Constructive Arithmetics in Ore Localizations with Enough - - PowerPoint PPT Presentation

constructive arithmetics in ore localizations with enough
SMART_READER_LITE
LIVE PREVIEW

Constructive Arithmetics in Ore Localizations with Enough - - PowerPoint PPT Presentation

Constructive Arithmetics in Ore Localizations with Enough Commutativity Johannes Hoffmann, Viktor Levandovskyy 1 RWTH Aachen University, Germany ISSAC 2018, New York 1 supported by DFG Transregio 195 Johannes Hoffmann (RWTH Aachen) Constructive


slide-1
SLIDE 1

Constructive Arithmetics in Ore Localizations with Enough Commutativity

Johannes Hoffmann, Viktor Levandovskyy1

RWTH Aachen University, Germany

ISSAC 2018, New York

1supported by DFG Transregio 195 Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 1 / 31

slide-2
SLIDE 2

Content

1

Motivation

2

The intersection problem in polynomial algebras

3

The closure problem

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 2 / 31

slide-3
SLIDE 3

Content

1

Motivation

2

The intersection problem in polynomial algebras

3

The closure problem

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 3 / 31

slide-4
SLIDE 4

Localization of commutative domains

Definition

A subset S of a ring R is called multiplicative set if 0 / ∈ S, 1 ∈ S and S is multiplicatively closed, that is, ∀s, t ∈ S : s · t ∈ S. Notation: [S] := the smallest multiplicative superset of S.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 4 / 31

slide-5
SLIDE 5

Localization of commutative domains

Definition

A subset S of a ring R is called multiplicative set if 0 / ∈ S, 1 ∈ S and S is multiplicatively closed, that is, ∀s, t ∈ S : s · t ∈ S. Notation: [S] := the smallest multiplicative superset of S.

Theorem (Classical)

Let S be a multiplicative set in a commutative domain R. Then S−1R := r s | r ∈ R, s ∈ S

  • =
  • s−1r | (s, r) ∈ S × R
  • is a commutative domain (with the usual addition and multiplication of

fractions).

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 4 / 31

slide-6
SLIDE 6

Commutative examples

Classical localizations

Let R be a commutative domain. Quot(R) :=

  • p

q | p, q ∈ R, q = 0

  • = (R \ {0})−1R

Rp :=

  • p

q | p, q ∈ R, q /

∈ p

  • = (R \ p)−1R, p prime ideal of R

Rf := p

f k | p ∈ R, k ∈ N0

  • = [f ]−1R, f ∈ R \ {0}

R Quot(R) Rf Rp

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 5 / 31

slide-7
SLIDE 7

The hierarchy of Ore localizations: localization of. . .

commutative domains

slide-8
SLIDE 8

The hierarchy of Ore localizations: localization of. . .

commutative domains commutative rings commutative

slide-9
SLIDE 9

The hierarchy of Ore localizations: localization of. . .

commutative domains commutative rings commutative arbitrary domains domains

slide-10
SLIDE 10

The hierarchy of Ore localizations: localization of. . .

commutative domains commutative rings commutative arbitrary domains domains arbitrary rings

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 6 / 31

slide-11
SLIDE 11

Left Ore sets, left denominator sets

Definition

Let S be a subset of a ring R. S satisfies the left Ore condition in R if ∀ s ∈ S, r ∈ R ∃ ˜ s ∈ S, ˜ r ∈ R : ˜ sr = ˜ rs.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 7 / 31

slide-12
SLIDE 12

Left Ore sets, left denominator sets

Definition

Let S be a subset of a ring R. S satisfies the left Ore condition in R if ∀ s ∈ S, r ∈ R ∃ ˜ s ∈ S, ˜ r ∈ R : ˜ sr = ˜ rs. Left Ore set = multiplicative set + left Ore condition

