Abstract:
One aspect of forest management is the scheduling ofharvest operations over time. Harvest scheduling may be part of a long-term strategic plan or a short-term operational exercise. Current demands and prices for a set of Iog products may be known but future demands and prices are usually uncertain. A common approach to the problem of quantifying uncertainty is the exploration of the effects of different price scenarios on a deterministic model. Only a small subset of possible future outcomes are explored and it is not possible to develop a plan that hedges against all scenarios and takes advantage of any variation in prices. This thesis describes a modelling system that is primarily aimed at developing operational plans where uncertainty in Iog product demands and prices are described as possible future outcomes. The model, which applies stochastic Benders decomposition to efficiently solve large linear programming problems, produces a hedged plan that makes the best decisions based on the information available.