Abstract An approach, termed partially stochastic linear programming, combines linear programming with subjective probability estimates to quantitatively recognize uncertainty in forestry decision making. The approach is applied to minimizing wood procurement costs over an industrial firm's planning period. The future availabilities of land and raw material are considered random variables rather than point estimates. Subjective probability distributions on each random variable are developed from information obtained from the firm's forestry personnel. Alternative resource availability situations are simulated and least-cost linear programming solutions obtained. The results are a distribution of solutions which can be described by its mean and variance. The approach provides the manager with decision-making information, such as the probability of the cost being above or below a specified level, which is not available from deterministic linear programs. Limitations of the approach along with suggestions for application and additional study are discussed. Forest Sci. 17:224-229.
Thompson et al. (Tue,) studied this question.