Stochastics in discrete logistics models: What can we do?

A large number of logistics problems involve integer programming. Adding stochastics to such a problem seems computationally prohibitive. On the other hand, in most cases, stochastics is an integral part of the underlying problem. And more importantly, without adding stochastics explicitly, we may lose important aspects of the solution we are searching for. In this presentation we give a new approach to such problems, and apply the approach to service network design.