~libadjoint/dolfin-adjoint/improved-optimisation : changes

~libadjoint/dolfin-adjoint/improved-optimisation » Changes from revision 359

From Revision 359 to 340

Rev	Summary	Authors	Date
359	merge from trunk merge from trunk	David Ham	11 years ago
358	Minimize/maximize now return the optimal control Minimize/maximize now return the optimal control	Simon Funke	11 years ago
357	- Add a maximization function - Add a maximization function - Minor fixes	Simon Funke	11 years ago
356	- optimisation -> optimization (follow the standard in progr... - optimisation -> optimization (follow the standard in programming) - bugfix for scipy.slsqp method, where the parameter full_output = True caused the optimisation modele to crash - new "scale" parameter which can be used to rescale the problem	Simon Funke	11 years ago
355	Add support for the CG and BFGS algorithms in scipy Add support for the CG and BFGS algorithms in scipy	Simon Funke	11 years ago
354	Add support for the truncated Newton algorithm in scipy Add support for the truncated Newton algorithm in scipy	Simon Funke	11 years ago
353	More documentation More documentation	Simon Funke	11 years ago
352	Correct the minimizer documentation Correct the minimizer documentation	Simon Funke	11 years ago
351	Add bounds support for multiple parameters to the optimisati... Add bounds support for multiple parameters to the optimisation module	Simon Funke	11 years ago
350	Add a callback to the reduced functional that can be used fo... Add a callback to the reduced functional that can be used for example to plot the controls at each optimisation iteration	Simon Funke	11 years ago
349	Merge from lp:~libadjoint/dolfin-adjoint/bug_1032291 Merge from lp:~libadjoint/dolfin-adjoint/bug_1032291	Simon Funke	11 years ago
348	Another fix to avoid PETSC ERRORS Another fix to avoid PETSC ERRORS	Simon Funke	11 years ago
347	Fix a hidden python reference bug that caused random crashes... Fix a hidden python reference bug that caused random crashes/values in the ReducedFunctional evaluations	Simon Funke	11 years ago
346	First attempt to fix a bug with the optimisation and multipl... First attempt to fix a bug with the optimisation and multiple parameters	Simon Funke	11 years ago
345	Remove another unnecessary argument from the optimize functi... Remove another unnecessary argument from the optimize function. The minimal call for an optimisation solve is now: >> optimize(reduced_functional) Also added support for optimizing multiple parameters.	Simon Funke	11 years ago
344	Some more tidying up Some more tidying up	Simon Funke	11 years ago
343	And thats it - all tests pass again And thats it - all tests pass again	Simon Funke	11 years ago
342	The optimal control mms test passes again The optimal control mms test passes again	Simon Funke	11 years ago
341	Major simplifications to the optimisation API. Major simplifications to the optimisation API. Instead of taking in the functional and a parameter, it now expects a ReducedFunctional object that contains this information. This reduced functional can be automaticall created by using the new reduced_functional class which reuses the existing annotation of the forward model to evaluate the functional. Only the optimisation/ test passes so far, more will follow	Simon Funke	11 years ago
340	A new "ReducedFunctional" class that, given a functional and... A new "ReducedFunctional" class that, given a functional and a parameter, automatically creates an evaluate function that evaluates the functional value for given parameters. This works by replacing the parameter value in the existing annotation and then replaying it.	Simon Funke	11 years ago

Older »