Abonnement à la biblothèque: Guest
TSFP DL Home Archives Comité de direction


Petros Koumoutsakos
Computational Science and Engineering Laboratory ETH Zurich CH-8092, Switzerland

Stefan Kern
Computational Science and Engineering Laboratory ETH Zurich CH-8092, Switzerland

Nikolaus Hansen
Computational Science and Engineering Laboratory, ETH Zürich Universitätsrasse 6, Zürich, CH-8092, Switzerland


Bioinpired optimization is concerned with the development and implementation of algorithms and devices based on our experience and understanding of nature. Fluid mechanics are one of the most prominent paradigms of this type of optimization as humanity has been always fascinated from the ways the majestic eagle exploits the wind and the wiggling sperm navigates in the seminal fluid. This fascination has led to mimesis and the development of engineering designs that immitate natural forms and functions. Besides mimicking the final design it is also possible to mimic the processes by which this is achieved leading to genetic algorithms and evolution strategies that can be cast into optimization problems. In this article we discuss bioinspired algorithms for flow optimization by describing some of the fundamental concepts of these techniques and by illustrating their advantages and drawbacks in selected case studies from our research. We discuss single and multi-objective optimization in noisy environments as they pertain to combustion in experimental test-rigs and turbomnachinery, and the use of evolutionary algorithms and local learning models for the optimization of expensive cost functions as applied to simulations of anguiliform swimmers.