RosenfeldGroebner - Maple Help
For the best experience, we recommend viewing online help using Google Chrome or Microsoft Edge.

Online Help

All Products    Maple    MapleSim


DifferentialAlgebra

  

RosenfeldGroebner

  

the RosenfeldGroebner function

 

Calling Sequence

Parameters

Options

Description

Examples

Calling Sequence

RosenfeldGroebner(sys, R, opts)

Parameters

sys

-

a list or a set of differential equations or inequations; all rational in the independent and dependent variables and their derivatives

R

-

a differential polynomial ring built by DifferentialRing specifying the independent and dependent variables and a ranking (ordering) for them

additionally

-

(optional) used together with arbitrary and relations to add to the arbitrary of the differential ring

arbitrary

-

(optional) a function or a name, or a list of them, specifying arbitrary objects, that either override those specified in the differential ring R or add to them when additionally is given

assumeisprime

-

(optional) used when sys is a regular differential chain and proceeds changing rankings regardless of whether sys is prime or not

attributes

-

(optional) to specify the attributes expected for the differential chains (cases) returned

discardlowerdimension

-

(optional) to discard inexistent cases based on their (differential) dimension

notation

-

(optional) to request the output to be in any of jet, tjet, diff or Diff notations

relations

-

(optional) to specify relations that are satisfied by the arbitrary objects

singsol

-

(optional) can be all (default), none or essential, to request the computation of all, none or only the essential singular cases

Options

• 

additionally is used together with arbitrary, to add variables to the list declared as arbitrary when creating the differential ring using DifferentialRing

• 

arbitrary = ...  a function or a name, or a list of them, specifying arbitrary objects. This option permits to run a computation with different arbitrary objects than those specified in the differential ring R passed to RosenfeldGroebner, or to add arbitrary objects to those in R when additionally is passed together.

• 

assumeisprime used when changing the ranking of a regular differential chain, this option tells RosenfeldGroebner to proceed disregarding whether the received differential chain is or not prime.

• 

attributes = L where L is a list involving some of the keywords differential, prime, primitive, squarefree, coherent, autoreduced, normalized. This option permits to modify the behavior of the function (e.g. if differential is omitted, the decomposition is purely algebraic ; if normalized is omitted, the returned regular differential chains do not need to be normalized). The presence of some of the attributes may imply, automatically, the presence of some other ones. For more details, see the Glossary.

• 

discardlowerdimension = nocase, safecase, odecase, pdecase. This option controls the splittings performed by RosenfeldGroebner by discarding any component whose dimension (differential dimension in the differential case) is lower than the number of input equations simplified with respect to their integrability conditions.

– 

nocase disables this option.

– 

safecase is the default value. The option is only activated in the non-differential case and in the case of a single input equation. These two cases are theoretically proven.

– 

odecase applies also the option to general ODE systems. The option implements a conjecture in this case [R50, Questions for investigation, 10].

– 

pdecase applies also the option to general PDE systems. The option implements a conjecture in this case.

• 

notation = jet, tjet, diff or Diff. Specifies the notation used for the result of the function call. If not specified, the notation of the input is used.

• 

memout = nonnegative. Specifies a memory limit, in MB, for the computation. Default is zero (no memory out).

• 

relations = ... where the right-hand-side is a set of equations satisfied by the arbitrary objects of the differential ring R or passed with the option arbitrary. Note that all the functions and names (symbolic variables) involved in these relations will automatically be taken as arbitrary, even if they were not declared as such in the differential ring R. Also, all the dependent variables of the differential ring R that are ranked lower than the objects found in the given relations, or within a same orderly sublist of the ranking, are also considered arbitrary.

• 

singsol = all, essential, none. This option controls the splittings performed by RosenfeldGroebner.

– 

all is the default value.

– 

essential ensures that the returned decomposition is irredundant whenever this is possible, i.e., whenever sys is made of only one equation.

– 

none makes RosenfeldGroebner return at most one regular differential chain. The returned chain is the first one that would have been computed without setting this option. In many cases, this first regular differential chain could be considered as the general component, though this notion is not algebraically well-defined in general.

Description

• 

Given a system of equations sys, containing equations (of the form algebraic or algebraic = algebraic) and inequations (of the form algebraic <> 0), RosenfeldGroebner(sys, R) splits the given system sys into cases, each one specified by their Equations and Inequations. One and only one of these cases is the general case, and the rest are singular cases that may exist only when sys is nonlinear. In each of the cases returned, there is no one equation that is the consequence of the other equations and all the integrability conditions are taken into account. The relation between the system sys and the systems of each of theses cases is the solutions, general and singular, of sys are given by the union of the general solutions of each of the returned systems. NOTE: the number of cases returned and the equations and inequations that appear in these cases depend on how the dependent and independent variables are ranked inside R.

