DynReAct PDP aims to strengthen the European steel industry by developing an advanced production planning system comprising the full exploitation of all available information coming from each plant involved to generate resource-optimized production plans for products of the highest reachable quality.
Therefore, it is foreseen to develop an open-source platform for dynamic production planning [RES1] able to provide valid routings for the whole production process across all plants in a reasonable time scale. Realized as service-oriented architecture (SOA) this platform enables a modular connection of multiple planning relevant components (see Figure).
The backbone of this platform forms the hierarchical planning system [RES2] developed in the former DynReAct project combining three different planning horizons to enable dynamic reactions to unforeseen events, such as breakdowns and order cancellations, while continuing to consider optimized mid-term and long-term planning strategies.
To improve the planning precision as well as the useability of the platform three further components will be implemented based on the DynReAct platform services:
- A Recommender system [RES3] being an intelligent planning monitoring system that provides autonomously recommendations to the planner in-case of suboptimal planning situations.
- A Coil-order-allocation system [RES4] providing optimized allocations of raw material to orders.
- A plant performance modelling system [RES5] providing real-time estimations of plant performances concerning multiple aspects as use of resources or product quality.
An industrial DynReAct prototype for production scheduling will be realized [RES6] at the multi-route processing chain of thyssenkrupp Rasselstein to demonstrate the applicability of the DynReAct platform in industrial practise that should reach TRL 7-8 at the end of the DynReAct PDP pilot and demonstration project.
Along with the software implementation we intend to publish a specification of the dynamic planning architecture and interfaces [RES7]. This will enable the development of interoperable software interfaces on different platforms, and it can also serve as a starting point for a formal standardization process. The main findings of the project, including benefits realized with the industrial prototype implementation and potential for transfer to other sectors, will be summarized in a whitepaper that we aim to publish at the end of the project [RES8].