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 7 / 31

slide-13
SLIDE 13

Left Ore sets, left denominator sets

Definition

Let S be a subset of a ring R. S satisfies the left Ore condition in R if ∀ s ∈ S, r ∈ R ∃ ˜ s ∈ S, ˜ r ∈ R : ˜ sr = ˜ rs. Left Ore set = multiplicative set + left Ore condition S is left reversible in R if ∀ s ∈ S, r ∈ R : rs = 0 ⇒ ∃˜ s ∈ S : ˜ sr = 0.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 7 / 31

slide-14
SLIDE 14

Left Ore sets, left denominator sets

Definition

Let S be a subset of a ring R. S satisfies the left Ore condition in R if ∀ s ∈ S, r ∈ R ∃ ˜ s ∈ S, ˜ r ∈ R : ˜ sr = ˜ rs. Left Ore set = multiplicative set + left Ore condition S is left reversible in R if ∀ s ∈ S, r ∈ R : rs = 0 ⇒ ∃˜ s ∈ S : ˜ sr = 0. Left denominator set = left Ore set + left reversibility

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 7 / 31

slide-15
SLIDE 15

Left Ore sets, left denominator sets

Definition

Let S be a subset of a ring R. S satisfies the left Ore condition in R if ∀ s ∈ S, r ∈ R ∃ ˜ s ∈ S, ˜ r ∈ R : ˜ sr = ˜ rs. Left Ore set = multiplicative set + left Ore condition S is left reversible in R if ∀ s ∈ S, r ∈ R : rs = 0 ⇒ ∃˜ s ∈ S : ˜ sr = 0. Left denominator set = left Ore set + left reversibility

Consequences of the left Ore condition on S in R

Finite collections of elements from S have common left multiples in S. Any right fraction rs−1 can be rewritten as a left fraction ˜ s−1˜ r.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 7 / 31

slide-16
SLIDE 16

Construction of the left Ore localization

Theorem (Ore, 1931)

The following is an equivalence relation on S × R: (s1, r1) ∼ (s2, r2) ⇔ ∃ ˜ s ∈ S, ˜ r ∈ R : ˜ ss2 = ˜ rs1 and ˜ sr2 = ˜ rr1

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 8 / 31

slide-17
SLIDE 17

Construction of the left Ore localization

Theorem (Ore, 1931)

The following is an equivalence relation on S × R: (s1, r1) ∼ (s2, r2) ⇔ ∃ ˜ s ∈ S, ˜ r ∈ R : ˜ ss2 = ˜ rs1 and ˜ sr2 = ˜ rr1 S−1R := (S × R/ ∼, +, ·) is a ring via + : S−1R × S−1R → S−1R, (s1, r1) + (s2, r2) := (˜ ss1, ˜ sr1 + ˜ rr2), where ˜ s ∈ S and ˜ r ∈ R satisfy ˜ ss1 = ˜ rs2, and · : S−1R × S−1R → S−1R, (s1, r1) · (s2, r2) := (˜ ss1, ˜ rr2), where ˜ s ∈ S and ˜ r ∈ R satisfy ˜ sr1 = ˜ rs2.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 8 / 31

slide-18
SLIDE 18

Partial classification of Ore localizations

Definition

Let K be a field and R a K-algebra, S a left denominator set in R.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 9 / 31

slide-19
SLIDE 19

Partial classification of Ore localizations

Definition

Let K be a field and R a K-algebra, S a left denominator set in R. Monoidal localization: S is generated as a monoid by countably many elements Example: [f1, . . . , fk]−1K[x], fi ∈ K[x] \ {0}

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 9 / 31

slide-20
SLIDE 20

Partial classification of Ore localizations

Definition

Let K be a field and R a K-algebra, S a left denominator set in R. Monoidal localization: S is generated as a monoid by countably many elements Example: [f1, . . . , fk]−1K[x], fi ∈ K[x] \ {0} Geometric localization: Let n ∈ N, K[x] := K[x1, . . . , xn], J an ideal in K[x], p a prime ideal in K[x]/J and S = (K[x]/J) \ p Example: K[x]p

