Systematischer Entwurf hierarchisch-hybrider Regelsysteme

在化学工业中，不连续操作的工厂广泛用于生产精细或特种化学品。一个典型的任务是在保证安全和质量下限的情况下实现利润最大化。为了达到这个规格，必须适当地操纵控制输入信号。从理论的角度来看，这是一个高度非平凡的控制问题，因为一些输入是实值或连续值信号，而其他输入只能取有限数量的离散值。在实践中，以食谱为基础的方法占上风。它们意味着确定一个固定的配方，该配方提供了关于每个所需产品的定义良好的任务的顺序和持续时间的信息。在通常称为“调度”的后续步骤中，必须将这些任务分配给可用的资源。这种方法有两个重要的缺点：解决方案显然远非最佳；此外，由于该方法是开环的，它只能在没有不可预见的干扰的情况下工作。拟建项目的主题是研究这类电厂的替代控制策略。特别是，通过将基于优化的方法与基于混合系统理论的基于抽象的方法相结合，我们期望能够计算出在控制复杂性的同时显著提高性能的闭环控制策略。

In the chemical industries discontinuously operated plants are widely used for the production of fine or specialty chemicals. A typical task is to maximize profit while guaranteeing safety and certain lower bounds for quality. To achieve this specification the control input signals have to be suitably manipulated. From a theoretical point of view this is a highly nontrivial control problem as some of the inputs are real- or continuous-valued signals, while others can only take a finite number of discrete values. In practice recipe-based approaches prevail. They imply determining a fixed recipe, which provides information about the sequence and duration of well-defined tasks for every desired product. In a subsequent step commonly called "scheduling" these tasks have to be assigned to available resources. This approach has two important shortcomings: solutions are clearly far from optimal; moreover, as the approach is open-loop, it will only work in the absence of unforseen disturbances. The topic of the proposed project is to investigate alternative control strategies for this type of plant. In particular, by combining an optimization based approach with an abstraction based method from hybrid systems theory we expect to be able to compute closed-loop control strategies that significantly improve performance while keeping a lid on complexity.