QPSolve:
PROC [c: RVector,
A: Matrix, lobd: RVector, x: RVector, iVar: IVector, nFR:
NAT]
RETURNS [minFR:
NAT]
Solve simple Quadratic Program: minimize ý xTx + cTx, subject to Ax = 0 and x e lobd.
Iterations start with initial feasible vector x, with x[i iFX] fixed on their bounds and the rest, x[i iFR], free. Elements [0..nFR) of iVar store iFR, and [nFR..n) store iFX.