R arm of morie.fn.socpts: minimise \(c'x\) subject to
\(||A_i x + b_i||_2 \le e_i'x + d_i\) for each cone, the affine bound
(e_i, d_i) given in domains. Generalised-logarithm barriers,
damped Newton centering with backtracking, t multiplied by 10 until
m / t < eps; phase I \(\min s\) with the bounds relaxed by
\(s\) finds a strictly feasible start.
Usage
Socpts(c, A, b, domains, x0 = NULL, eps = 1e-10)
Arguments
- c
Objective.
- A
List of cone matrices.
- b
List of cone offsets.
- domains
List of list(e, d) affine right-hand sides.
- x0
Optional phase I start.
- eps
Duality-gap tolerance.
Value
List with x, objective, lhs, rhs,
slack, feasible.
References
Boyd, S. and Vandenberghe, L. (2004). Convex Optimization.
Cambridge University Press.
Lobo, M. S., Vandenberghe, L., Boyd, S. and Lebret, H. (1998).
Applications of second-order cone programming. Linear Algebra and its
Applications 284, 193-228.
