INOSIM Presentation At PEMT 2025 (Frankfurt)
In the process industry, offline schedules frequently become suboptimal or infeasible due to unforeseen disruptions, ranging from minor timing shifts to major events like machine failures or rush orders. To address this, online scheduling is required to react to real-time plant data and events.
At this years Annual Meeting of Process Engineering and Materials Technology (PEMT 2025, November 10-11, Frankfurt/Main, Germany), INOSIM employee Engelbert Pasieka presented a poster on Evolutionary Algorithm-Based Real-Time Scheduling via Simulation-Optimization. He introduced a reactive real-time scheduling framework based on Simulation-Optimization (SO) using an Evolutionary Algorithm (EA). The approach utilizes a high-fidelity Discrete-Event Simulator to both construct and evaluate feasible schedules, ensuring that complex constraints and side effects are naturally respected without the need for explicit gradients.
Engelbert Pasieka of INOSIM`s Innovation Department brought inspiring news to the visitors in Frankfurt. Engelbert presented the results of his research as part of his doctoral studies at TU Dortmund University (Chair of Professor Dr.-Ing. Sebastian Engell).
The system operates on a continuous rolling horizon, providing “anytime performance” by producing a valid “best-so-far” schedule whenever a response is required. A key feature is population continuity, where the EA population is carried over and repaired after disturbances, allowing the system to reuse good patterns and adapt quickly rather than starting from scratch.
The method was validated on a large multi-stage, multi-product batch plant (6 stages, 30 product types) subject to dynamic disturbances, including rush orders and maintenance events. The optimization objective was to minimize tardiness. Results indicate that the proposed SO approach significantly outperforms strategies without replanning and approaches the performance of a clairvoyant (perfect information) benchmark. The system effectively shortens recovery times and produces implementable plans with minimal manual modification.
Download our Poster to get more details of the topics discussed in the presentation
Do you have any questions or would like to know more about this topic? Please contact us.
General Inquiries 
