Soft-computing algorithms as a tool for the planning of cyclically interlinked production lines

The planning and configuration of complex flexible production lines—cyclically interlinked systems—cover numerous aspects and various possibilities. It is not possible to determine the applicability and characteristics for each of these variations by applying already existing planning methods. This paper outlines an innovative approach to this problem, which is based on the application of learning algorithms such as artificial neural networks. The iterative design process of a production line is initiated by a customer request. An initial layout is usually planned by an engineer and considered in a forecast system, which classifies the layout as stable or unstable. In case of stable layouts, the performance and cost values are visualized with operating figures. On this basis, the engineer is able to evaluate the layout. If the result is not satisfactory, new production line layouts are generated by modifying the overruled layout and another iteration begins.