The polynomials \(F_{9,1}\) and \(F_{9,2}\) are constructed to be invariant under the action of \(\operatorname{Sp}(6)\) on the 14-dimensional irreducible representation \(V(0,0,1)\) whose highest weight is the fundamental dominant weight \(\omega_3\). We check this below as follows.
The symplectic group is generated by transformations known as symplectic transvections. See below for a definition and a reference.
We test that \(F_{9,1}\) and \(F_{9,2}\) are invariant under some transvections and their products. First, we test this for a transvection whose inputs are simple enough that the action can be checked by hand. Then, we test invariance under transvections with respect to random vectors.
Let \(B\) be a symplectic form on a vector space \(V\) of dimension \(2n\). That is, \(B\) is a nondegenerate alternating form.
Let \(c\) be a scalar, and let \(w\) be a vector in \(V\). A symplectic transvection is a transformation of the form \( \tau(v) = v+c B(v,w) w\) for some scalar \(c\) and some vector \(w \in V\). These generate the symplectic group. See for instance Artin, Geometric Algebra, Section III.5, especially formula (3.41), Definition 3.14, and Theorem 3.25.
We use the alternating form \(B\) corresponding to the matrix \[ \left[ \begin{array}{rr} 0 & I_n \\ -I_n & 0 \end{array}\right]. \]
First, consider the symplectic transformation \(g_1\) corresponding to the choices \(c = 1\) and \(w = e_1\). We have \(B(e_i,w)=0\) when \(i \neq 4\), and \(B(e_4,w) = -1\). Thus the matrix of \(g_1\) is \[ \left[ \begin{array}{rrrrrr} 1 & 0 & 0 & -1 & 0 & 0 \\ 0 & 1 & 0 & 0 & 0 & 0 \\ 0 & 0 & 1 & 0 & 0 & 0 \\ 0 & 0 & 0 & 1 & 0 & 0 \\ 0 & 0 & 0 & 0 & 1 & 0 \\ 0 & 0 & 0 & 0 & 0 & 1 \end{array} \right]. \] The basis we use for the 14-dimensional vector space \(V(0,0,1)\) is as follows. We write \(e_{i_{1}i_{2}i_{3}}\) for the wedge product \(e_{i_1} \wedge e_{i_2} \wedge e_{i_3}\). \[ \mathscr{B} = \{ e_{123}, e_{234},e_{134}-e_{235},-e_{124}-e_{236}, -e_{135},e_{125}-e_{136}, e_{126}, e_{156}, e_{146}-e_{256}, -e_{145}-e_{356}, -e_{246}, e_{245}-e_{346},e_{345},e_{456}\}. \] The action of \(g_1\) on this basis is as follows. \[ \begin{array}{rcl} g_1.e_{123} &= & e_{123}\\ g_1.e_{234} &= & -e_{123}+ e_{234}\\ g_1.(e_{134}-e_{235}) &= & e_{134}-e_{235}\\ g_1.(-e_{124}-e_{236}) &= & -e_{124}-e_{236}\\ g_1.(-e_{135}) &= & -e_{135}\\ g_1.(e_{125}-e_{136}) &= & e_{125}-e_{136}\\ g_1.e_{126} &= & e_{126}\\ g_1.e_{156} &= & e_{156}\\ g_1.(e_{146}-e_{256}) &= & e_{146}-e_{256}\\ g_1.(-e_{145}-e_{356}) &= & -e_{145}-e_{356}\\ g_1.(e_{245}-e_{346}) &= & e_{125}-e_{136}+e_{245}-e_{346}\\ g_1.e_{345} &= & e_{135}+e_{345}\\ g_1.e_{456} &= & -e_{156}+e_{456} \end{array} \] This matches the matrix rhog1 computed in the Macaulay2 session below. First, we load some files and and functions that will be useful.Macaulay2, version 1.25.11
Type "help" to see useful commands
i1 : load "MukaiModels.m2";
i2 : load "InvariantPolynomialsforMukaiModels.m2";
i3 : matrixFromHyperplaneEquations = (L) -> (
R:=ring(L_0);
eqnMatrix:=matrix apply(L, f -> apply(numgens R, i -> coefficient(R_i,f)));
transpose gens ker eqnMatrix
);
i4 : Q = FHsymplecticFormMatrix(3,QQ);
6 6
o4 : Matrix QQ <-- QQ
We enter \(c=1 \) and \(w=e_1\), and compute the element \(g_1 \in \operatorname{Sp}(6)\).
