Evolving Submodels for Column Generation in Cutting and Packing for Glulam Production

This study introduces a novel optimization approach for the pressing phase in glued laminated timber (glulam) production, integrating column generation techniques with an evolution-ary algorithm within a real-world glulam factory context. At its core, an evolutionary strategy refines parameters in a sub model dedicated to developing cutting patterns via column generation. These patterns are then utilized in a separate, overarching 2D packing problem, targeting the optimization of the pressing process while adhering to pressing capacity constraints. The solutions' effectiveness feeds back into the evolutionary strategy, guiding the optimization of its parameters. This iterative, multi-layered framework presents a significant advancement in glulam manufacturing, aiming to enhance production efficiency, reduce waste, and expedite customer order fulfillment.