Biased search trees

We consider the problem of storing items from a totally ordered set in a search tree so that the access time for a given item depends on a known estimate of the access frequency of the item. We describe two related classes of biased search trees whose average access time is within a constant factor of the minimum and that are easy to update under insertions, deletions and more radical update operations. We present and analyze efficient update algorithms for biased search trees. We list several applications of such trees.