i5 : c1 = 1
o5 = 1
i6 : v1 = {1,0,0,0,0,0}
o6 = {1, 0, 0, 0, 0, 0}
o6 : List
i7 : g1 = symplecticTransvectionMatrix(Q,c1,v1)
o7 = | 1 0 0 -1 0 0 |
| 0 1 0 0 0 0 |
| 0 0 1 0 0 0 |
| 0 0 0 1 0 0 |
| 0 0 0 0 1 0 |
| 0 0 0 0 0 1 |
6 6
o7 : Matrix QQ <-- QQ
i8 : isInFultonHarrisSp2n(g1)
o8 = true
i9 : rhog1 = restrictToSubspace(exteriorPower(3,g1),MyBasisVw3C3)
o9 = | 1 -1 0 0 0 0 0 0 0 0 0 0 0 0 |
| 0 1 0 0 0 0 0 0 0 0 0 0 0 0 |
| 0 0 1 0 0 0 0 0 0 0 0 0 0 0 |
| 0 0 0 1 0 0 0 0 0 0 0 0 0 0 |
| 0 0 0 0 1 0 0 0 0 0 0 0 -1 0 |
| 0 0 0 0 0 1 0 0 0 0 0 1 0 0 |
| 0 0 0 0 0 0 1 0 0 0 -1 0 0 0 |
| 0 0 0 0 0 0 0 1 0 0 0 0 0 -1 |
| 0 0 0 0 0 0 0 0 1 0 0 0 0 0 |
| 0 0 0 0 0 0 0 0 0 1 0 0 0 0 |
| 0 0 0 0 0 0 0 0 0 0 1 0 0 0 |
| 0 0 0 0 0 0 0 0 0 0 0 1 0 0 |
| 0 0 0 0 0 0 0 0 0 0 0 0 1 0 |
| 0 0 0 0 0 0 0 0 0 0 0 0 0 1 |
14 14
o9 : Matrix QQ <-- QQ
We check that the symplectic Grassmannian is invariant under \(g_1\).
i10 : S = QQ[z_0..z_13];
i11 : I1 = {-z_5^2+z_4*z_6-z_0*z_7,
z_3*z_5-z_2*z_6-z_0*z_8,
-z_3*z_4+z_2*z_5-z_0*z_9,
-z_3^2+z_1*z_6-z_0*z_10,
z_2*z_3-z_1*z_5-z_0*z_11,
-z_2^2+z_1*z_4-z_0*z_12};
i12 : I2 = {-z_11^2+z_10*z_12-z_1*z_13,
z_9*z_11-z_8*z_12-z_2*z_13,
-z_9*z_10+z_8*z_11-z_3*z_13,
-z_9^2+z_7*z_12-z_4*z_13,
z_8*z_9-z_7*z_11-z_5*z_13,
-z_8^2+z_7*z_10-z_6*z_13};
i13 : I3 = {z_1*z_7+z_2*z_8+z_3*z_9-z_0*z_13,
z_1*z_8+z_2*z_10+z_3*z_11,
z_1*z_9+z_2*z_11+z_3*z_12,
z_2*z_7+z_4*z_8+z_5*z_9,
z_2*z_8+z_4*z_10+z_5*z_11-z_0*z_13,
z_2*z_9+z_4*z_11+z_5*z_12,
z_3*z_7+z_5*z_8+z_6*z_9,
z_3*z_8+z_5*z_10+z_6*z_11,
z_3*z_9+z_5*z_11+z_6*z_12-z_0*z_13};
i14 : SpGr36 = ideal(flatten {I1,I2,I3});
o14 : Ideal of S
i15 : F1 = map(S,S,transpose rhog1);
o15 : RingMap S <-- S
i16 : F1(SpGr36)==SpGr36
o16 = true
Next, we choose a random matrix \(M_{9,1} \in \operatorname{Mat}_{9 \times 14}\) and consider the point in \(\operatorname{Gr}(9,14)\) corresponding to its row space.
