Objective function

mpini · May 7, 2014, 06:01

Hi all,

What is the fastest way to implement a new objective function?

What about the continuous adjoint in this case? How to update it accordingly?

Thanks,
Matteo

economon · June 5, 2014, 15:42

Hi Matteo,

The derivation of the continuous adjoint equations and suitable boundary conditions (and resulting surface sensitivity) depends on the particular objective function chosen (and not all are mathematically possible). This is a very interesting research topic for us, as it requires some mathematical development for obtaining new continuous adjoint formulations. We are always interested in new objective functions. For example, we're working on temperature-based functionals and off-body functionals for minimizing sonic boom noise.

So, unfortunately, there is no short answer to your question, but you can check out the following subroutine in solver_adjoint_mean.cpp if you want to see how some of the current objective functions are included: void CAdjEulerSolver::SetForceProj_Vector(CGeometry *geometry, CSolver **solver_container, CConfig *config).

Perhaps we can discuss this more in the future (possibly in person). Cheers,
Tom

May 7, 2014, 06:01	Objective function	#1
mpini New Member Matteo Pini Join Date: May 2014 Posts: 2 Rep Power: 0	Hi all, What is the fastest way to implement a new objective function? What about the continuous adjoint in this case? How to update it accordingly? Thanks, Matteo

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
channelFoam for a 3D pipe	AlmostSurelyRob	OpenFOAM	3	June 24, 2011 13:06
latest OpenFOAM-1.6.x from git failed to compile	phsieh2005	OpenFOAM Bugs	25	February 9, 2010 04:37
Compilation errors in ThirdPartymallochoard	feng_w	OpenFOAM Installation	1	January 25, 2009 06:59
Problem with compile the setParabolicInlet	ivanyao	OpenFOAM Running, Solving & CFD	6	September 5, 2008 20:50
Please help about the VTKFoam	liugx212	OpenFOAM Running, Solving & CFD	0	November 18, 2005 18:27

June 5, 2014, 15:42		#2
economon Super Moderator Thomas D. Economon Join Date: Jan 2013 Location: Stanford, CA Posts: 271 Rep Power: 14	Hi Matteo, The derivation of the continuous adjoint equations and suitable boundary conditions (and resulting surface sensitivity) depends on the particular objective function chosen (and not all are mathematically possible). This is a very interesting research topic for us, as it requires some mathematical development for obtaining new continuous adjoint formulations. We are always interested in new objective functions. For example, we're working on temperature-based functionals and off-body functionals for minimizing sonic boom noise. So, unfortunately, there is no short answer to your question, but you can check out the following subroutine in solver_adjoint_mean.cpp if you want to see how some of the current objective functions are included: void CAdjEulerSolver::SetForceProj_Vector(CGeometry geometry, CSolver solver_container, CConfig config). Perhaps we can discuss this more in the future (possibly in person). Cheers, Tom