• 

More technically, RosenfeldGroebner(sys, R) returns a representation of the radical of the differential ideal generated by sys, as an intersection of radical differential ideals saturated by the multiplicative family generated by the inequations found in sys. This representation consists of a list of regular differential chains with respect to the ranking of R.

• 

The system sys may involve rational differential fractions, that is the ratio between differential polynomials. In that case, the denominators of these rational fractions are automatically considered inequations.

• 

The second argument, R representing a differential ring, is constructed using DifferentialRing, and specifies the dependent and independent variables and a ranking for them.

• 

The differential polynomials in sys are expected to have coefficients in the field Q(x1, ..., xp), obtained by adjoining the independent variables to the field of the rational numbers. Arbitrary symbolic variables and functions can also be part of the coefficient field, for that purpose you can use the option arbitrary = ... when declaring the differential ring or pass this option directly to RosenfeldGroebner.

  

REMARK: By default, all symbolic variables and functions that were not stated as dependent variables when defining the differential ring are automatically considered arbitrary (equivalent to have indicated them as arbitrary to DifferentialRing). When that happens a related warning message is displayed on the screen.

• 

The regular differential chains in the output of RosenfeldGroebner are assumed to hold the attributes: differential, autoreduced, primitive, squarefree, normalized and coherent. For more details on attributes, see the Glossary.

• 

The returned list may also be empty. In that case sys has no solutions and so the radical of the differential ideal generated by sys is the unit ideal, see [R50, chapter II, Theorem of zeros]. Also, if sys is empty, the returned representation involves a single regular differential chain also empty, which describes the zero ideal.

Recomputing ideals changing the ranking

• 

RosenfeldGroebner can also receive as first argument a prime regular differential chain - say regchain - computed with some ranking, by RosenfeldGroebner or by PretendRegularDifferentialChain, plus an indication of the form blocks = ..., derivations = ... that specifies a new ranking for either or both the dependent (blocks) and independent (derivations) variables, and it recomputes regchain with respect to the new ranking. The returned differential chain defines the same prime differential ideal as regchain and holds the same attributes as regchain. To specify the new ranking for the dependent and/or independent variables you can also directly pass, anywhere after the first argument, a differential ring computed with DifferentialRing. In all cases the new ranking must involve the same dependent and independent variables of the old ranking with which regchain got computed.

• 

NOTE: The method used only applies to prime differential ideals. To know whether a regular differential chain is or not prime use the Is command, as in Is(prime, regchain). If regchain passed to RosenfeldGroebner is not prime, you can nevertheless request RosenfeldGroebner to proceed assuming that regchain is prime by passing the optional argument assumeisprime, but the resulting differential ideal is not guaranteed to represent the same ideal as regchain.

Examples

Main usage

withDifferentialAlgebra

BelongsTo&comma;DifferentialRing&comma;Equations&comma;Get&comma;Inequations&comma;Is&comma;NormalForm&comma;PowerSeriesSolution&comma;ReducedForm&comma;RosenfeldGroebner&comma;Tools

(1)
• 

NOTE: Prior to any computations with the DifferentialAlgebra package, the appropriate differential indeterminates (dependent variables) and derivation variables (independent variables) with the ranking for them must be defined with the command DifferentialRing as shown in the following examples.

Removing consequences and simplifying taking into account the integrability conditions

  

When tackling a system of equations with RosenfeldGroebner, in each of the cases returned, there is no one equation that is the consequence of the other equations of the case and all the integrability conditions are taken into account. If during this simplification process the system proves to be inconsistent (say arriving at 1=0) an empty list is returned meaning the system has no solution. To illustrate this consider the following equations

p1diffvx&comma;y&comma;y+ux&comma;y

p1yvx&comma;y+ux&comma;y

(2)

p2diffux&comma;y&comma;xdiffux&comma;y&comma;y

p2xux&comma;yyux&comma;y

(3)
  

Indicate the dependent and independent variables and an ordering (ranking) for them. Because the purpose in this example is to only remove consequences and take into account integrability conditions, use an orderly ranking, that could be u&comma;v or v&comma;u (see DifferentialRing)