We compute the value of \(F_{9,1}\) on \(M_{9,1}\) and \(g_1.M_{9,1}\).
i17 :
M91 = matrix {{8, 10, 2, 0, 7, 6, 9, 5, -9, -10, -7, 6, 7, 3}, {1, 0, -8, 4, -4, -4, 6, -5, -2, 7, -4, 7, 9, 2}, {-3, 8, 1, -6, 3, -2, -5, -7, 9, 5, -9, -6, 5, -9}, {-8, -6, 9, 2, -10, 2, 4, 4, 0, 3, -8, -7, -10, -1}, {-1, 8, -3, -1, -9, 0, 10, 10, 8, -10, 4, 6, 4, -10}, {-8, 3, -6, 9, -4, 6, 6, 9, 5, 10, 2, 0, -7, 0}, {7, -3, 9, -2, 1, 4, -6, 7, 6, -6, 0, -10, -6, 8}, {-5, 3, 7, 9, -10, 2, 8, -7, 9, 3, -1, 2, -2, 10}, {-2, -4, 9, -7, 5, 10, -6, 1, 8, 4, 0, 10, 6, -7}}
o17 = | 8 10 2 0 7 6 9 5 -9 -10 -7 6 7 3 |
| 1 0 -8 4 -4 -4 6 -5 -2 7 -4 7 9 2 |
| -3 8 1 -6 3 -2 -5 -7 9 5 -9 -6 5 -9 |
| -8 -6 9 2 -10 2 4 4 0 3 -8 -7 -10 -1 |
| -1 8 -3 -1 -9 0 10 10 8 -10 4 6 4 -10 |
| -8 3 -6 9 -4 6 6 9 5 10 2 0 -7 0 |
| 7 -3 9 -2 1 4 -6 7 6 -6 0 -10 -6 8 |
| -5 3 7 9 -10 2 8 -7 9 3 -1 2 -2 10 |
| -2 -4 9 -7 5 10 -6 1 8 4 0 10 6 -7 |
9 14
o17 : Matrix ZZ <-- ZZ
i18 : F91(M91)
o18 = -69066628702801276747468144869007195496549388986880
o18 : QQ
i19 : F91(transpose(rhog1*transpose(M91)))
o19 = -69066628702801276747468144869007195496549388986880
o19 : QQ
Next, we construct some additional elements in \(
\operatorname{Sp}(6)\) by choosing random scalars and vectors and
using them to construct transvections.
i20 : c2 = 1/10;
i21 : v2 = {1, 8/9, 9/5, 9, 7/10, 7/10};
i22 : g2 = symplecticTransvectionMatrix(Q,c2,v2)
o22 = | 19/10 7/100 7/100 -1/10 -4/45 -9/50 |
| 4/5 239/225 14/225 -4/45 -32/405 -4/25 |
| 81/50 63/500 563/500 -9/50 -4/25 -81/250 |
| 81/10 63/100 63/100 1/10 -4/5 -81/50 |
| 63/100 49/1000 49/1000 -7/100 211/225 -63/500 |
| 63/100 49/1000 49/1000 -7/100 -14/225 437/500 |
6 6
o22 : Matrix QQ <-- QQ
i23 : isInFultonHarrisSp2n(g2)
o23 = true
i24 : rhog2 = restrictToSubspace(exteriorPower(3,g2),MyBasisVw3C3)
o24 = | 9397/4500 -1/10 8/45 9/25 -32/405 -8/25 -81/250 0 0 0 0 0 0 0 |
| 81/10 1297/4500 8/5 81/25 0 0 0 0 0 0 -81/250 8/25 -32/405 0 |
| -63/100 7/100 563/500 -63/500 4/5 81/50 0 0 81/250 -4/25 0 9/50 -4/45 0 |
| -63/100 7/100 -14/225 239/225 0 4/5 81/50 0 -4/25 32/405 -9/50 4/45 0 0 |
| 49/1000 0 7/50 0 8837/4500 -63/250 0 -81/250 0 -9/25 0 0 -1/10 0 |
| 49/1000 0 7/100 7/100 -14/225 19/10 -63/500 4/25 9/50 4/45 0 1/10 0 0 |
| 49/1000 0 0 7/50 0 -28/225 8263/4500 -32/405 -8/45 0 -1/10 0 0 0 |
| 0 0 0 0 49/1000 -49/500 49/1000 7703/4500 -7/50 -7/50 0 0 0 -1/10 |
| 0 0 -49/1000 49/1000 0 -63/100 63/100 -4/5 437/500 14/225 -7/100 -7/100 0 4/45 |
| 0 0 49/1000 -49/1000 63/100 -63/100 0 -81/50 63/500 211/225 0 -7/100 -7/100 9/50 |
| 0 49/1000 0 63/50 0 0 81/10 0 -8/5 0 163/4500 28/225 0 -32/405 |
| 0 -49/1000 -63/100 -63/100 0 -81/10 0 0 -81/50 -4/5 63/500 1/10 14/225 -4/25 |
| 0 49/1000 63/50 0 81/10 0 0 0 0 -81/25 0 63/250 737/4500 -81/250 |
| 0 0 0 0 0 0 0 81/10 -63/50 -63/50 49/1000 49/500 49/1000 -397/4500 |