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 9 / 31

slide-21
SLIDE 21

Partial classification of Ore localizations

Definition

Let K be a field and R a K-algebra, S a left denominator set in R. Monoidal localization: S is generated as a monoid by countably many elements Example: [f1, . . . , fk]−1K[x], fi ∈ K[x] \ {0} Geometric localization: Let n ∈ N, K[x] := K[x1, . . . , xn], J an ideal in K[x], p a prime ideal in K[x]/J and S = (K[x]/J) \ p Example: K[x]p Rational localization: T ⊆ R is a K-subalgebra, S = T \ {0} Special case: R is generated by a set X of variables and T is generated by a subset of X ⇒ S−1R is essential rational Example: (K[x] \ {0})−1K[x, y] ∼ = K(x)[y]

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 9 / 31

slide-22
SLIDE 22

Previously on ISSAC’17

Setup: a left Ore set S in a (not necessarily commutative) domain R. Goal: provide algorithms for basic arithmetic in S−1R. Restrictions: R is a G-algebra, S belongs to one of the types abovea. Key problem: intersection of left ideal with submonoid Result: library olga.lib for Singular:Plural Johannes Hoffmann and Viktor Levandovskyy. A Constructive Approach to Arithmetics in Ore Localizations. In Proc. ISSAC’17, pages 197–204. ACM Press, 2017.

aNote that further computability restrictions apply. Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 10 / 31

slide-23
SLIDE 23

Addressing the key problem

The intersection problem

Let S be a left denominator set in a ring R and I a left ideal in R. The intersection problem is to decide whether I ∩ S = ∅ and to compute an element contained in this intersection when the answer is negative.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 11 / 31

slide-24
SLIDE 24

Addressing the key problem

The intersection problem

Let S be a left denominator set in a ring R and I a left ideal in R. The intersection problem is to decide whether I ∩ S = ∅ and to compute an element contained in this intersection when the answer is negative.

Recent result (Posur, 2018)

The intersection problem is a main ingredient for solving linear systems

  • ver localizations of commutative rings.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 11 / 31

slide-25
SLIDE 25

Addressing the key problem

The intersection problem

Let S be a left denominator set in a ring R and I a left ideal in R. The intersection problem is to decide whether I ∩ S = ∅ and to compute an element contained in this intersection when the answer is negative.

Recent result (Posur, 2018)

The intersection problem is a main ingredient for solving linear systems

  • ver localizations of commutative rings.

Follow-up question

What can we do in the commutative setting?

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 11 / 31

slide-26
SLIDE 26

Content

1

Motivation

2

The intersection problem in polynomial algebras Polynomial algebras I: essential rational intersection Polynomial algebras II: geometric intersection Polynomial algebras III: finitely generated monoidal intersection

3

The closure problem

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 12 / 31

slide-27
SLIDE 27

The intersection problem in polynomial algebras

Definition

Let K be a field, n ∈ N, K[x] := K[x1, . . . , xn] and J an ideal in K[x]. Then we consider the polynomial algebra R := K[x]/J.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 13 / 31

slide-28
SLIDE 28

The intersection problem in polynomial algebras

Definition

Let K be a field, n ∈ N, K[x] := K[x1, . . . , xn] and J an ideal in K[x]. Then we consider the polynomial algebra R := K[x]/J.

What can we do?

Let I be an ideal of R and S a multiplicative set in R. Then we can solve the intersection problem I ∩ S in either of the following cases: S−1R is monoidal and S is finitely generated. S−1R is geometric. S−1R is essential rational. We can also decide whether a multiplicative submonoid of R contains 0.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 13 / 31

slide-29
SLIDE 29

Toolbox

Setting

K is a field, K[x] := K[x1, . . . , xn] and K[y] := K[y1, . . . , ym], I = K[x]h1, . . . , hk and J = K[y]g1, . . . , gl are ideals, f1, . . . , fn ∈ K[y]. Consider the homomorphism ϕ : K[x]/I → K[y]/J, xi → fi.

Computing the kernel of a polynomial algebra homomorphism

1

Let H := K[x,y]h1, . . . , hk, g1, . . . , gl, x1 − f1, . . . , xn − fn.

2

Compute ker(ϕ) = H ∩ K[x] by eliminating y1, . . . , ym.