RDifferentialRingblocks=u&comma;v&comma;derivations=x&comma;y

Rdifferential_ring

(4)
  

Consider now the system constructed with p1 and p2 in this way:

sysexpandp1=0&comma;p2=0&comma;diffp1p2&comma;x=0&comma;diffp2&comma;y=0

sysyvx&comma;y+ux&comma;y=0&comma;xux&comma;yyux&comma;y=0&comma;2xyvx&comma;yxux&comma;y2xyvx&comma;yyux&comma;y+xux&comma;y2xux&comma;yyux&comma;y+yvx&comma;y2x2ux&comma;yyvx&comma;y2xyux&comma;y+ux&comma;y2x2ux&comma;yux&comma;y2xyux&comma;y=0&comma;2xyux&comma;y2y2ux&comma;y=0

(5)
  

By construction, we know that the third and fourth equations, diff(p1*p2, x) = 0 and diff(p2, y) = 0, bring no additional information. They are consequences of p1=0,p2=0 and so they are simplified away in the system returned

simplified_sysRosenfeldGroebnersys&comma;R

simplified_sysregular_differential_chain

(6)

Equationssimplified_sys

xux&comma;yyux&comma;y&comma;yvx&comma;y+ux&comma;y

(7)
  

Detection of inconsistencies in a differential system

  

Add now a third equation

p3diffvx&comma;y&comma;x+ux&comma;yy

p3xvx&comma;y+ux&comma;yy

(8)
  

What RosenfeldGroebner tells us about the system {p1, p2, p3}

RosenfeldGroebnerp1&comma;p2&comma;p3&comma;R

(9)
  

This result indicates that the system bears a contradiction; p1, p2, and p3 have no common solution. In this case it is possible to visually detect the contradiction:

diffp1=0&comma;xdiffp3=0&comma;yp2=0

1=0

(10)
  

The three equations cannot be satisfied simultaneously. The system is inconsistent; it has no solution.

Solving ordinary differential systems: Elimination ranking

  

To perform a chain resolution of a system of ordinary differential equations, use an elimination ranking. Consider the differential system defined by the following set S of differential polynomials in the unknown functions xt,yt,zt.

Sdiffxt&comma;txtxt+yt&comma;diffyt&comma;t+ytxt+yt&comma;diffxt&comma;t2+diffyt&comma;t2+diffzt&comma;t21

S&DifferentialD;&DifferentialD;txtxtxt+yt&comma;&DifferentialD;&DifferentialD;tyt+ytxt+yt&comma;&DifferentialD;&DifferentialD;txt2+&DifferentialD;&DifferentialD;tyt2+&DifferentialD;&DifferentialD;tzt21

(11)
  

Using an elimination ranking where zt>yt>xt, you obtain differential chains containing an equation in xt alone, an equation determining yt in terms of xt, and finally an equation determining zt in terms of yt and xt. Such an elimination ranking is represented by z&comma;y&comma;x

RDifferentialRingblocks=z&comma;y&comma;x&comma;derivations=t&colon;

GRosenfeldGroebnerS&comma;R

Gregular_differential_chain&comma;regular_differential_chain

(12)
  

In the result above we see two cases. The equations of the first case satisfying this elimination ranking mentioned are

EquationsG1&comma;solved

&DifferentialD;&DifferentialD;tzt2=&DifferentialD;&DifferentialD;txt42&DifferentialD;&DifferentialD;txt3xt2+2&DifferentialD;&DifferentialD;txt2xt4xt4xt4&comma;yt=&DifferentialD;&DifferentialD;txt+xt2xt&comma;&DifferentialD;2&DifferentialD;t2xt=2&DifferentialD;&DifferentialD;txtxt

(13)
  

The equations of the second case also satisfying the elimination ranking proposed are

EquationsG2&comma;solved

&DifferentialD;&DifferentialD;tzt2=yt4+1&comma;&DifferentialD;&DifferentialD;tyt=yt2&comma;xt=0

(14)

Constrained systems: Orderly and mixed rankings

  

Consider the set of equations describing the motion of a pendulum, that is, a point mass m suspended from a massless rod of length l under the influence of gravity g, in Cartesian coordinates x,y. The Lagrangian formulation leads to two second order differential equations with an algebraic constraint. T is the Lagrangian multiplier.

