QP.Solve: Quick Quadratic Programming Solver.

In fnoorian/quadprogpp: Quick Quadratic Programming Solver in C++

Quick Quadratic Programming Solver.

Description

QP.Solve solves a quadratic programming in form of:

min	1/2 x^\mathrm{T} G x + g_0^\mathrm{T} x
s.t.
	C_E^\mathrm{T} x + c_{e0} = 0
	C_I^\mathrm{T} x + c_{i0} \ge 0

Usage

QP.Solve(G, g0, 
         CI = matrix(0, length(g0), 0), ci0 = vector(),
         CE = matrix(0, length(g0), 0), ce0 = vector())

Arguments

G	n by n matrix appearing in the quadratic function to be minimized.
g0	vector on length n appearing in the quadratic function to be minimized.
CI	n by m constraints matrix. Can be an empty matrix.
ci0	constraints constants, with size m. Can be an empty vector.
CE	n by p equalities matrix. Can be an empty matrix.
ce0	equalities vector, with size p. Can be an empty vector.

Value

A vector containing the solution of the quadratic programming problem.

See Also

quadprog.solve.QP, quadprog package

Examples

##
## Assume we want to minimize: 1/2 x^T x + d^T %*% x
## under the constraints:      A^T x + b0 >= 0
## with d = (0,-5,0)^T
##     b0 = (8,-2,0)^T
## and 
##          (-4  2  0) 
##      A = (-3  1 -2)
##          ( 0  0  1)
## we can use QP.Solve as follows:
##
Dmat <- diag(3)
dvec <- c(0,-5,0)
CI   <- matrix(c(-4,-3,0,2,1,0,0,-2,1),3,3)
ci0  <- c(8,-2,0)
QP.Solve(Dmat, dvec, CI, ci0)

## This is comparable to using solve.QP from quadprog package:
## solve.QP(Dmat, -dvec, CI, -ci0)

fnoorian/quadprogpp documentation built on Aug. 21, 2024, 2:02 a.m.