ABSTRACT
Assembly tasks are a big challenge for nowadays planning systems. Depending on the problem domain, a planner has to deal with a huge number of objects which can be combined in several ways. Uncertainty about the outcome of actions and the availability of parts to be assembled even worsens the problem. As a result, classical approaches have shown to be of little use for reactive (online) planning during an assembly, due to the huge computational complexity. The approach proposed in this paper bypasses this problem by calculating the complex planning problems prior to the actual assembly. During assembly the precalculated solutions are then used to provide fast decisions allowing an efficient execution of the assembly ....
