lp:~libadjoint/dolfin-adjoint/applications

Branch merges

Propose for merging

No branches dependent on this one.

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Related blueprints

586. By Patrick Farrell on 2014-01-07

Some fixes to the MPEC code.

585. By Patrick Farrell on 2013-12-20

Try to implement Bendsoe's description of the PBM method. Failing miserably.

584. By Patrick Farrell on 2013-12-20

Bizarre. We really can't get around this hump. Arc length continuation?

583. By Simon Funke on 2013-12-20

Remove the (t - 1)**2 * dx term from the Laplacian. It is not needed for non-zero initial guesses.

582. By Simon Funke on 2013-12-20

Some improvements to the Stokes topology solver:

1) (Re-)activate the Strong-Wolfe line search. Apart from being more rebust, this has the main advantage that it detects if the Newton step is not a descent direction in which case it immediately aborts. This avoids doing many Newton iterations on subproblem which diverge, something that happened regularly before.

2) Check if a Newton step produces an infeasible solution (i.e. not 0 < rho < 1). If this test fails we discard the Newton step and increase mu. This is mainly a safeguard check.

581. By Simon Funke on 2013-12-20

Some minor improvements.

580. By Patrick Farrell on 2013-12-19

Attempt continuation in mu to solve the Laplace topology optimisation problem. Haven't ran it for long enough to see if it works.

579. By Simon Funke on 2013-12-19

Implement Example4. Also converges nicely to the expected solution.

578. By Simon Funke on 2013-12-19

Add DirichletBC support for the line search method.

577. By Simon Funke on 2013-12-19

Switch to the custom Newton solver. Still works nicely.