A Descent Algorithm for Minimizing Polyhedral Convex Functions

From MaRDI portal





The problem under consideration consists of minimizing a polyhedral convex function, i.e. \[ \min_{x\in {\mathcal X}}\max_{1\leq i\leq N}\{f^T_ix-\alpha_i\}, \] where \(f_i\in\mathbb{R}^P\), \(\alpha_i\in\mathbb{R}\), \(i=1,\ldots,N\), and \(\mathcal X\) is a polyhedral convex set. Special problem types are parameter estimation in the maximum norm, minimization of a piecewise linear function, and the rank regression problem. The authors use a continuation approach to construct a descent algorithm. Internal computations are based on orthogonal factorizations. The algorithm is then specified to the mentioned problem types and its finiteness is proved. Some numerical experiments based on randomly generated problems are reported.











This page was built for publication: A Descent Algorithm for Minimizing Polyhedral Convex Functions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3321363)