Solution Approaches for the Dynamic Stacking Problem

Publication, 2020


S. Raggl, A. Beham, S. Wagner, M. Affenzeller - Solution Approaches for the Dynamic Stacking Problem - Proceedings of the Genetic and Evolutionary Computation Conference Companion - GECCO '20, Virtuell, Austria, 2020, pp. 1652-1660


In this paper we describe a dynamic stacking problem as it arises in a more complex form in the steel industry. We describe a simulation model and the simulated processes that are implemented. The model covers a gantry crane that performs relocations of blocks among three types of stacks. In the real world the crane operators and dispatchers solve complex stacking problems targeting to minimize relocation effort while adhering to many constraints, to various time windows, and to satisfy quality demands. While our model does not include all the real-world constraints, the challenges and benefits of solving this problem as a dynamic optimization problem in contrast to a static formulation become apparent. We adapt an existing solution approach from a static formulation that is based on the block relocation problem and compare it with a hand-written rule-based approach. Furthermore, we devise a hybrid approach in the form of an open-ended evolutionary algorithm.