Log classification in the hardwood timber industry: method and value analysis

Natural resources industries, such as the forest product industry, must deal with variable input material, which affects their efficiency and their ability to accurately predict output yields. In order to address this, the industry can use technologies that adapt to variable input, or plan its operations taking variability into account. In the Canadian softwood lumber industry, the first approach is used with the adoption of advanced technologies that adapt sawing patterns to logs’ and work-in-process characteristic using scanners technology. Another approach to deal with material variability is input material classification. Specific characteristics can be measured to classify input material and reduce variability within each class. However, whether the process involves logs, mining ores or recycled papers, material classification has both a value and a cost. This paper first proposes a method based on classification tree analysis to classify hardwood logs. Next, using agent-based simulation, it analyses the value of different classification strategies, from detailed, to no classification at all. Results show in the context of the Québec hardwood lumber industry that the benefit of detailed classification is offset by its cost, while a relatively simple classification strategy dramatically improves output yield at relatively low cost.