“Cutting pattern optimization in sawmill”
Wednesday 31 May 2023 h. 14:30 - Room 2BC30 - Enrico Vicario (Padova, Dip. Mat.)
Abstract
Optimizing the cutting of wood from log to board is a step within the lumber production chain that has great potential for optimization. This processing in industry is carried out in lines with machines that are increasingly automated and have varying degrees of flexibility in cutting execution. In this work, a specific problem of generating optimal cutting patterns was modeled in order to address it as a Mixed Integer Linear Problem (MILP). The mathematica formulation has been tested on real instances and the result was compared, in terms of obtained value and computation time, with an ad-hoc greedy heuristics. An original method is also presented to approach a more general cutting pattern problem, based on the integration of optimization techniques and the use of Convolutional Neural Networks (CNN) for reliable board valorization.
