dolfin team mailing list archive

[Anders Logg <question142516@xxxxxxxxxxxxxxxxxxxxx>]

Question #142516 on DOLFIN changed:
https://answers.launchpad.net/dolfin/+question/142516

    Status: Open => Answered

Anders Logg proposed the following answer:
B1;2600;0cOn Sun, Jan 23, 2011 at 01:10:28AM -0000, Patricio Jara wrote:
> New question #142516 on DOLFIN:
> https://answers.launchpad.net/dolfin/+question/142516
>
> I would like to include a complex number in the definition of the bilinear form.
>
> For example in the classical toy problem (Poisson's Equation) -Au=f
> where A is the Laplace operator, the coercive bilinear form is
> a(u,v)=(f,v). Moreover, a(u,v)+b(u,v) is also coercive and solves
> the problem -(A+bI)u=f where b is any complex number with positive
> real part. However, every time I add a complex number in the
> definition of the bilinear form, fenics gives me an errror. Perhaps
> I do not know how to do it correctly.  Thanks

Complex numbers are not supported, but there's any easy work-around.

Just write your equation as a system of two coupled real-valued PDEs
by equating the real and imaginary parts of your equation,
respectively. Use a DOLFIN mixed function space to define your trial
and test functions.

--
Anders

-- 
You received this question notification because you are a member of
DOLFIN Team, which is an answer contact for DOLFIN.