Pmdiffyt&comma;t&comma;t+Ttyt+g&comma;mdiffxt&comma;t&comma;t+Ttxt&comma;xt2+yt2l2

Pm&DifferentialD;2&DifferentialD;t2yt+Ttyt+g&comma;m&DifferentialD;2&DifferentialD;t2xt+Ttxt&comma;xt2+yt2l2

(15)
  

Orderly Ranking

• 

To find the lowest order equations vanishing on the zeros of the system modeling the pendulum and in particular all the algebraic constraints (that is, the equations of order 0), use an orderly ranking T&comma;x&comma;y

RDifferentialRingblocks=T&comma;x&comma;y&comma;derivations=t&comma;arbitrary=m&comma;l&comma;g&colon;

GRosenfeldGroebnerP&comma;R

Gregular_differential_chain&comma;regular_differential_chain

(16)

EquationsG1&comma;solved

&DifferentialD;&DifferentialD;tTt=3&DifferentialD;&DifferentialD;tytgl2&comma;&DifferentialD;&DifferentialD;tyt2=Ttyt2l2Ttl4+yt3gytl2gml2&comma;xt2=yt2+l2

(17)

EquationsG2&comma;solved

Tt=ytgl2&comma;xt=0&comma;yt2=l2

(18)
  

The second component in the output of RosenfeldGroebner corresponds to the equilibria of the pendulum. From the first component of the output, it can be seen that the actual motion is described by two first order differential equations with a constraint: the solution depends on only two arbitrary constants.

  

Mixed Ranking

  

To obtain the differential system satisfied by x and y alone, eliminate T with regards to x,y, so this ranking is T&comma;x&comma;y

RMDifferentialRingR&comma;blocks=T&comma;x&comma;y

RMdifferential_ring

(19)

GMRosenfeldGroebnerP&comma;RM

GMregular_differential_chain&comma;regular_differential_chain

(20)

EquationsGM1&comma;solved

Tt=&DifferentialD;&DifferentialD;tyt2ml2+yt3gytl2gyt2l2l4&comma;&DifferentialD;2&DifferentialD;t2yt=&DifferentialD;&DifferentialD;tyt2ytml2yt4g+2yt2l2gl4gyt2ml2ml4&comma;xt2=yt2+l2

(21)

EquationsGM2&comma;solved

Tt=ytgl2&comma;xt=0&comma;yt2=l2

(22)
  

Note here that the motion is given by a second order differential equation in yt. Then the Lagrange multiplier is given explicitly in terms of yt and its first derivative.

Solving partial differential system: Lexicographic ranking

  

To solve overdetermined systems of partial differential equations it is interesting to compute first the ordinary differential equations satisfied by the solutions of the system. For this purpose, a lexicographic ranking must be chosen. Consider the differential system S defining the infinitesimal generators of the symmetry group of the Burgers equations.

Burgers_pdediffux&comma;t&comma;x=diffux&comma;t&comma;t&comma;tux&comma;tdiffux&comma;t&comma;t

Burgers_pdexux&comma;t=2t2ux&comma;tux&comma;ttux&comma;t

(23)

BurgersSymmetryPDEtools:-DeterminingPDEBurgers_pde&comma;ux&comma;t&comma;ξ&comma;φ&comma;ηx&comma;t&comma;u

BurgersSymmetry3x3ξx&comma;t&comma;u=0&comma;tηx&comma;t&comma;u=2x2ξx&comma;t&comma;u2&comma;uηx&comma;t&comma;u=xξx&comma;t&comma;u2&comma;xηx&comma;t&comma;u=2x2ξx&comma;t&comma;uu2&comma;tφx&comma;t&comma;u=xξx&comma;t&comma;u2&comma;uφx&comma;t&comma;u=0&comma;xφx&comma;t&comma;u=xξx&comma;t&comma;uu2+ηx&comma;t&comma;u&comma;tξx&comma;t&comma;u=0&comma;uξx&comma;t&comma;u=0

(24)
  

Seek the ordinary differential equations with respect to t satisfied by the solutions of this system enclosing the dependent variables (in this example, all of them) using lex

RDifferentialRingblocks=lexη,φ,ξ&comma;derivations=u&comma;x&comma;t

Rdifferential_ring

(25)

GRosenfeldGroebnerBurgersSymmetry&comma;R

Gregular_differential_chain

(26)

EquationsG1&comma;solved

uηx&comma;t&comma;u=tφx&comma;t&comma;u&comma;uφx&comma;t&comma;u=0&comma;uξx&comma;t&comma;u=0&comma;xηx&comma;t&comma;u=tηx&comma;t&comma;uu&comma;xφx&comma;t&comma;u=tφx&comma;t&comma;uu+ηx&comma;t&comma;u&comma;xξx&comma;t&comma;u=2tφx&comma;t&comma;u&comma;2t2ηx&comma;t&comma;u=0&comma;2t2φx&comma;t&comma;u=0&comma;tξx&comma;t&comma;u=0

(27)
  

There are three independent ordinary differential equations with respect to t (the last three in the output above). All other ordinary equations in t vanishing on the solutions of BurgersSymm can be written as linear combination of these three and their derivatives.

