Christian Schoppmeyer

An environment for the efficient testing and implementation of robust NMPC

In the last years many research studies have presented simulation or experimental results using Nonlinear Model Predictive Control (NMPC). The computation times needed for the solution of the resulting nonlinear optimization problems are in many cases no longer an obstacle due to the advances in algorithms and computational power. However, NMPC is not yet an industrial reality as its linear counterpart is. Two reasons for this are the lack of good tool support for the development of NMPC solutions and the fact that it is difficult to ensure the robustness of NMPC to plant-model mismatch. In this paper, we address both these issues. The main contribution is the development of an environment for the efficient implementation and testing of NMPC solutions, offering flexibility to test different algorithms and formulations without the need to re-encode the model or the algorithm. In addition, we present and discuss the approach of multi-stage robust NMPC to systematically deal with the robustness issue. The benefits of our approach are illustrated by experimental results on a laboratory plant.

Efficient Scheduling of Batch Plants Using Reachability Tree Search for Timed Automata with Lower Bound Computations

Abstract In this paper we discuss an extension of a recent approach to solve batch scheduling problems using reachability analysis for timed automata (TA) by embedding lower bound computations in the reachability algorithm. We propose an extension of the embedded LP-formulation to handle scheduling problems with NIS policy and an improved minimum remaining processing time (MRPT) procedure to compute the lower bounds. The proposed bounding procedures are tested on typical batch scheduling problems. The comparative study shows that the MRPT-based bounding procedure is efficient and increases the overall performance significantly in comparison to the LP-based bounding procedure.

Modeling and Solving Batch Scheduling Problems with Various Storage Policies and Operational Policies using Timed Automata

Abstract In this contribution we present the conceptual ideas on modeling deterministic batch scheduling problems with various storage policies and operational policies using timed automata (TA) and on solving them using reachability analysis for TA. We propose a new reduction technique to improve the efficiency of the search algorithm for scheduling problems with non-intermediate storage policy. The performance of the proposed modeling approach and the reduction technique is evaluated by investigating it on benchmark instances from OR literature and on typical batch scheduling problems. The comparative study shows that the proposed modeling approach handles the crucial aspects of the storage policies and operational policies in straightforward way and the new reduction technique increases the overall performance of the solution algorithm.

Timed automata based scheduling for a miniature pipeless plant with mobile robots

In this contribution we present a conceptual idea on modeling a scheduling problem in a miniature pipeless plant with mobile robots by timed automata (TA) and solving it using reachability analysis. Two TA-based tools, TAOpt and UPPAAL, are evaluated on the miniature pipeless plant, and on job shop benchmark instances from the OR literature. The comparative study shows that the proposed modeling approach handles the crucial constraints of the miniature pipeless plant in an effective and straightforward way.

Effective Batch Scheduling with Sequence-dependent Changeovers Using Reachability Analysis of Timed Automata Combined with Lower Bound Computations

Abstract In this contribution we discuss an extension of a recent approach to solve batch scheduling problems using reachability analysis for timed automata (TA) with embedded lower bound computations. We propose two bounding procedures embedded in the reachability algorithm to handle scheduling problems with sequence-dependent changeovers: (i) a MILP formulation (originally proposed by Manne 1960) extended with additional constraints to model setup and changeover operations and (ii) an improved minimum remaining processing time (MRPT) procedure. The efficiency of the proposed bounding procedures is evaluated on job shop problems with sequence-dependent changeovers. The comparative study shows that the MRPT-based bounding procedure is efficient and increases the overall performance significantly in comparison to the MILP-based bounding procedure.

Embedding of Timed Automata-based Schedule Optimization into Recipe Driven Production

Abstract In this contribution, we discuss the conceptual idea of a hierarchical control system with an integration of schedule optimization and process control. We propose to embed timed automata (TA)-based schedule optimization into recipe driven production based on sequential control logic combined with interlocks in order to handle uncertainties in the operation on the different levels of the control hierarchy, depending on the type and on the impact of the uncertainty. The efficiency and robustness of the proposed operations management system is evaluated on a lab-scale multiproduct pipeless batch plant using different scenarios of uncertainties.

Optimal Management of Shuttle Robots in a High-Rise Warehouse Using Timed Automata Models

Abstract In warehousing systems, the speed of the removal and distribution of goods often represents a bottleneck as it limits the achievable throughput and, thus, the efficiency of the complete warehousing system. Consequently, an optimal management of the distribution of goods can lead to large efficiency and financial gains. In this contribution, it is shown for an accurate model of an industrial real-world high-rise warehouse (HRW), which is connected to a polymer production plant, that rigorous optimization of the distribution procedure can lead to a large efficiency increase in comparison to standard techniques. To this end, the distribution problem is modelled using timed automata (TA) in which the jobs arrive one after the other and are not known in advance. The TA-based approach, which was originally developed for standard scheduling problems, is extended to a reactive moving window approach that employs a cost-optimal reachability algorithm to compute optimal plans for the goods distribution. We show that compared to a typical first-come first-served dispatching rule, this approach can lead to throughput gains of more than 40%.

Optimal Management of Shuttle Robots in a Laboratory Automation System of a Cement Plant

Abstract To meet the quality requirements in the production of cement, samples are taken in the production process, transported to a central lab and analysed partially automatically. The measurements are used to adjust the parameters of the production process. An optimal management of the laboratory system can lead to significant throughput gains and shorter analysis times. In this contribution, we describe the optimal scheduling of the robots-based laboratory automation system a cement plant. The TA-based approach is embedded into a reactive scheduling approach to handle the different uncertainties in the system, e.g. manually inserted jobs. We show that compared to a priority-based dispatching rule, this approach leads to significant throughput gains while meeting all process-related timing restrictions.