New discrete Online-Optimization methods for in-house logistic systems using various reoptimization models will be evaluated in theory and simulation.

Many logistic systems in production planning require a reliable online control. In this project, a special class of online optimization algorithms is being studied: reoptimization algorithms. Whenever the state of the system changes, an offline optimization problem (auxiliary problem) is solved, based on the data currently available. This auxiliary problem may be the offline version of the original online optimization problem (standard auxiliary problem): the same objective is optimized subject to the same constraints, but only the already known input data are used. This can lead to undesired behavior.

The goal of this project is to find suitable auxiliary problems to obtain, e.g., a stable online behavior of the system. We conjecture that auxiliary problems with low sensitivity are appropriate.