2
$\begingroup$

I need to define a multivariable differential equation of the type

$$ \mathcal{D}=\Theta_1^2-z_1(-\Theta_1+\Theta_2+\Theta_3+\Theta_4)(1+\Theta_1+\Theta_2+\Theta_3+\Theta_4)\\ \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;-z_1(z_2+z_3+z_4)(1+\Theta_1+\Theta_2+\Theta_3+\Theta_4)(2+\Theta_1+\Theta_2+\Theta_3+\Theta_4) $$ where $\Theta_k=z_k\partial_{z_k}$.
I have tried to define each paranthesis as their own operator and compose all the operators with one another:

Θ[f_, k_] := k*D[f, k]
d1[k_, z1_, z2_, z3_, z4_] := (k # + Θ[#, z1] + Θ[#, z2] + Θ[#, z3] + Θ[#, z4]) &
d2[k_, z1_, z2_, z3_, z4_] := (Θ[#, z1] + Θ[#, z2] + Θ[#, z3] + Θ[#, z4] - 2 Θ[#, k]) &
D1[z1_, z2_, z3_, z4_] := (Θ[Θ[#, z1], z1]-z1 d2[z1, z1, z2, z3, z4]@d1[1, z1, z2, z3, z4]-z1 d1[1, z1, z2, z3, z4]@d1[2, z1, z2, z3, z4]) &
D1[z1, z2, z3, z4]@f[z]

Which produces the undesired output

(-z1)*((-z1)*(Derivative[0, 1][Θ][#1, z1] & ) + z2*(Derivative[0, 1][Θ][#1, z2] & ) + z3*(Derivative[0, 1][Θ][#1, z3] & ) + 
    z4*(Derivative[0, 1][Θ][#1, z4] & )) - z1*((2*#1 + Θ[#1, z1] + Θ[#1, z2] + Θ[#1, z3] + Θ[#1, z4] & ) + 
    z1*(Derivative[0, 1][Θ][#1, z1] & ) + z2*(Derivative[0, 1][Θ][#1, z2] & ) + z3*(Derivative[0, 1][Θ][#1, z3] & ) + 
    z4*(Derivative[0, 1][Θ][#1, z4] & )) + z1*(Derivative[1, 0, 0, 0][f][z1, z2, z3, z4] + z1*Derivative[2, 0, 0, 0][f][z1, z2, z3, z4])

How can one define a differential operator that is the product of two or more (generally) different differential operators?

Improve this question
$\endgroup$
1
  • 2
    $\begingroup$ See section "Some noncommutative algebraic manipulation" here. It is important that one avoid explicit Times for other than scalars since for example x and d/dx (as D[#,x]&) do not commute as operators. $\endgroup$
    – Daniel Lichtblau
    May 27 at 15:23

2 Answers 2

Reset to default
3
$\begingroup$

I too wish there was a nicer way to manipulate operators. For now, this will have to do:

opRules = Prepend[Join[Function[{Θk, zk}, (c_. Θk^j_.)[f][args__] /; 
                           FreeQ[c, Θ1 | Θ2 | Θ3 | Θ4] :> c Nest[zk D[#, zk] &, f[args], j]] @@@ 
                  Transpose[{{Θ1, Θ2, Θ3, Θ4}, {z1, z2, z3, z4}}],
                       Function[{Θp, zp, id}, (c_. Θp)[f][args__] /; 
                            FreeQ[c, Θ1 | Θ2 | Θ3 | Θ4] :> 
                            c zp Apply[Derivative, UnitVector[4, id[[1]]] +
                                                   UnitVector[4, id[[2]]]][f][args]] @@@
        Transpose[{Times @@@ Subsets[{Θ1, Θ2, Θ3, Θ4}, {2}],
                   Times @@@ Subsets[{z1, z2, z3, z4}, {2}], 
                   Subsets[Range[4], {2}]}]],
                  (c_ /; FreeQ[c, Θ1 | Θ2 | Θ3 | Θ4])[f][args__] :> c f[args]];

Through[Operate[Composition[Through,
   Expand[Θ1^2 - z1 (-Θ1 + Θ2 + Θ3 + Θ4) (1 + Θ1 + Θ2 + Θ3 + Θ4) - 
          z1 (z2 + z3 + z4) (1 + Θ1 + Θ2 + Θ3 + Θ4) (2 + Θ1 + Θ2 + Θ3 + Θ4)]], 
   f[z1, z2, z3, z4]]] /. opRules // Simplify
Improve this answer
$\endgroup$
1
$\begingroup$

You have to take care that the derivative operator is not evaluated too soon. For this you may use another name and rename it in the end. E.g. we may use DD instead of D.

We first define a helper function: t to ease the writing of Θk:

t[i_] := Symbol["z" <> ToString[i]] DD[#, sym[i]];

Then we write the operator with "DD":

op = t[1]^2 - 
  z1 (-t[1] + t[2] + t[3] + t[4]) (1 + t[1] + t[2] + t[3] + t[4]) - 
  z2 (z2 + z3 + z4) (1 + t[1] + t[2] + t[3] + t[4]) (2 + t[1] + t[2] +
      t[3] + t[4])

enter image description here

Finally we change "DD" to "D" and define a function:

op = Evaluate[op] & /. DD -> D

enter image description here

We may try our operator on a function f[z1, z2, z3, z4]:

op[f[z1, z2, z3, z4]]

![enter image description here

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.