English

A Triangle Algorithm for Semidefinite Version of Convex Hull Membership Problem

Optimization and Control 2019-05-21 v2 Computational Complexity

Abstract

Given a subset S={A1,,Am}\mathbf{S}=\{A_1, \dots, A_m\} of Sn\mathbb{S}^n, the set of n×nn \times n real symmetric matrices, we define its {\it spectrahull} as the set SH(S)={p(X)(Tr(A1X),,Tr(AmX))T:XΔn}SH(\mathbf{S}) = \{p(X) \equiv (Tr(A_1 X), \dots, Tr(A_m X))^T : X \in \mathbf{\Delta}_n\}, where Δn{\bf \Delta}_n is the {\it spectraplex}, {XSn:Tr(X)=1,X0}\{ X \in \mathbb{S}^n : Tr(X)=1, X \succeq 0 \}. We let {\it spectrahull membership} (SHM) to be the problem of testing if a given bRmb \in \mathbb{R}^m lies in SH(S)SH(\mathbf{S}). On the one hand when AiA_i's are diagonal matrices, SHM reduces to the {\it convex hull membership} (CHM), a fundamental problem in LP. On the other hand, a bounded SDP feasibility is reducible to SHM. By building on the {\it Triangle Algorithm} (TA) \cite{kalchar,kalsep}, developed for CHM and its generalization, we design a TA for SHM, where given ε\varepsilon, in O(1/ε2)O(1/\varepsilon^2) iterations it either computes a hyperplane separating bb from SH(S)SH(\mathbf{S}), or XεΔnX_\varepsilon \in \mathbf{\Delta}_n such that p(Xε)bεR\Vert p(X_\varepsilon) - b \Vert \leq \varepsilon R, RR maximum error over Δn\mathbf{\Delta}_n. Under certain conditions iteration complexity improves to O(1/ε)O(1/\varepsilon) or even O(ln1/ε)O(\ln 1/\varepsilon). The worst-case complexity of each iteration is O(mn2)O(mn^2), plus testing the existence of a pivot, shown to be equivalent to estimating the least eigenvalue of a symmetric matrix. This together with a semidefinite version of Carath\'eodory theorem allow implementing TA as if solving a CHM, resorting to the {\it power method} only as needed, thereby improving the complexity of iterations. The proposed Triangle Algorithm for SHM is simple, practical and applicable to general SDP feasibility and optimization. Also, it extends to a spectral analogue of SVM for separation of two spectrahulls.

Keywords

Cite

@article{arxiv.1904.09854,
  title  = {A Triangle Algorithm for Semidefinite Version of Convex Hull Membership Problem},
  author = {Bahman Kalantari},
  journal= {arXiv preprint arXiv:1904.09854},
  year   = {2019}
}

Comments

18 pages

R2 v1 2026-06-23T08:46:18.948Z