Using Tabu Search to generate stand-level, rule-based bucking patterns

A Tabu Search (TS) heuristic was developed for generating easily implementable bucking rules that are applicable to entire stands (as opposed to individual stem classes), and for selecting the rule-set that provides the best feasible solution (given log prices and market constraints). The tabu restrictions used for guiding the search process are based on the frequency of the export logs' attributes, and the profit corresponding to the local optima found during the search. Parallel tabu lists of different time-spans were considered. The aspiration criterion for revoking a tabu status consisted of improving upon the best feasible solution. The model was applied to a set of radiata pine stands, considering different market constraints and price sets. The TS approach leads to bucking outcomes that are reasonably close, in profit terms, to the solutions obtained by an integer programming model designed for selecting the optimal set of stem-specific bucking patterns.