Lls2qp

Purpose

Converts an lls problem to a new problem based on the formula below. Only the objective function is affected. The

problem can be of any type with an LLS objective.

${\begin{array}{lc}\min \limits _{x}&f(x)={\frac {1}{2}}||Cx-d||=\\&0.5*(y-Cx)'*(y-Cx)=0.5*(y'y-2y'Cx+x'C'Cx)=\\&0.5*y'y+0.5*x'Fx+c'x\\\\&F=C'C\\&c'=-y'C\\&const={\frac {1}{2}}y'y\\\end{array}}$

Calling Syntax

qpProb = lls2qp(Prob, IntVars)

Inputs

Input	Description
Prob.LS.C	The linear matrix in 0.5 * ||y - Cx||.
Prob.LS.y	The constant vector in 0.5 * ||y - Cx||.

Outputs

Output	Description
qpProb	The converted problem.

Description

If the problem is a linear least squares problem a qp problem is created. The new problem may have integer variables. Create the problem with llsAssign then use this routine.

If the problem has nonlinear constraints an nlp is created. The new problem may have integer variables. Create the problem with conAssign or minlpAssign, the set the fields Prob.LS.C and Prob.LS.y