Definitions of D_c and D_d in the augmented system

[pbmadabhushi]

I'm curious to know what D_c and D_d in IPOPT code (see screenshot) map to in the IPOPT implementation paper? Any direction will be helpful.

[svigerske]

They do not seem to appear in (13) in the paper.

But as far as I see, they are also set to zero in the implementation: NULL is passed for D_c and D_d in

Ipopt/src/Algorithm/IpPDFullSpaceSolver.cpp

Lines 520 to 522 in 391b771

	retval = augSysSolver_->Solve(&W, 1.0, &sigma_x, delta_x, &sigma_s, delta_s, &J_c, NULL, delta_c, &J_d,
	NULL, delta_d, *augRhs_x, *augRhs_s, *rhs.y_c(), *rhs.y_d(), *sol->x_NonConst(), *sol->s_NonConst(),
	*sol->y_c_NonConst(), *sol->y_d_NonConst(), check_inertia, numberOfEVals);

And NULL means a 0-matrix here.

Maybe there was some thought once that they could be useful.