Initial work will concentrate on mining generalised association rules and generalised sequential patterns. These techniques, along with numerous others, are described at http://www.almaden.ibm.com/cs/quest/publications.html.
Part of the challenge of this project is to adapt these techniques for our purposes. Sometimes, we do not need the full generality of some particular method or we need to do something that is somewhat different (e.g., different pruning heuristics). As mentioned earlier, speed will be a primary consideration and it may be that the algorithms given in research papers are not the fastest ones for our particular data set, so you would need to develop new ones.