I am trying to solve this pde numerically with Mathematica:

NDSolve[{2 f[x, y] + I*D[f[x, y], y, y] - I*D[f[x, y], x, x] - 
2 x*D[f[x, y], y] + 2 x*D[f[x, y], x] + 2 y*D[f[x, y], y] - 
2 y*D[f[x, y], x] + 2*f[x, y]*x^2 - 4 f[x, y]*x*y + 
2*f[x, y]*y^2 - 2 I*x*D[f[x, y], y] - 2 I*x*D[f[x, y], x] + 
2 I*y*D[f[x, y], y] + 2 I*y*D[f[x, y], x] == 0, 
f[x, 0] == Exp[-x], f[x, 1] == 1}, f, {x, 0, 1}, {y, 0, 1}]

However, even if I chop of parts of my equation to make it easier to solve, I keep getting the error message:

 NDSolve::femdpop: The FEMStiffnessElements operator failed.

What does this mean? I've found this post here, but I have not found the meaning of my error message.

As a side note, I am using Mathematica 10.

  • $\begingroup$ may be it does not like the complex part. (all those I's in there). Is this quantum mechanics problem? No beep sound is generated when the "I's" are removed. $\endgroup$
    – Nasser
    Mar 30, 2015 at 4:50
  • $\begingroup$ This means that the FEM had an issue with discretizing the PDE. I filed this as a bug but as @Nasser suggests it seems to have something to do with complex values. $\endgroup$
    – user21
    Mar 30, 2015 at 7:40
  • $\begingroup$ @Nasser Yes, this is a quantum mechanics problems. I'm trying to solve the equation for quantum Brownian motion numerically, with all constants set to 1 and by assuming that the solution is separable. $\endgroup$
    – user85503
    Mar 31, 2015 at 2:13
  • $\begingroup$ Since FEM is unable to discretize the PDE, is there any method which will solve the differential equation? $\endgroup$
    – user85503
    Mar 31, 2015 at 2:16

1 Answer 1


This bug has been fixed as of Mathematica 10.4.0.

enter image description here


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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