Improving dynamic treatment unit forest planning with cellular automata heuristics

We present a model for conducting dynamic treatment unit (DTU) forest planning using a heuristic cellular automata (CA) approach. The clustering of DTUs is driven by entry costs associated with treatments, thus we directly model the economic incentive to cluster. The model is based on the work presented in the literature but enhanced by adding a third phase to the CA algorithm where DTUs are mapped in high detail. The model allows separate but nearby forest areas to be included in the same DTU and shares the entry cost if they are within a defined distance. The model is applied to a typical long-term forest planning problem for a 1 182 ha landscape in northern Sweden, represented by 4 218 microsegments with an average size of 0.28 ha. The added phase increased the utility by 1.5-32.2%. The model produced consistent solutions-more than half of all microsegments were managed with the same treatment program in 95% of all solutions when multiple solutions were found.