Computing singular solutions

  

Consider the following Clairaut partial differential equation.

ux&comma;y=xdiffux&comma;y&comma;x+ydiffux&comma;y&comma;y+diffux&comma;y&comma;xdiffux&comma;y&comma;y

ux&comma;y=xxux&comma;y+yyux&comma;y+xux&comma;yyux&comma;y

(28)
  

Represent it by the equation p with a more compact syntax, called jet notation.

pu+xux+yuy+uxuy

pxux+yuy+uxuyu

(29)
  

Before any manipulation of this equation, define a differential polynomial ring it belongs to:

RDifferentialRingblocks=u&comma;derivations=x&comma;y

Rdifferential_ring

(30)
  

The Clairaut equation under consideration has a singular solution. This is shown in the output of RosenfeldGroebner by the existence of two cases: the general case (always exists unless the system is inconsistent) and an additional - therefore - singular case; each one represented by a regular_differential_chain

ClairautRosenfeldGroebnerp&comma;R

Clairautregular_differential_chain&comma;regular_differential_chain

(31)
  

The equations and inequations of each egular_differential_chain can be accessed using the Equations and Inequations commands. The general case is represented by the differential chain that involve some additional inequations while the second regular differential chain represents the singular case

EquationsClairaut1,InequationsClairaut1

xux+yuy+uxuyu,uy+x

(32)

EquationsClairaut2,InequationsClairaut2

xy+u,

(33)
  

Analysis of singular solutions

  

All singular cases are present in the output of RosenfeldGroebner while only the essential singular zeros are present when the option singsol = essential is used. For first order differential equations:

– 

The essential singular solutions are envelopes of the nonsingular (general) solutions; that is: they are not particular cases of the general solution.

– 

The other singular solutions are particular cases of the general solution.

• 

Illustrate this analysis on the equation Wp(cf. WeierstrassP):

Wpdiffwz&comma;z24wz3+g2wz+g3

Wp&DifferentialD;&DifferentialD;zwz24wz3+g2wz+g3

(34)
• 

Case g2327g32 :

RDifferentialRingblocks=w&comma;derivations=z&comma;arbitrary=g2&comma;g3

Rdifferential_ring

(35)

EquationsRosenfeldGroebnerWp&comma;R

&DifferentialD;&DifferentialD;zwz24wz3+g2wz+g3&comma;4wz3g2wzg3

(36)
  

There are three singular solutions: wz=ri, where ri is one of the roots of 4r3g2rg3. There are all three essential singular solutions as we can see using the option singsol = essential

EquationsRosenfeldGroebnerWp&comma;R&comma;singsol=essential

&DifferentialD;&DifferentialD;zwz24wz3+g2wz+g3&comma;4wz3g2wzg3

(37)
  

As such they are envelopes of the nonsingular (general) solution of Wp. This can also be observed by plotting the real zeros for g2=4,g3=−1:

singular_solutionsplotfsolve4r3+4r1&comma;r&comma;z=3..3&comma;color=red&comma;thickness=2&colon;

general_solution_sample_1seqplotsodeplotdsolvediffwz&comma;z&comma;z6wz2+2&comma;wic=1&comma;Dwic=1.0&comma;wz&comma;type=numeric&comma;3..3&comma;color=green&comma;ic=0..4&colon;

general_solution_sample_2seqplotsodeplotdsolvediffwz&comma;z&comma;z6wz2+2&comma;wic=2&comma;Dwic=5.0&comma;wz&comma;type=numeric&comma;ic0.15..ic+1.7&comma;color=blue&comma;ic=3..1&colon;

plotsdisplaysingular_solutionsuniongeneral_solution_sample_1uniongeneral_solution_sample_2&comma;title=g2 = 4, g3 = - 1\n red: singular solutions\n green and blue: samples of general solutions&comma;view=3..3&comma;32..3

