Comparison of Semidefinite Solvers for Topology Optimization of Cantilever Trusses Subject to Fundamental Eigenvalue Constraint

• Published in: Advanced Materials Research Vol. 1144; pp. 172 - 177
• Main Authors: Tyburec, Marek, Zeman, Jan
• Format: Journal Article
• Language: English
• Published: Zurich Trans Tech Publications Ltd 07.03.2017
• Subjects: Design optimization; Eigenvalues; Free vibration; Optimization; Solvers; Three dimensional models; Topology optimization; Trusses; Eigenvalues Constraints; Semidefinite Programming; Truss Topology Optimization; Ground Structure; Minimum Compliance; Free Vibrations
• ISBN: 3035710929; 9783035710922
• ISSN: 1022-6680; 1662-8985; 1662-8985
• DOI: 10.4028/www.scientific.net/AMR.1144.172

Optimal design of trusses is a widely developed field of structural optimization. In topology optimization the infinite design space is usually discretized into a finite ground structure defining the set of all feasible joints and bars. In this paper we adopt semidefinite programming (SDP) formulation of the topology optimization that enables us to obtain a globally optimal least compliant truss subjected to the constraints on structural volume and on the fundamental eigenvalue of free vibrations. For a model problem of 3D cantilever truss, we present a comparison of six solvers, namely CSDP, MOSEK, PENLAB, SDPA, SDPT3 and SeDuMi, from which it follows, that SDPA delivers the best computational performance.