arXiv subjects

Simon Wessing

Publications and source records attributed to Simon Wessing.

3 recordsLinked to original sources

arXiv2017

Experimental Analysis of a Generalized Stratified Sampling Algorithm for Hypercubes

Stratified sampling is a fast and simple method to generate point sets with uniform distribution in hypercubes. However, for the most common paraxial stratfication it has the prominent drawback that the number of sampled points in n dimensions has to be an n-th power of an integer number. This exponential growth makes its application unattractive or even infeasible in high dimensions. We present a stratification procedure that eliminates this problem by a recursive binary partitioning of the hypercube. The algorithm runs in linear time and tries to minimize the hyperboxes' deviation from the cubic shape. We analyze the properties of the algorithm using discrepancy and covering radius, and directly in practical applications, comparing it to quasirandom and other sampling methods. We also discuss a potential combination with Latin hypercube sampling, which positively affects discrepancy.

stat.CO↗

arXiv2017

The True Destination of EGO is Multi-local Optimization

Efficient global optimization is a popular algorithm for the optimization of expensive multimodal black-box functions. One important reason for its popularity is its theoretical foundation of global convergence. However, as the budgets in expensive optimization are very small, the asymptotic properties only play a minor role and the algorithm sometimes comes off badly in experimental comparisons. Many alternative variants have therefore been proposed over the years. In this work, we show experimentally that the algorithm instead has its strength in a setting where multiple optima are to be identified.

math.OC↗

arXiv2016

Towards a Systematic Development Process of Optimization Methods

The ultimate goal of all optimization methods is to solve real-world problems. For a successful project execution, knowledge about optimization and the application has to be pooled. As it is too inefficient to highly train one person in both fields, a team of experts is usually put together. Unfortunately, communication errors must be expected when several people collaborate. In this work, we deal with the avoidance and the repair of these communication errors. The tools proposed in this regard are, among others, the algorithm engineering cycle, checklists for structuring communication, and knowledge databases. The discussion is enriched with examples from continuous optimization, but most tools have a much wider applicability, even beyond optimization.

math.OC↗