Compact labeling schemes for ancestor queries

We consider the following problem. Give a rooted tree T, label the nodes of Tin the most compact way such that given the labels of two nodes one can determine in constant time, by looking only at the labels, if one node is an ancestor of the other. The best known labeling scheme is rather straightforward and uses labels of size at most 2 log n, where n is the number of vertices In the tree. Our main result in this paper is a labeling scheme with maximum label size close to 3/2 log n. Our motivation for studying this problem is enhancing the performance of Web search engines. In the context of this application each indexed document is a tree and the labels of all trees are maintained in main memory. Therefore even small improvements in the maximum label size are important. There are no lower bounds known for this problem except for an obvious lower bound of log n that follows from the fact that different vertices must have different labels. The guestion whether one can find even shorter labels remains an intriguing open guestion.