This computes $Hom_R(M, R)$.
i1 : R = QQ[x,y,z]/ideal(x^2-y*z); |
i2 : m = ideal(x,y,z); o2 : Ideal of R |
i3 : dualize(m) o3 = ideal x o3 : Ideal of R |
i4 : I = ideal(x,y); o4 : Ideal of R |
i5 : dualize(I) o5 = ideal (z, x) o5 : Ideal of R |
i6 : dualize(I^2) o6 = ideal z o6 : Ideal of R |
i7 : dualize(I^3)
2
o7 = ideal (z , x*z)
o7 : Ideal of R
|
If Strategy => IdealStrategy, then dualize assume the module is isomorphic to an ideal, embeds it as an ideal, and computes the dual by forming a colon. ModuleStrategy simply computes the Hom. The default Strategy for modules is ModuleStrategy, and the default Strategy for ideals is IdealStrategy. This is because there is overhead using the opposite strategy (involving embedding modules as ideals). Frequently IdealStrategy is faster, but not always. Consider first a D4 singularity in characteristic 2.
i8 : R = ZZ/2[x,y,z]/ideal(z^2-x*y*z-x^2*y-x*y^2); |
i9 : m = ideal(x,y,z); o9 : Ideal of R |
i10 : J = m^9; o10 : Ideal of R |
i11 : M = J*R^1; |
i12 : time dualize(J, Strategy=>IdealStrategy);
-- used 0.240044 seconds
o12 : Ideal of R
|
i13 : time dualize(J, Strategy=>ModuleStrategy);
-- used 3.57903 seconds
o13 : Ideal of R
|
i14 : time dualize(M, Strategy=>IdealStrategy);
-- used 3.81413 seconds
|
i15 : time dualize(M, Strategy=>ModuleStrategy);
-- used 0.0138112 seconds
|
i16 : time embedAsIdeal dualize(M, Strategy=>ModuleStrategy);
-- used 0.0110163 seconds
o16 : Ideal of R
|
For monomial ideals in toric rings, frequently ModuleStrategy appears faster.
i17 : R = ZZ/7[x,y,u,v]/ideal(x*y-u*v); |
i18 : I = ideal(x,u); o18 : Ideal of R |
i19 : J = I^15; o19 : Ideal of R |
i20 : time dualize(J, Strategy=>IdealStrategy);
-- used 0.319259 seconds
o20 : Ideal of R
|
i21 : time dualize(J, Strategy=>ModuleStrategy);
-- used 0.0306731 seconds
o21 : Ideal of R
|
KnownDomain is an option for dualize. If it is false (default is true), then the computer will first check whether the ring is a domain, if it is not then it will revert to ModuleStrategy. If KnownDomain is set to true for a non-domain, then the function can return an incorrect answer.
i22 : R = QQ[x,y]/ideal(x*y); |
i23 : J = ideal(x,y); o23 : Ideal of R |
i24 : dualize(J, KnownDomain=>true) o24 = ideal x o24 : Ideal of R |
i25 : dualize(J, KnownDomain=>false) o25 = ideal (y, x) o25 : Ideal of R |
The object dualize is a method function with options.