Using View-Based Architecture Descriptions to Aid in Automated Runtime Planning for a Smart Factory

Abstract

Smart factories are highly flexible production sites, that can adapt to fulfill a variety of production needs. Particularly, this shall allow for fulfilling production orders that are unknown during the design phase of the factory. To this end, smart factories reconfigure themselves during runtime to allow the execution of new and previously unknown production orders while aiming for an optimal use of resources. In the adaption a variety of factors must be taken into account, not only can the supply chain between different machines be changed to allow performing production steps in a new sequence (as needed by the production order), but the individual machines also can adapt themselves to exhibit different capabilities. A challenging task for such smart factories is the examination whether a certain production order is producible and which configuration is optimal for fulfilling this production order. To determine the optimal configuration among a huge variety of reconfiguration possibilities, many pieces of information must be taken into account. For instance, the capabilities of the machines, their workload, the possible sequences of production steps or constraints such as time and costs. To cope with the complexity of smart factory production planning, we employ a view-based engineering approach for the development of embedded systems. This paper contributes a report of the application of the view-based architecture descriptions to the engineering of a smart factory and illustrates that this approach also has considerable benefits for production planning within a smart factory.