14 14
o24 : Matrix QQ <-- QQ
i25 : c3 = 4;
i26 : v3 = {9/2, 4/3, 5/7, 4, 3/10, 5/7};
i27 : c4 = 8/7
8
o27 = -
7
o27 : QQ
i28 : v4 = {4/7, 1/4, 7/3, 3/7, 8/3, 5/8};
i29 : g3 = symplecticTransvectionMatrix(Q,c4,v4)*symplecticTransvectionMatrix(Q,c3,v3)
o29 = | 1438019/15435 222083/25725 363386/21609 -177923/1715 -1425824/46305 -387494/21609 |
| 66494/2205 14753/3675 68429/12348 -8318/245 -133393/13230 -18614/3087 |
| 88616/945 22142/1575 25412/1323 -1586/15 -89426/2835 -30116/1323 |
| 407096/5145 61952/8575 103709/7203 -151021/1715 -407906/15435 -109736/7203 |
| 130856/1323 34061/2205 179050/9261 -16397/147 -128183/3969 -227266/9261 |
| 14767/441 3229/735 157345/24696 -1849/49 -59473/5292 -20348/3087 |
6 6
o29 : Matrix QQ <-- QQ
i30 : isInFultonHarrisSp2n(g3)
o30 = true
i31 : rhog3 = restrictToSubspace(exteriorPower(3,g3),MyBasisVw3C3)
o31 = | 1588801199/7563150 -3937896/16807 14429651/108045 -2619212/36015 -86311/4410 -603586/64827 -319977856/756315 1092242/27783 5846968/21609 -90158/9261 7824968/16807 -241316/7203 3721/6174 0 |
| 3198472/16807 -1605299401/7563150 4341004/36015 -7079272/108045 -288/343 -84928/2401 -56349728/151263 0 15690448/64827 -11824/1029 309975584/756315 -3902630/64827 -560257/30870 1092242/27783 |
| 43675574/540225 -5396448/60025 40448/1323 -781358/77175 633238/12005 -1279952/6615 -5371696/108045 -7845224/64827 30116/1323 -5712074/138915 2001736/36015 -474982/2205 -19579718/324135 2923484/21609 |
| -102143/7203 453541/28812 -42029/12348 257/8575 -440/343 539258/15435 34296/12005 5912/1029 10946/3087 133393/13230 -124444/36015 586559/15435 3355/3087 -45079/9261 |
| 728552/1225 -5484672/8575 564784/2401 1117248/8575 1087203128/6806835 -336707008/324135 -512072/77175 -1033224632/2268945 -4003472/36015 -25976192/108045 0 -13102272/12005 -16389624/84035 7824968/16807 |
| 47816737/432180 -312156/2401 4033439/43218 -3172717/25725 -5082757/259308 6112112/46305 -266602996/540225 65806253/972405 36635546/108045 1238249/46305 6551136/12005 177923/1715 114985/28812 -120658/7203 |
| 10223215/403368 -3553225/134456 103675/7203 30036/2401 -6050/3087 -478063/43218 13193354/1260525 -186731/92610 -312392/36015 -6710/3087 -209366/16807 -114985/14406 0 3721/6174 |
| 332929/24696 0 -1087645/28812 318504/1715 43109765/1210104 -38259241/129654 44438966/77175 -1267944329/13613670 -3026420/7203 -3354557/43218 -5484672/8575 -624312/2401 -3553225/134456 2253026/84035 |
| -124055/7203 1087645/57624 -157345/24696 -6247/5145 -47300/7203 16021/441 378592/180075 397958/108045 4912/1029 495511/37044 -461506/180075 34201/882 103675/14406 -1317989/648270 |
| 1276901/15435 -159252/1715 230980/9261 -34061/2205 915331/21609 -2004788/9261 -4479112/77175 -45952316/324135 263698/9261 -1007981/27783 558624/8575 -250571/1029 -1043633/21609 17290246/108045 |
