Randomized search heuristics form a widely applied class of black-box optimization techniques, designed with the hope to produce high-quality solutions for a great variety of problems and without requiring any problem-specific knowledge. Many different types of randomized search heuristics exist. After discussing their main working principles, we will describe the state of the art in analyzing randomized search heuristics by theoretical means. We will highlight a few examples that demonstrate how these analyses have influenced the design of new algorithmic ideas. To stimulate a discussion with the audience of this seminar, we will then summarize where the current challenges are.
