Message-ID: <98059e39-8b68-52f1-e30a-c1c09f92bcb5@bluewin.ch>
Date: 2021-05-17T19:09:54Z
From: Pascal Kündig
Subject: Solving a quadratically constrained linear program with inital values
Hi everyone,
I'm looking for an R-function that solves a quadratically constrained
linear program of the form:
min(x) -\mu' x
subject to
x' \Sigma x <= s
1'x <= 1
-1'x <= -1
Ix <= u
-Ix <= -b
while considering a given starting value for the vector x.
The above problem results from a larger program of the same structure
and by setting the constraint that some elements of the solution vector
\tilde{x} of this larger program have to be 0 if they lie below a
certain threshold. The starting value for the vector x is therefore a
subvector of \tilde{x}. \Sigma is symmetric but not necessarily positive
definite.