Observation

Computing kernels is equivalent to computing preimages of ideals.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 14 / 31

slide-30
SLIDE 30

Polynomial algebras I: essential rational

Setting

K a field, K[x] := K[x1, . . . , xn] J an ideal in K[x], I an ideal in R := K[x]/J r ∈ {1, . . . , n}, K[t] := K[t1, . . . , tr] ˆ S = K[x1 + J, . . . , xr + J] ⊆ R, S = ˆ S \ {0}

Essential rational intersection in polynomial algebras

1

Let ϕ : K[t] → R, ti → xi.

2

Compute the preimage ϕ−1(I) =: K[t]w1, . . . , wk.

3

If ϕ(wi) = 0 for some i return ϕ(wi).

4

Otherwise return 0.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 15 / 31

slide-31
SLIDE 31

Polynomial algebras II: geometric

Setting

K a field, K[x] := K[x1, . . . , xn] J an ideal in K[x], I an ideal in R := K[x]/J p a prime ideal in R, S = R \ p

Geometric intersection in polynomial algebras

1

Let π : K[x] ։ R.

2

Compute the preimage q := π−1(p).

3

If NF(h|q) = 0 for some generator h of I return h + J.

4

Otherwise return 0.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 16 / 31

slide-32
SLIDE 32

Polynomial algebras III: finitely generated monoidal

Setting

K a field, K[x] := K[x1, . . . , xn] J an ideal in K[x], I an ideal in R := K[x]/J f1, . . . , fk ∈ K[x], K[t] = K[t1, . . . , tk], S = [f1 + J, . . . , fk + J]

Decide whether 0 ∈ S

1

Let ψ : K[t] → R, ti → fi + J.

2

Compute ker(ψ) ⊆ K[t].

3

Compute the saturation M := ker(ψ) : t1 · . . . · tm∞.

4

If 1 ∈ M return “yes”, otherwise return “no”.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 17 / 31

slide-33
SLIDE 33

Polynomial algebras III: finitely generated monoidal

Setting

K a field, K[x] := K[x1, . . . , xn] J an ideal in K[x], I an ideal in R := K[x]/J f1, . . . , fk ∈ K[x], K[t] = K[t1, . . . , tk], S = [f1 + J, . . . , fk + J]

Finitely generated monoidal intersection

1

Let ψ : K[t] → R, ti → fi + J.

2

Compute the preimage L := ψ−1(I) ⊆ K[t].

3

If ψ(L) = {0} return 0.

4

Compute ker(ψ) ⊆ K[t].

5

Let ϕ : K[t] → K[t, q±1] := K[t, q1, q−1

1 , . . . qk, q−1 k ], ti → qiti.

6

Compute a reduced GB of the monomial ideal W := K[t,q±1]L ∩ K[t].

7

If NF(w| ker(ψ)) = 0 for some monomial gen. w of W return w + J.

8

Otherwise return 0.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 18 / 31

slide-34
SLIDE 34

Content

1

Motivation

2

The intersection problem in polynomial algebras

3

The closure problem Commutative decomposition closure Central closures for G-algebras

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 19 / 31

slide-35
SLIDE 35

The closure problem

Definition

Let S be a left Ore set in a ring R and M a left R-module. The left Ore localization of M at S is S−1M := S−1R ⊗R M. The homomorphism ε = εS,R,M : M → S−1M, m → (1, m) := (1, 1) ⊗ m is called the localization map of M with respect to S.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 20 / 31

slide-36
SLIDE 36

The closure problem

Definition

Let S be a left Ore set in a ring R and M a left R-module. The left Ore localization of M at S is S−1M := S−1R ⊗R M. The homomorphism ε = εS,R,M : M → S−1M, m → (1, m) := (1, 1) ⊗ m is called the localization map of M with respect to S.

The closure problem

Given a left R-submodule P of M, find the local closure or S-closure PS := ε−1(S−1P).

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 20 / 31

slide-37
SLIDE 37

The closure problem

Definition

Let S be a left Ore set in a ring R and M a left R-module. The left Ore localization of M at S is S−1M := S−1R ⊗R M. The homomorphism ε = εS,R,M : M → S−1M, m → (1, m) := (1, 1) ⊗ m is called the localization map of M with respect to S.

