Approximation algorithms in the successive hitting set model

Abstract

We introduce the successive Hitting Set model, where the set system is not given in advance but a set generator produces the sets that contain a specific element from the universe on demand. Despite incomplete knowledge about the set system, we show that several approximation algorithms for the conventional Hitting Set problem can be adopted to perform well in this model. We describe, and experimentally investigate, several scenarios where the new model is beneficial compared to the conventional one.