[Csdp] Problems

[Brian Borchers]

>I would like solve several problems in the form
>
>min 1/2 x^T F x
>
>suject to f(x)<= 0
>
>where f(x) = Phi(x)-c; and Phi = grad(f) . x and F is positive defined.
>
>in a c++ FEM elastoplasticity code.

I can't make any sense of your problem.  Is x a vector in R^n?  How can
a vector be "positive defined?"  Does f map R^n to R?  How can 
f(x)=grad(f(x))-c if the left hand side of the equation is a scalar and the
right hand side is a vector?

>is possible solve in efficient way this kind of problems using Sedumi?

After you've clarified your question, you might ask in the SeDuMi discussion
forum.  However, calling SeDuMi from within a C++ code would be extremely
difficult.  In any case, this mailing list is for discussion of the CSDP 
solver.