STATISTICALLY COMPENSATED OPTIMIZATION SYSTEM

A method of optimizing the yield of useful material cut from a piece of raw material, such as lumber, includes considering different cutting options for cutting the piece, identifying different potential yields obtainable for each cutting option, determining the relative probability of obtaining each potential yield, and determining the-relative value of each cutting option, preferably by averaging the potential yields for that option where each potential yield is statistically weighted by its relative probability. By comparing the relative values, the highest-valued cutting option is identified and cutting of the piece is performed in accordance with such option. The relative probabilities for the respective possible yields are preferably determined by empirically measured statistical attributes that characterize the uncertain variables that affect these yields and by grouping the uncertain variables into statistically significant classes, each class corresponding to a respective yield, where the relative frequency of each class is determined from the underlying statistical attributes. Alternatively, the relative probabilities are determined by using statistical attributes that are based on a statistically expected distribution.