Convert a LQ CUTEr problem into a matrix representation.
[Q,p,r,C,b,me] = cute2qp(N,cl,cu,equatn)
An integer, the dimension of the unknown vector.
A vector, the CUTEr contraints lower bounds (see csetup).
A vector, the CUTEr contraints upper bounds (see csetup).
A boolean array whose i-th component is %t if the i-th constraint is an equation and %f else (see csetup)
A symetric square matrix, the quadratic term of the objective.
A vector, the linear term of the objective.
A scalar, the constant part of the objective.
A matrix, the linear term of the constraints
A vector, the constant term of the constraints
An integer, the number of equality constraints.
If the current CUTEr problem is a linear quadratic problem, this fonction translate it into the LQ problem matrix form:
Which is suitable for the quapro
,
qld
and qpsolve
solvers.
probpath=TMPDIR+'/AVGASA'; sifdecode(get_sif_path()+'/sif/AVGASA.SIF',probpath) buildprob(probpath) //initialize problem routines data structures [XSTART,bl,bu,v,cl,cu,equatn,linear] = csetup(probpath+'/OUTSDIF.d', [%t %t %f]); [Q,p,r,C,b,me] = cute2qp(N,cl,cu,equatn) [xopt,lagr,info]=qld(Q,p,C,b,bl,bu,me) | ![]() | ![]() |