• 

Case g23=27g32; parametrize the equation with g as in

Wpssubsg2=3g2&comma;g3=g3&comma;Wp

Wps&DifferentialD;&DifferentialD;zwz24wz3+3g2wz+g3

(38)

RsDifferentialRingR&comma;arbitrary=g&colon;

EquationsRosenfeldGroebnerWps&comma;Rs

&DifferentialD;&DifferentialD;zwz24wz3+3g2wz+g3&comma;2wz+g&comma;wzg

(39)
  

There are two singular solutions: wz=g and wz=g2 from which only wz=g is an essential singular solution of Wps as is shown using the option singsol = essential

EquationsRosenfeldGroebnerWps&comma;Rs&comma;singsol=essential

&DifferentialD;&DifferentialD;zwz24wz3+3g2wz+g3&comma;wzg

(40)
  

Therefore wz=g is an envelope of the nonsingular, general solution of Wps. On the contrary, the singular solution wz=g2 of Wps is not essential and therefore it is a limiting case of the general solution of Wps. Observe the situation for g=−1:

singular_solutionsplot1&comma;12&comma;color=red&comma;thickness=2&colon;

general_solution_sample_1plotseq1+32tanh12Czsqrt62&comma;C=1..1&comma;z=3..3&comma;color=green&colon;

general_solution_sample_2plotseq1+32coth12Czsqrt62&comma;C=1..1&comma;z=3..3&comma;color=blue&colon;

plotsdisplaysingular_solutions&comma;general_solution_sample_1&comma;general_solution_sample_2&comma;view=3..3&comma;32..3&comma;title=g2 = 3, g3 = - 1\n red: singular solutions\n green and blue: samples of general solutions

• 

Case g2=0,g3=0. The equation is:

Wp0subsg2=0&comma;g3=0&comma;Wp

Wp0&DifferentialD;&DifferentialD;zwz24wz3

(41)

R0DifferentialRingblocks=w&comma;derivations=z&colon;

EquationsRosenfeldGroebnerWp0&comma;R0

&DifferentialD;&DifferentialD;zwz24wz3&comma;wz

(42)

EquationsRosenfeldGroebnerWp0&comma;R0&comma;singsol=essential

&DifferentialD;&DifferentialD;zwz24wz3

(43)
  

So there is only one singular solution, wz=0, and it is not essential. Observe its limiting property with respect to the nonsingular (general) solution.

singular_solutionsplot0&comma;s=4..4&comma;color=red&comma;thickness=2&colon;

general_solution_sampleplotseq1zC2&comma;C=2..1&comma;z=4..4&comma;color=blue&comma;blue&comma;blue&colon;

plotsdisplaygeneral_solution_sample&comma;singular_solutions&comma;view=4..4&comma;12..2&comma;title=g2 = 0, g3 = 0\n red: singular solutions\n blue: samples of general solutions

A two steps strategy changing rankings

  

The following two examples illustrate a good strategy for computing a representation of the radical of the differential ideal generated by a given system sys, with respect to some target ranking. This strategy usually leads to simpler computations and fewer regular differential chains. It often applies, because, many differential polynomial system do actually define prime differential ideals. The strategy involves two steps.

• 

First, search an easy ranking (may involve trial an error), different from the target one, with respect to which, sys is almost a regular differential chain (if possible, a ranking with respect to which the leading ranks have degree 1). Call now RosenfeldGroebner with this easier rank to generate the ideal associated to the given system of equations.

• 

Second, run RosenfeldGroebner over the regular differential chain(s) computed at the first step, but now with respect to the target ranking.

Euler equations for an incompressible fluid

  

In this example the two steps strategy is applied to Euler equations for an incompressible fluid in two dimensions. We want to prove that the radical I of the differential ideal generated by this system is prime, and, has differential dimension zero.

• 

To prove that I is prime, by definition, it is sufficient to show that I can be presented by a single regular differential chain C, whose leading ranks have degree 1.

• 

To prove that I has differential dimension zero, it is sufficient to compute C with respect to an orderly ranking, and, to check that each dependent variable has a derivative among the leading derivatives of C [K73, chapter II, section 12, Theorem 6].

