An approach to considering uncertainty in developing long-term, least-cost wood procurement policies

TR Number
Journal Title
Journal ISSN
Volume Title
Virginia Tech

An approach was developed to consider the uncertainty which is intrinsic to forestry decisions. The approach was termed a partial stochastic linear program because uncertainty was considered by introducing variation into one element of the linear programming model (the right hand side). To implement this approach, subjective evaluations were made, regarding the amount of uncertainty associated with the values in question.

This approach was applied to a wood procurement problem which had been previously solved as deterministic. The previous problem was a case study of an integrated forest products firm with the objective of minimizing the present value of wood procurement over a 20-year study using linear programming. The management of this firm was required to make subjective estimates of the variation associated with each available source of supply. The original case study was then reformulated as a partial stochastic linear program.

The solutions of the partial stochastic approach were compared to the deterministic solution. This comparison showed the procurement policies suggested by both approaches were much the same. However, the stochastic approach differed in that management could obtain information about the sensitivity of a policy or a source and establish trade-off relationships between the cost of one policy and the uncertainty of another policy. The questions of the extent of model building and the implications for future study in this area are also considered.

forestry decisions