This project investigates pruning techniques in optimal state space planning. In particular, we consider techniques based on partial order reduction and symmetry reduction.
Partial order reduction has been originally proposed in the area of computer aided verification. The basic idea is to tackle the state explosion problem by avoiding a blow-up that is induced by independent actions. Recent results have shown that such techniques can be powerful in the area of automated planning as well. In SPOSSS, we investigate partial order reduction techniques for computer aided verification and for planning both theoretically and practically.
Similar to partial order reduction, techniques based on symmetry reduction have been originally proposed in the area of computer aided verification. In contrast to partial order reduction which exploits symmetric action sequences, techniques based on symmetry reduction compute a quotient structure of the state space, where symmetrical states are considered equivalent. In SPOSSS, we investigate symmetry reduction techniques and their applications to planning.