A full Nesterov–Todd step infeasible-interior-point algorithm for Cartesian P*(κ) horizontal linear complementarity problems over symmetric cones. (1st February 2016)

• Record Type: Journal Article
• Authors: Mohammadi, N.; Mansouri, H.; Zangiabadi, M.; Asadi, S.
• Is Part Of: Optimization. Volume 65:Number 2(2016)
• Page Start: 539
• Page End: 565
• Publication Date: 2016-02-01
• Subjects: interior-point methods -- horizontal linear complementarity problems -- Euclidean Jordan algebras -- symmetric cones; Mathematical optimization -- Periodicals; 519.7
• DOI: 10.1080/02331934.2015.1062011
• Languages: English
• ISSNs: 0233-1934

Abstract

Abstract : Euclidean Jordan algebra is a commonly used tool in designing interior-point algorithms for symmetric cone programs. In this paper, we present a full Nesterov–Todd (NT) step infeasible interior-point algorithm for horizontal linear complementarity problems over Cartesian product of symmetric cones. Since the algorithm uses only full-NT feasibility and centring steps, it has the advantage that no line searches are needed. The complexity result obtained here for symmetric cones using NT directions coincides with the best bound obtained for horizontal linear complementarity problems.