AbstractLog bucking is an industrial problem of subdividing longer logs into smaller logs such that potential revenue is maximized. The model has been solved using dynamic programming predominately, and branch-and-bound to a lesser extent. This research considers the effect of modelling assumptions that have been used to accommodate the solution techniques. Three levels of increasing model detail are considered. An examination of the nature and amount of error is made by using a model of lesser detail instead of a more precise model. The main experiment is used to show the differences in solution accuracy for the three dynamic programming strategies when compared with the accuracy of solutions generated by using branch-and-bound. The final analysis considers the cost in terms of time to solution for each solution technique to produce highly comparable solution accuracy.