Improvement of the Logistic Performance of Cluster-Oriented Decentralized Control in Material Flow Networks in Manufacturing

几年来，制造和物流业务的控制从集中式计划转向实时分散控制，由“智能对象”实现。关于分散控制方法，研究迄今主要集中在开发新的控制算法，提供一定的“智能”，以提高物流目标的实现。此外，还有很多努力来开发所需的技术，以实现分散控制，例如物联网或工业4.0。然而，物流系统的底层网络结构及其对物流绩效的影响尚未得到广泛研究，尽管拓扑结构必须与算法和技术沿着被认为是分散控制的关键成功因素。例如，高度连接的系统通常提供更多的路由选择，而稀疏连接系统中的路由路径必须在更多的对象之间共享。这两种现象有一个明显的影响，在物流系统的路由，因此需要考虑在这些系统的控制方法的设计。该项目的目标是开发一种基于网络的方法，将物流系统划分为集群或模块，这些集群或模块应该由不同类型的分散控制来管理，以改善物流关键数字的实现。

Since several years, the control of operations in manufacturing and logistics moves from centralized planning towards decentralized control in real-time, enabled by 'intelligent objects'. Regarding decentralized control approaches, research has so far focused primarily on developing new control algorithms which provide a certain 'intelligence' to improve the achievement of logistic targets. Beside, there was a lot of effort to develop the required technologies to enable decentralized control, such as the Internet of Things or Industry 4.0. However, the underlying network structure of material flow systems and its influence on logistic performance has not been extensively studied, although the topology has to be considered along with algorithms and technology as a crucial success factor of decentralized control. For example, highly connected systems usually provide much more routing alternatives, while in turn the routing paths in sparsely connected systems have to be shared among much more objects. Both phenomena have a distinct impact on routing in material flow systems and therefore need to be considered during the design of control approaches for these systems. The goal of this project is the development of a network-based approach to divide a material flow system into clusters or modules, which are supposed to be governed by distinct types of decentralized control, in order to improve the achievement of logistic key figures.