The closure problem

Given a left R-submodule P of M, find the local closure or S-closure PS := ε−1(S−1P).

Lemma

PS = {m ∈ M | ∃s ∈ S : sm ∈ P} =: LSatM

S (P).

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 20 / 31

slide-38
SLIDE 38

Example: symbolic powers of prime ideals

Observation

Powers of prime ideals need not be primary.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 21 / 31

slide-39
SLIDE 39

Example: symbolic powers of prime ideals

Observation

Powers of prime ideals need not be primary.

Definition

Let p be a prime ideal in a commutative ring R. The ideal p(k) :=

  • f ∈ R | ∃s ∈ R \ p : sf ∈ pk

is called the k-th symbolic power of p.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 21 / 31

slide-40
SLIDE 40

Example: symbolic powers of prime ideals

Observation

Powers of prime ideals need not be primary.

Definition

Let p be a prime ideal in a commutative ring R and S = R \ p. The ideal p(k) :=

  • f ∈ R | ∃s ∈ R \ p : sf ∈ pk

= (pk)S is called the k-th symbolic power of p.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 21 / 31

slide-41
SLIDE 41

Example: symbolic powers of prime ideals

Observation

Powers of prime ideals need not be primary.

Definition

Let p be a prime ideal in a commutative ring R and S = R \ p. The ideal p(k) :=

  • f ∈ R | ∃s ∈ R \ p : sf ∈ pk

= (pk)S is called the k-th symbolic power of p.

Lemma

The ideal p(k) is the smallest p-primary ideal containing pk.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 21 / 31

slide-42
SLIDE 42

Commutative decomposition closure: preparation

Lemma

Let S be a multiplicative set in a commutative ring R and q a primary ideal

  • f R. Then

qS =

  • R,

if q ∩ S = ∅, q, if q ∩ S = ∅.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 22 / 31

slide-43
SLIDE 43

Commutative decomposition closure: preparation

Lemma

Let S be a multiplicative set in a commutative ring R and q a primary ideal

  • f R. Then

qS =

  • R,

if q ∩ S = ∅, q, if q ∩ S = ∅.

Lemma

Let S be a left Ore set in a ring R, M a left R-module and P1, . . . , Pk some left R-submodules of M. Then  

k

  • j=1

Pj  

S

=

k

  • j=1

PS

j .

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 22 / 31

slide-44
SLIDE 44

Commutative decomposition closure: algorithm

Let S be a multiplicative set in a commutative ring R and I an ideal in R.

Commutative decomposition closure

1

Obtain a decomposition of I into primary ideals: I = k

j=1 qj.

2

For each j set ˜ qj :=

  • qj,

if qj ∩ S = ∅, R, if qj ∩ S = ∅.

3

Return I S = k

j=1 ˜

qj.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 23 / 31

slide-45
SLIDE 45

Commutative decomposition closure: algorithm

Let S be a multiplicative set in a commutative ring R and I an ideal in R.

Commutative decomposition closure

1

Obtain a decomposition of I into primary ideals: I = k

j=1 qj.

2

For each j set ˜ qj :=

  • qj,

if qj ∩ S = ∅, R, if qj ∩ S = ∅.

3

Return I S = k

j=1 ˜

qj.

Observation

We need to solve several intersection problems with primary ideals.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 23 / 31

slide-46
SLIDE 46

G-algebras (PBW algebras, algebras of solvable type)

Definition

For a field K, The K-algebra A := K is called a G-algebra, if:

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 24 / 31

slide-47
SLIDE 47

G-algebras (PBW algebras, algebras of solvable type)

Definition

For a field K, n ∈ N The K-algebra A := Kx1, . . . , xn

  • is called a G-algebra, if:

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 24 / 31

slide-48
SLIDE 48

G-algebras (PBW algebras, algebras of solvable type)

Definition

For a field K, n ∈ N and 1 ≤ i < j ≤ n The K-algebra A := Kx1, . . . , xn | {xjxi = xixj : 1 ≤ i < j ≤ n} is called a G-algebra, if:

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 24 / 31

slide-49
SLIDE 49

G-algebras (PBW algebras, algebras of solvable type)

