Vortrag auf der Tagung der Deutschen Gesellschaft für Operations Research

Das Projekt Re²Pli war mit einem Beitrag auf der jährlichen International Conference on Operations Research vertreten, die gemeinsam von der Deutschen Gesellschaft für Operations Research (GOR), der Österreichischen Gesellschaft für Operations Research (ÖGOR) und der Schweizerischen Gesellschaft für Operations Research (SVOR/ASRO) veranstaltet wird. Die Tagung fand vom 29. August bis 1. September 2023 in Hamburg statt.
Im Stream "Energie und Umwelt" präsentierte Sascha Burmeister einen neuen Ansatz für das Green Multi-Objective Flexible Job Shop Problem unter Berücksichtigung von Energiespeichern.

Abstract (Englisch):

Environmental sustainability has become increasingly important in recent years, as many countries set targets to reduce the carbon emissions. A major consumer of energy is the manufacturing industry. For manufacturers it is essential to find ways to align their production with the current energy mix and consume energy at times with high renewable energy production without jeopardizing production deadlines. In addition to schedules based on the energy mix, energy storage systems can be used to compensate for fluctuations in renewable energy sources. In the literature, the Green Flexible Job Shop Scheduling Problem (FJSP) is concerned with resource and environmental aspects in addition to the economic objective of the minimum makespan. However, existing approaches neglect the volatile dynamic energy mix, a multi-criteria objective, or the consideration of modern energy storage systems. We aim to close this research gap and propose an algorithm based on the Non-dominated Sorting Genetic Algorithm (NSGA-II) with the goal of minimizing both, makespan and emissions. In order to design low-emission production, the algorithm constructs production schedules while considering the charging and discharging of an energy storage system. We evaluate the approach in computational experiments using prominent FJSP-benchmark instances from the literature with real emissions. We investigate the trade-off between a short makespan and low carbon emissions based on solutions on the approximated Pareto front and discuss the value of energy storage systems for sustainable production.