Genetic Algorithm Embedded into a Quality-Oriented Workflow of Methods for the Development of a Linear Drive used in Intralogistic Systems

Abstract

This paper presents the results of the DFG-project (Deutsche Forschungsgemeinschaft) Q-ELF (“Qualitätsorientierter Methodenworkflow für die Produktneuentwicklung eines Linearantriebs in der Fördertechnik”) carried out in cooperation of the TU Dortmund University (support code KU 1307/12-1) with the BUW Wuppertal (support code WI 1234-11/1). The project continues the former project SFB 696 (Sonderforschungsbereich) regarding the Demand Compliant Design (DeCoDe) and the corresponding system model that strengthens the knowledge management to create high-quality mechatronical systems. In contrast to the SFB, which comprised the reverse engineering of a belt conveyor, Q-ELF applied a workflow of methods for quality oriented development on a new product. The DeCoDe ensures a methodical development that connects different engineering domains. This connection is important because the most problems and malfunctions arise at the interface of different domains due to their different notations for example. This approach also enables a methodical comparison of different competing concepts to pick the best suited one. A genetic algorithm is presented to further decrease the design-space. The project was carried out to develop linear drives for intralogistic systems.

Archives of Control Sciences

“…The hybrid optimization algorithm (Fig. 2), which is a combination of genetic and gradient Newton algorithms was used to minimize the objective function [1,9,15,[17][18][19]24]. The results of calculations of λ and F i using the genetic algorithm are the starting point for the gradient algorithm.…”