Automatic and Interactive Tuning of Algorithms

Providing tools for algorithm tuning (and the related statistical analysis) is the main topic of this tutorial. This tutorial provides the necessary background for performing algorithm tuning with state-of-the-art tools. We will discuss pros and cons of manual, interactive, and automatic tuning of randomized algorithms such as Genetic Algorithms, Differential Evolution, Particle Swarm, and Evolution Strategies. Moreover, we highlight the important components of experimental work such as proper setup, visualization, and reporting and refer to the most prominent mistakes that may occur, giving examples for failed and successful experiments.