Definition

For a field K, n ∈ N and 1 ≤ i < j ≤ n consider the constants ci,j ∈ K \ {0} The K-algebra A := Kx1, . . . , xn | {xjxi = ci,jxixj : 1 ≤ i < j ≤ n} is called a G-algebra, if:

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 24 / 31

slide-50
SLIDE 50

G-algebras (PBW algebras, algebras of solvable type)

Definition

For a field K, n ∈ N and 1 ≤ i < j ≤ n consider the constants ci,j ∈ K \ {0} and polynomials di,j ∈ K[x1, . . . , xn]. The K-algebra A := Kx1, . . . , xn | {xjxi = ci,jxixj + di,j : 1 ≤ i < j ≤ n} is called a G-algebra, if:

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 24 / 31

slide-51
SLIDE 51

G-algebras (PBW algebras, algebras of solvable type)

Definition

For a field K, n ∈ N and 1 ≤ i < j ≤ n consider the constants ci,j ∈ K \ {0} and polynomials di,j ∈ K[x1, . . . , xn]. The K-algebra A := Kx1, . . . , xn | {xjxi = ci,jxixj + di,j : 1 ≤ i < j ≤ n} is called a G-algebra, if: there exists a monomial total well-ordering < on K[x1, . . . , xn] such that for any 1 ≤ i < j ≤ n either di,j = 0 or the leading monomial of di,j with respect to < is smaller than xixj, and

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 24 / 31

slide-52
SLIDE 52

G-algebras (PBW algebras, algebras of solvable type)

Definition

For a field K, n ∈ N and 1 ≤ i < j ≤ n consider the constants ci,j ∈ K \ {0} and polynomials di,j ∈ K[x1, . . . , xn]. The K-algebra A := Kx1, . . . , xn | {xjxi = ci,jxixj + di,j : 1 ≤ i < j ≤ n} is called a G-algebra, if: there exists a monomial total well-ordering < on K[x1, . . . , xn] such that for any 1 ≤ i < j ≤ n either di,j = 0 or the leading monomial of di,j with respect to < is smaller than xixj, and {xα1

1

· . . . · xαn

n

| αi ∈ N0} is a K-basis of A.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 24 / 31

slide-53
SLIDE 53

G-algebras (PBW algebras, algebras of solvable type)

Definition

For a field K, n ∈ N and 1 ≤ i < j ≤ n consider the constants ci,j ∈ K \ {0} and polynomials di,j ∈ K[x1, . . . , xn]. The K-algebra A := Kx1, . . . , xn | {xjxi = ci,jxixj + di,j : 1 ≤ i < j ≤ n} is called a G-algebra, if: there exists a monomial total well-ordering < on K[x1, . . . , xn] such that for any 1 ≤ i < j ≤ n either di,j = 0 or the leading monomial of di,j with respect to < is smaller than xixj, and {xα1

1

· . . . · xαn

n

| αi ∈ N0} is a K-basis of A.

Remark

G-algebras are Noetherian domains. There exists a Gröbner basis theory for G-algebras plus implementation (most extensive in Singular:Plural).

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 24 / 31

slide-54
SLIDE 54

Examples of G-algebras

Weyl algebras (Kx1, . . . , xn, ∂1, . . . , ∂n | ∀i : ∂ixi = xi∂i + 1) Shift algebras (Kx1, . . . , xn, s1, . . . , sn | ∀i : sixi = (xi + 1)si) q-Weyl algebras (Kx, ∂ | ∀i∃qi ∈ K ∗ : ∂ixi = qixi∂i + 1) q-Shift algebras (Kx, s | ∀i∃qi ∈ K ∗ : sixi = qixisi) Integration algebras (Kx, I | ∀i : Iixi = xiIi + I 2

i )

Universal enveloping algebras of finite-dimensional Lie algebras Many quantum groups Tensor products over K of G-algebras . . .

