Multiscale annealing and robustness: Fast heuristics for large scale nonlinear optimization
Large scale (generally nonlinear) optimization problems correspond to many applications; see e.g. neural networks, pattern recognition etc. Evidently, these optimization problems are complicated. However, many of them can be phrased in a grouping, clustering or quantization framework, which result in nonlinear optimization problems. The paper deals with a stochastic version of annealing techniques. In the paper, after introduction, first the basic notion are recalled and equations for multiscale optimization for different grouping criteria are formally derived. The, first part is followed by the case when an unknown distribution must be replaced by empirical date. The large deviation technique is employed in this part. The paper is completed by numerical example and bibliographic notes. For the entire collection see [Zbl 0971.00007].
This page was built for publication: Multiscale annealing and robustness: Fast heuristics for large scale nonlinear optimization
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2762582)
