Using Evolutionary Algorithms in the Conceptual Design of Selfoptimizing Systems

The machines of tomorrow will be capable of self-optimization. This means that they are able to adjust to changing surroundings and conditions on their own. New systems require new developmental methods and strategies. The approach, which is presented here, is based on evolutionary algorithms in order to design the principle solution of a self-optimizing system. The evolutionary algorithm produces the active structure automatically. The partial model shape, behavior and function can be derived from it. The representation for these self-optimizing systems and the approach, how to design a concept with evolutionary algorithms is, what is described in this article. The functionality of this method will be verified by an example. This contribution was developed in the course of the Collaborative Research Centre 614 “SelfOptimizing Concepts and Structures in Mechanical Engineering” (Speaker: Prof. Gausemeier) funded by the German Research Foundation (DFG) under grant number SFB 614 (see www.sfb614.de for further details).