Recent result (Heinle, L., 2012-2017)

G-algebras are finite factorization domains. Factorization in G-algebras is algorithmic and implemented in ncfactor.lib in Singular:Plural.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 25 / 31

slide-55
SLIDE 55

Central saturation

Definition

Let R be a ring, q ∈ Z(R), k ∈ N and I a left R-submodule of Rk. The central quotient of I by q is I : q :=

  • f ∈ Rk | qf = fq ∈ I
  • .

The central saturation of I by q is I : q∞ :=

  • j∈N0

(I : qj) =

  • f ∈ Rk | ∃n ∈ N0 : qnf ∈ I
  • .

The central saturation index of I by q is the smallest n ∈ N0 ∪ {∞} such that (I : qn) = (I : q∞), denoted by Satindex(I, q).

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 26 / 31

slide-56
SLIDE 56

Central monoidal closure

Task

Let A be a G-algebra and I a left A-submodule of Ak. Let g1, . . . , gk ∈ Z(A), then S := [g1, . . . , gk] is a left Ore set in A. Goal: compute I S.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 27 / 31

slide-57
SLIDE 57

Central monoidal closure

Task

Let A be a G-algebra and I a left A-submodule of Ak. Let g1, . . . , gk ∈ Z(A), then S := [g1, . . . , gk] is a left Ore set in A. Goal: compute I S.

Solution

1

Set g := k

j=1 gj, then T := [g] is a left Ore set in A.

2

Compute I : g∞ = I S via central saturation.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 27 / 31

slide-58
SLIDE 58

Central essential rational closure: problem

Task

Let K be a field and consider a G-algebra A := Kx1, . . . , xn

  • central

, y1, . . . , ym | Relations such that x generates a sub-G-algebra B ⊆ Z(A) of A. Then S := B \ {0} is a left Ore set in A and B. Let I be a left R-submodule of Ar. Goal: compute I S.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 28 / 31

slide-59
SLIDE 59

Central essential rational closure: problem

Task

Let K be a field and consider a G-algebra A := Kx1, . . . , xn

  • central

, y1, . . . , ym | Relations such that x generates a sub-G-algebra B ⊆ Z(A) of A. Then S := B \ {0} is a left Ore set in A and B. Let I be a left R-submodule of Ar. Goal: compute I S.

Example

For intuition: think of D[s] = K[s1, . . . , sn] ⊗K A m

2 (K), where A m 2 (K) is a

Weyl algebra generated by x1, . . . , x m

2 , ∂1, . . . , ∂ m 2 . Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 28 / 31

slide-60
SLIDE 60

Central essential rational closure: algorithm

For an ordering ≤ on A, denote by ≤POT the position-over-term ordering

  • n Ak based on ≤.

Let ε : Ar → S−1Ar, m → (1, m).

Algorithm

1

Let ≤ = (≤1, ≤2) be an antiblock-ordering on A and = ≤POT.

2

Compute a Gröbner basis G of I with respect to .

3

Let h :=

  • g∈G lc2(ε(g)) ∈ K[x] \ {0}, where 2 = ≤2POT.

4

Compute k := Satindex(I : h).

5

Return (I : hk).

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 29 / 31

slide-61
SLIDE 61

Algorithmic conclusion

We can solve the intersection problem I ∩ S. . .

ISSAC’17: in G-algebras R, where S−1R is. . .

◮ finitely generated monoidal with commuting generators (if I ∩ S = ∅), ◮ geometric in a Weyl algebra, or ◮ essential rational (if elimination is possible).

ISSAC’18: in polynomial algebras R = K[x]/J, where S−1R is

◮ finitely generated monoidal with commuting generators, ◮ geometric, or ◮ essential rational.

We can solve the closure problem I S. . .

in commutative rings, if I is effectively decomposable and we can solve the intersection problem q ∩ S for every involved primary ideal in G-algebras, where S−1R is. . .

◮ finitely generated monoidal with central generators, or ◮ central essential rational.

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 30 / 31

slide-62
SLIDE 62

The latest version of Singular is available at:

http://www.singular.uni-kl.de

The latest version of olga.lib is available at:

http://www.math.rwth-aachen.de/~Johannes.Hoffmann/singular.html

Johannes Hoffmann (RWTH Aachen) Constructive Ore Arithmetics ISSAC 2018, New York 31 / 31