  

The target ranking is orderly, but a direct computation does not allow us to conclude.

unassignp&colon;

p1v1t+v1v1x+v2v1y+px&colon;

p2v2t+v1v2x+v2v2y+py&colon;

p3v1x+v2y&colon;

target_rankingderivations=x&comma;y&comma;t,blocks=v1&comma;v2&comma;p

target_rankingderivations=x&comma;y&comma;t,blocks=v1&comma;v2&comma;p

(44)

RDifferentialRingtarget_ranking&colon;

idealRosenfeldGroebnerp1&comma;p2&comma;p3&comma;R

idealregular_differential_chain&comma;regular_differential_chain&comma;regular_differential_chain

(45)
  

However, there exists an easy ranking, with respect to which, the RosenfeldGroebner algorithm does not split into cases. This easy ranking is not orderly with respect to the dependent variables but an elimination one (we use the keyword lex) and also the ordering for the independent variables changes from x&comma;y&comma;t to x&comma;y&comma;t, but it serves the purpose of our two-steps strategy. Let us perform the first step.

easy_rankingderivations=t&comma;x&comma;y,blocks=lexp,v1,v2

easy_rankingderivations=t&comma;x&comma;y,blocks=lexp,v1,v2

(46)

SDifferentialRingeasy_ranking&colon;

idealRosenfeldGroebnerp1&comma;p2&comma;p3&comma;S

idealregular_differential_chain

(47)
  

Let us, perform the second step, rewriting ideal using the original ranking R. The output proves that I is prime and has differential dimension zero.

idealRosenfeldGroebnerideal1&comma;R

idealregular_differential_chain

(48)

Tools:-LeadingRankideal

py,y&comma;v1x&comma;v1y&comma;v2x

(49)

A parametric dynamical system

  

This example illustrates the use of RosenfeldGroebner for studying parametric dynamical systems. Indeed, with respect to any orderly ranking on the state variables, a dynamical system is a regular differential chain, which describes a prime differential ideal.

  

Here, RosenfeldGroebner is used to compute the so-called input-output relation of a dynamical system. This relation is used, in control theory, to study the global identifiability, and, simplify the estimation of the unknown parameters from measures.

  

First, one defines a parametric dynamical system of two ODE. The two state variables x1 and x2 are grouped in one block. This ensures that the derivatives of x1 and x2 are ranked orderly. The parameters of the system are considered arbitrary

paramske&comma;Ve&comma;k1,2&comma;k2,1

paramske&comma;Ve&comma;k1,2&comma;k2,1

(50)

RDifferentialRingderivations=t&comma;blocks=x1&comma;x2&comma;arbitrary=params

Rdifferential_ring

(51)

edoAdiffx1t&comma;t=k1,2x1t+k2,1x2tVex1tke+x1t

edoA&DifferentialD;&DifferentialD;tx1t=k1,2x1t+k2,1x2tVex1tke+x1t

(52)

edoBdiffx2t&comma;t=k1,2x1tk2,1x2t

edoB&DifferentialD;&DifferentialD;tx2t=k1,2x1tk2,1x2t

(53)
  

This call to RosenfeldGroebner bundles the two ODEs in a regular differential chain. The parameters do not generate splitting cases because they were declared arbitrary when entering the differential ring

idealRosenfeldGroebneredoA&comma;edoB&comma;R

idealregular_differential_chain

(54)
  

The leading ranks of the regular differential chain have degree 1. The differential ideal is thus prime.

Tools:-LeadingRankideal1

&DifferentialD;&DifferentialD;tx1t&comma;&DifferentialD;&DifferentialD;tx2t

(55)
  

The input-output relation is thus obtained by eliminating x2. This is achieved by a change of ranking for the dependent variables over ideal, from x1&comma;x2 to x2&comma;x1, using RosenfeldGroebner. The input-output relation is the differential polynomial of IOideal whose leading derivative is a derivative of x1. Note that you only need to pass to RosenfeldGroebner the piece of information that is changing and there is no need to re-enter the information of the original differential ring, embedded in the first argument, the ideal to be rewritten

IOidealRosenfeldGroebnerideal1&comma;blocks=x2&comma;x1

IOidealregular_differential_chain

(56)

IOrelEquationsIOideal&comma;leader=derivativex1t&comma;notation=jet1

IOrelVekek2,1x1+Vek2,1x12+ke2k1,2x1t+ke2k2,1x1t+2kek1,2x1x1t+2kek2,1x1x1t+k1,2x12x1t+k2,1x12x1t+Vekex1t+ke2x1t,t+2kex1x1t,t+x12x1t,t

(57)

Optional arguments

Changing arbitrary, blocks and adding relations

  

During a call to RosenfeldGroebner it is also possible to change (replace or add to) the arbitrary and/or blocks stated when constructing the differential ring with DifferentialRing. You can also state relations for the arbitrary, allowing for a more detailed study of a system. As an example, consider again the problem of the singular cases of the Weierstrass differential equation,

