solve_QP1QC: Solve (non-convex) quadratic program with 1 quadratic...

In aaronjfisher/qp1qc: Solves a possibly non-convex quadratic program with 1 quadratic constraint

Description

Solves a possibly non-convex quadratic program with 1 quadratic constraint. Either A_mat or B_mat must be positive definite, but not necessarily both (see Details, below).

Usage

solve_QP1QC(A_mat, a_vec, B_mat, b_vec, k, tol = 10^-7,
  eigen_B_mat = NULL, eigen_A_mat = NULL, verbose = TRUE)

Arguments

A_mat	see details below
a_vec	see details below
B_mat	see details below
b_vec	see details below
k	see details below
tol	a calculation tolerance variable used at several points in the algorithm.
eigen_B_mat	(optional) the precalculated result eigen(B_mat).
eigen_A_mat	(optional) the precalculated result eigen(A_mat).
verbose	show progress from calculation

Details

Solves a minimization problem of the form:

min_{x} x^T A_mat x + a_vec^T x

such that x^T B_mat x + b_vec^T x + k ≤q 0,

where either A_mat or B_mat must be positive definite, but not necessarily both.

Value

a list with elements

• soln - the solution for x

• constraint - the value of the constraint function at the solution

• objective - the value of the objective function at the solution

aaronjfisher/qp1qc documentation built on Aug. 22, 2020, 2:35 p.m.