Heuristic Optimization of Rough-Mill Yield with Production Priorities
MetadataShow full item record
Efficient lumber utilization at the saw is a key issue in the woodworking industry because of shrinking supply and increasing raw material prices. In this thesis, the formulation of the cross-cut-first method of cutting defects out of lumber, as a one-dimensional stock cutting problem, is discussed, with the objective to maximize yield. Monte-Carlo simulations were used for generating boards of a given grade, to test and compare alternative solution procedures. A fast heuristic for solving the above problem is introduced to enable a real-time computerized implementation. The heuristic is shown to compare favorably with the algorithmic solution obtained, using Kolesar's knapsack algorithm, in terms of solution time and yield. The system is also capable of automatically assigning cutting priorities to reflect demand and production needs. In addition, the cut-rip defect removal strategy is introduced. The results of a set of experiments designed to evaluate the heuristic procedure as applied to the cut-off and cut-rip strategies are reviewed, in comparison to the operator's performance. Finally, the implementation issues on an automated cut-off saw are presented.