We develop an algorithm for merging plans that are represented in a richly expressive language. Speci cally, weare concerned with plans that have i quantitative temporal constraints, ii actions that are not instantaneous, but rather have temporal extent, and iii conditional branches. Given a set S of such plans, our algorithm nds a set of constraints that jointly ensure that the plans in S are mutually consistent, if such a set of constraints exists. The algorithm has three phases. In the rst, it employs a new data structure, a conditional simple temporal network CSTN, to identify con icts between the plans. Next, it uses an approach developed by Yang 1997 to suggest a potential resolution of the identi ed conicts. Finally, the CSTN is again used to check whether the proposed resolution observes all the temporal constraints. We have implemented our approach, and we present preliminary experimental evidence about domain factors that in uence its performance.
Ioannis Tsamardinos, Martha E. Pollack, John F. Ho