2\lVert\alpha_O\rVert^2LσL^\sigma
A2n−1A_{2n-1}, n≥2n\ge222Bn\sqrt{2}B_n2, 42Zn\sqrt{2}\mathbb{Z}^n
A2nA_{2n}, n≥2n\ge222Bn\sqrt{2}B_n2, 42Zn\sqrt{2}\mathbb{Z}^n
Dn+1D_{n+1}, n≥3n\ge32CnC_n2, 4DnD_n
D4D_43G2G_22, 6A2A_2
E6E_62F4F_42, 4D4D_4

Table 2.

Since a root system is determined up to isometry by the Gram matrix of its simple roots, Φσ\Phi^\sigma is isometric to the realization of the introduction, except for BnB_n, which appears scaled by 2\sqrt{2}, so that Triv⁡(2Bn)=2Triv⁡(Bn)\operatorname{Triv}(\sqrt{2}B_n)=2\operatorname{Triv}(B_n). For A2n−1A_{2n-1} and Dn+1D_{n+1}, the subalgebra of the simple Lie algebra of type Φ\Phi fixed by the automorphism induced by σ\sigma has type CnC_n and BnB_n respectively [2 Proposition 8.3 e)], the dual of Φσ\Phi^\sigma. A root system and its dual have the same reflection hyperplanes, hence the same regular vectors; it is the lattice LσL^\sigma that selects Φσ\Phi^\sigma among the two.

Proposition 1. Let Φ\Phi and σ\sigma be as above. Then Triv⁡(Φσ)≥Triv⁡(Φ)\operatorname{Triv}(\Phi^\sigma)\ge\operatorname{Triv}(\Phi), with equality if and only if some vector of LL in the open dominant chamber of norm Triv⁡(Φ)\operatorname{Triv}(\Phi) is fixed by σ\sigma. If the optimal solution for Φ\Phi is unique up to the action of W(Φ)W(\Phi), then equality holds, the optimal solution for Φσ\Phi^\sigma is unique up to the action of W(Φσ)W(\Phi^\sigma), and the optimal vector vv in the dominant chamber is the same for Φ\Phi and Φσ\Phi^\sigma. Its coefficient on αO\alpha_O is the common coefficient of vv on the roots of OO.

Proof. For x∈Vσx\in V^\sigma and β∈O\beta\in O, we have ⟨x,β⟩=⟨x,σ(β)⟩\langle x,\beta\rangle=\langle x,\sigma(\beta)\rangle, so ⟨x,β⟩\langle x,\beta\rangle is constant on OO and ⟨x,αO⟩=∣O∣⟨x,β⟩\langle x,\alpha_O\rangle=|O|\langle x,\beta\rangle. Hence the open dominant chamber of Φσ\Phi^\sigma, defined by ⟨x,αO⟩>0\langle x,\alpha_O\rangle>0 for all orbits OO, is C∘∩VσC^\circ\cap V^\sigma, where C∘C^\circ is the open dominant chamber of Φ\Phi. Since Triv⁡\operatorname{Triv} can be computed in the open dominant chamber,

Triv⁡(Φσ)=min⁡{∥x∥2:x∈Lσ∩C∘}≥min⁡{∥x∥2:x∈L∩C∘}=Triv⁡(Φ),\operatorname{Triv}(\Phi^\sigma)=\min\{\lVert x\rVert^2:x\in L^\sigma\cap C^\circ\}\ge\min\{\lVert x\rVert^2:x\in L\cap C^\circ\}=\operatorname{Triv}(\Phi),

with equality if and only if a minimizer of the right-hand side lies in VσV^\sigma. Every W(Φ)W(\Phi)-orbit meets C∘C^\circ in at most one point, so if the optimal solution for Φ\Phi is unique up to W(Φ)W(\Phi), there is a unique minimizer v∈L∩C∘v \in L \cap C^\circ. Since σ\sigma preserves LL, C∘C^\circ and the norm, σ(v)\sigma(v) is also a minimizer, so σ(v)=v\sigma(v)=v and equality holds. Any minimizer for Φσ\Phi^\sigma in Lσ∩C∘L^\sigma\cap C^\circ is then a minimizer for Φ\Phi, hence equal to vv. Finally, writing v=∑iciαiv=\sum_i c_i\alpha_i, the relation σ(v)=v\sigma(v)=v shows that cic_i is constant on each orbit OO, with value cOc_O say, and v=∑OcOαOv=\sum_O c_O\alpha_O.

By Theorem 4, or by Theorems 1 and 2 and Remark 4, the optimal solutions for AnA_n, DnD_n and E6E_6 are unique, so Proposition 1 gives the equalities above. It also gives a second proof of the values of Triv⁡\operatorname{Triv} for BnB_n, CnC_n, F4F_4 and G2G_2, and of the uniqueness of their optimal solutions. The last assertion of Proposition 1 explains the coefficients in Figure 1, and the fact that the coefficients ci=i(2n+1−i)/2c_i=i(2n+1-i)/2 of the optimal vector of BnB_n are those of A2n−1A_{2n-1} and A2nA_{2n} for i≤ni\le n in Theorem 1. For Dn+1D_{n+1}, the space VσV^\sigma is the hyperplane xn+1=0x_{n+1}=0, which explains why the optimal vector of Dn+1D_{n+1} is that of CnC_n with a coordinate 0 appended. Note that D4D_4 admits diagram automorphisms of order 2 and 3, giving Triv⁡(C3)=Triv⁡(G2)=Triv⁡(D4)=14\operatorname{Triv}(C_3)=\operatorname{Triv}(G_2)=\operatorname{Triv}(D_4)=14, while E7E_7 and E8E_8 have no nontrivial diagram automorphism.

Tool and computational resource disclosure

Following the Leiden Declaration on Artificial Intelligence and Mathematics [3], we disclose the automated tools used in this work. The optimal vectors for the exceptional types were computed with programs written in GAP, using the SCIP optimization suite [6] for E6E_6, E7E_7 and E8E_8 and the enumeration of short lattice vectors for F4F_4 and G2G_2. The AI coding assistant Claude Code (Anthropic), with the large language model Claude Opus 5.5, was used to draft and edit parts of the text and of the proofs, including the treatment of the types BnB_n, CnC_n, F4F_4 and G2G_2, the comparison with the Weyl vector, Theorem 4 and § 4, to search the literature, and to check the results numerically by independent enumerations. All statements and proofs were checked by the author, who takes full responsibility for the content of the paper.

MATHIEU DUTOUR SIKIRIĆ, MSM PROGRAMMING, KARLOVAČKA CESTA 28B, 10452 KLINCA SELO

Email address: [email protected]

References

  1. [1]J. E. Humphreys, Reflection groups and Coxeter groups, Cambridge Studies in Advanced Mathematics 29, Cambridge University Press, Cambridge, 1990.DOI
  2. [2]V. G. Kac, Infinite-dimensional Lie algebras, 3rd edition, Cambridge University Press, Cambridge, 1994.
  3. [3]Leiden Declaration on Artificial Intelligence and Mathematics, 2 June 2026, https://leidendeclaration.ai/, doi:10.5281/zenodo.20302944.DOI
  4. [4]E. Moorhouse, E₈, http://ericmoorhouse.org/e8/.
  5. [6]S. Bolusani, M. Besançon, K. Bestuzheva, et al., The SCIP Optimization Suite 9.0, Technical Report, Optimization Online, 2024. https://www.scipopt.org.arxiv.org/abs/2402.17702

Paper details

vlnc:W_5bam6xhv
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents