simpleqp

Matlab implementation of Sequential coordinate-wise algorithm.

Objective

The quadratic programming (QP) task with the non-negativity constraints is

Minimize f(x) = 0.5 * x'*H*x + x'*f w.r.t. x
subject to the constraint x >= 0.

where, H is symmetric and positive semidefinite N x N matrix and f is a N dimensional column vector.

Usage

[x,fval] = qpsca(H,f)

Example

H = [1 -1; -1 2];
f = [-2; -6];
[x,fval] = qpsca(H,f)

x =

    9.9688
    7.9844
    
fval =

    -33.9998

Comparision to QuadProg

H = [1 -1; -1 2];
f = [-2; -6];
A = [-1 0; 0 -1];
b = [0; 0];
[x,fval] = quadprog(H,f,A,b)

x =

    10
    8
    
fval =

    -34