Approximation algorithm for hotlink assignment in the greedy model

Abstract

Link-based information structures such as the web can be enhanced through the addition of hotlinks. Assume that each node in the information structure is associated with a weight representing the access frequency of the node by users. In order to access a particular node, the user has to follow a path leading to it from the root node. By adding new hotlinks to the tree, it may be possible to reduce the access cost of the system, namely, the expected number of steps needed to reach a leaf from the root, assuming the user can decide which hotlinks to follow in each step. The hotlink assignment problem involves finding a set of hotlinks maximizing the gain in the expected cost. The paper addresses this problem in the more realistic greedy user model recently introduced in [3], and presents a polynomial time 2-approximation algorithm for the hotlink assignment problem on trees. © Springer-Verlag Berlin Heidelberg 2004.