| 369800/16807 -9229985/403368 94600/7203 79586/7203 0 -20640/2401 465656/50421 -288/343 -828812/108045 -880/343 -13855246/1260525 -243857/43218 -6050/3087 -7853/13230 |
| -761272/7203 53511083/432180 -1819739/21609 8515732/77175 10320/2401 -407096/5145 46986416/108045 -42464/2401 -98051056/324135 -334106/15435 -259737284/540225 -2279512/46305 2916997/259308 -32009633/972405 |
| 39178952/77175 -4672744/8575 13488016/64827 8958224/77175 12522376/84035 -93972832/108045 0 -56349728/151263 -10743392/108045 -68814832/324135 -512072/77175 -294901328/324135 -1233871888/6806835 856635688/2268945 |
| 0 332929/24696 -248110/7203 2553802/15435 369800/16807 -1522544/7203 39178952/77175 -8081672/252105 -24242740/64827 -1456562/21609 -6216358/11025 -22537451/129654 -15249835/1210104 -452838161/13613670 |
14 14
o31 : Matrix QQ <-- QQ
i32 : F2 = map(S,S,transpose rhog2);
o32 : RingMap S <-- S
i33 : F2(SpGr36)==SpGr36
o33 = true
i34 : F3 = map(S,S,transpose rhog3);
o34 : RingMap S <-- S
i35 : F3(SpGr36)==SpGr36
o35 = true
We check that moving \(M_{9,1}\) by these elements does not change the
value of \(F_{9,1}\).
i36 : F91(transpose(rhog2*transpose(M91)))
o36 = -69066628702801276747468144869007195496549388986880
o36 : QQ
i37 : F91(transpose(rhog3*transpose(M91)))
o37 = -69066628702801276747468144869007195496549388986880
o37 : QQ
Next, we check that moving the point in \(\operatorname{Gr}(9,14)\) corresponding to the cuspidal curve of genus 9 with nonagonal symmetry does not change the value of \(F_{9,1}\).
i38 : LCusp9 = {z_3-(135/2)*z_4,z_9-(1/432)*z_10,z_5-45*z_12,z_6 + 36*z_11,z_0+1/5*z_13};
i39 : F91(matrixFromHyperplaneEquations LCusp9)
7406920142307
o39 = -------------
64
o39 : QQ
i40 : F91(transpose(rhog1*transpose(matrixFromHyperplaneEquations LCusp9)))
7406920142307
o40 = -------------
64
o40 : QQ
i41 : F91(transpose(rhog2*transpose(matrixFromHyperplaneEquations LCusp9)))
7406920142307
o41 = -------------
64
o41 : QQ
i42 : F91(transpose(rhog3*transpose(matrixFromHyperplaneEquations LCusp9)))
7406920142307
o42 = -------------
64
o42 : QQ
i44 : F92(M92)
309054206607
o44 = ------------
2
o44 : QQ
i45 : F92(transpose(rhog1*transpose(M92)))
309054206607
o45 = ------------
2
o45 : QQ
i46 : F92(transpose(rhog2*transpose(M92)))
309054206607
o46 = ------------
2
o46 : QQ
i47 : F92(transpose(rhog3*transpose(M92)))
309054206607
o47 = ------------
2
o47 : QQ
i48 : LTD9 = {z_3-(135/2)*z_4,z_9-(1/432)*z_10,z_5-45*z_12,z_6 + 36*z_11};
i49 : F92(matrixFromHyperplaneEquations LTD9)
1323
o49 = - ----
16
o49 : QQ
i50 : F92(transpose(rhog1*transpose(matrixFromHyperplaneEquations LTD9)))
1323
o50 = - ----
16
o50 : QQ
i51 : F92(transpose(rhog2*transpose(matrixFromHyperplaneEquations LTD9)))
1323
o51 = - ----
16
o51 : QQ
i52 : F92(transpose(rhog3*transpose(matrixFromHyperplaneEquations LTD9)))
1323
o52 = - ----
16
o52 : QQ