OPTIMIZATION BY TREES ON SIMPLE ADAPTIVE LANDSCAPES

We evaluate the optimizing ability (rate of adaptation) of trees on si mple adaptive landscapes. At points away from a peak, there is a stron g negative relationship between rate of adaptation and tree precision P, a relationship that is independent of the size of the tree. P measu res the variability among trial solutions generated by the tree: high precision trees have low variability, low precision trees have high va riability. Near a peak. the situation reverses, with high precision tr ees showing higher rates of adaptation than low precision trees, howev er, for all trees, the absolute rate of adaptation is uniformly low. O n multiple-peak landscapes, the probability of crossing an adaptive va lley from a lower peak to a higher peak is also negatively correlated with tree precision. These results suggest that under a wide range of conditions, trees with low precision are, on average, the best optimiz ers.