Nikolaus Hansen

Inria (research centre Saclay – Île-de-France)
Randomised Optimisation group (RandOpt)
Centre for Applied Mathematics (CMAP, UMR 7641)
Ecole polytechnique, Institut Polytechnique de Paris
Location: Campus d'Ecole polytechnique (static map), Laboratories building wing 0 (entrance north-east, south-west), room 00 20 05 (2nd floor)
91128 Palaiseau, France

mail	to

Research objective. We want to understand stochastic search (AKA randomized optimization) in continuous, high-dimensional search spaces. When improving known or developing new algorithms, we strongly care about whether and how they are useful in practice. Typical research means are statistical machine learning techniques, Markov chain analysis, meaningful comparison methodologies, and adaptive parameter control, eventually bridging the gap to the practitioners.

• Publications
  • via the HAL open archive (from 2008 only, most with pdf)
  • via the HAL open archive by year (from 2008 only, most with pdf)
  • via Google Scholar
  • via DBLP (Digital Bibliography and Library Project with many further links)
  • via a manually updated list (most with pdf, some with source code)
• Talks (seminars, tutorials,... most including slides)
• CMA-ES (Covariance Matrix Adaptation Evolution Strategy)
  • A written tutorial (arXiv:1604.00772)
  • A short introduction with further links and a commented publication list
  • A longer introduction on Wikipedia
  • Source code page
  • The mathematical foundation of CMA-ES can be found from information geometry, see Information-Geometric Optimization Algorithms: A Unifying Picture via Invariance Principles (arXiv:1106.3708)
  • Related stuff
    • On the CMSA (Covariance Matrix Self-Adaptation) Evolution Strategy (2012)
    • On self-adaptation and derandomized self-adaptation (2002)
• Benchmarking continuous optimization algorithms
  • The COCO platform (COmparing Continuous Optimizers) for benchmarking real-parameter black-box optimization algorithms (new code at github)
  • Slides (4MB): A few overview results from the GECCO BBOB workshops.
  • Two galleries of all results from 2013.
  • BBOB-2010: Black-Box Optimization Benchmarking of real-parameter optimization algorithms for GECCO 2010
  • BBOB-2009: Black-Box Optimization Benchmarking of real-parameter optimization algorithms for GECCO 2009
  • Paper (pdf): Comparison of 31 Algorithms on the COCO testbed (BBOB-2009)
  • CEC-2005: Comparison of Evolutionary Algorithms for the 2005 IEEE Congress on Evolutionary Computation

---

last change $Date: 2019-04-05 $