Wpdiffwz&comma;z24wz3+g2wz+g3

Wp&DifferentialD;&DifferentialD;zwz24wz3+g2wz+g3

(58)
  

but this time take g2 as a variable of the problem and only g3 as arbitrary, so in the differential ring rank g2 within blocks, after w

RDifferentialRingblocks=w&comma;g2&comma;derivations=z&comma;arbitrary=g3

Rdifferential_ring

(59)
  

This problem now splits into three cases

idealRosenfeldGroebnerWp&comma;R

idealregular_differential_chain&comma;regular_differential_chain&comma;regular_differential_chain

(60)

Equationsideal&comma;solved

&DifferentialD;&DifferentialD;zwz2=4wz3g2wzg3&comma;wz3=g2wz4+g34&comma;wz=g2218g3&comma;g23=27g32

(61)
  

From the equations of each case we see that the new case exists only when the invariants g2 and g3 are related by g23=27g32, corresponding to the discriminant of the WeierstrassP function equal to zero. This is the same third singular case presented in a previous section above without derivation.

  

Suppose now you are interested in taking both g3 and g2 as arbitrary, or only g2, or none of them. For that purpose, use the option arbitrary and additionally. This adds g2 to the existing arbitrary set of the differential ring R

EquationsRosenfeldGroebnerWp&comma;R&comma;additionally&comma;arbitrary=g2&comma;solved

&DifferentialD;&DifferentialD;zwz2=4wz3g2wzg3&comma;wz3=g2wz4+g34

(62)
  

This makes only g2 arbitrary, replacing the arbitrary = g3 entered when stating the differential ring R

EquationsRosenfeldGroebnerWp&comma;R&comma;arbitrary=g2&comma;solved

&DifferentialD;&DifferentialD;zwz2=4wz3g2wzg3&comma;wz3=g2wz4+g34&comma;wz=3g32g2&comma;g32=g2327

(63)
  

This makes none of g2 and g3 arbitrary, overriding arbitrary = g3 in the differential ring R

EquationsRosenfeldGroebnerWp&comma;R&comma;arbitrary=&comma;solved

&DifferentialD;&DifferentialD;zwz2=4wz3g2wzg3&comma;wz3=g2wz4+g34&comma;wz=g2218g3&comma;g23=27g32&comma;wz=0&comma;g2=0&comma;g3=0

(64)
  

Suppose now you are interested in the cases when both g2 and g3 are arbitrary and satisfy certain relations, say the discriminant of the WeierstrassP function equal to zero, that is, 27g32=g23. Before proceeding note that as explained in the description, all the objects involved in these relations will automatically be taken as arbitrary (you can optionally pass arbitrary = [g2, g3]). So the input is

EquationsRosenfeldGroebnerWp&comma;R&comma;relations=27g32=g23&comma;solved

&DifferentialD;&DifferentialD;zwz2=4wz3g2wzg3&comma;g23=27g32&comma;wz=g229g3&comma;g23=27g32&comma;wz=3g32g2&comma;g23=27g32

(65)
  

Finally this results in g2 being added to the arbitrary objects declared when constructing the differential ring R and also states the relation 27g32=16 for g3

EquationsRosenfeldGroebnerWp&comma;R&comma;additionally&comma;arbitrary=g2&comma;relations=27g32=16&comma;solved

&DifferentialD;&DifferentialD;zwz2=4wz3g2wzg3&comma;g32=1627&comma;wz3=g2wz4+g34&comma;g32=1627

(66)

Requesting different attributes

  

The attributes = L option permits to modify the behavior of the function. Here, one omits the differential attribute. Therefore, the PDE system is regarded as a non - differential system. The decomposition is purely algebraic: the derivative ux,y occurring in the third polynomial, is not simplified by the first polynomial.

RDifferentialRingderivations=x&comma;y&comma;blocks=a&comma;v&comma;u&colon;

systux24u=0&comma;ux,yvyu+a=0&comma;vx,xux=0&colon;

idealRosenfeldGroebnersyst&comma;R&comma;attributes=normalized

idealregular_differential_chain

(67)

Equationsideal&comma;solved

a=ux,yvy+u&comma;vx,x=ux&comma;ux2=4u

(68)

Getattributes&comma;ideal

autoreduced&comma;primitive&comma;normalized

(69)

See Also

DifferentialAlgebra

DEtools[rifsimp]

PDEtools[casesplit]