UMOP 1.2 Software Demonstration

UMOP (Jensen 2000; Jensen & Veloso 2000; Jensen, Veloso, & Bowling 2001) is a universal planning system for multi-agent and non-deterministic domains. As input UMOP takes a planning problem described in the Nondeterministic Agent Domain Language (NADL). The NADL problem is translated to a search problem in a finite transition system. The transition system is represented by a conjunctive partitioned transition relation encoded symbolically as a set of reduced ordered Binary Decision Diagrams (BDDs) (Bryant 1986). Version 1.2 of UMOP includes the three original BDDbased universal planning algorithms: weak, strong, and strong cyclic planning where the actions of the uncontrollable environment agents are abstracted (Cimatti, Roveri, & Traverso 1998a; 1998b; Cimatti et al. 2001). An execution of a strong universal plan is guaranteed to reach states covered by the plan and terminate in a goal state after a finite number of steps. An execution of a strong cyclic plan is also guaranteed to reach states covered by the plan and terminate in a goal state, if it terminates. However, a goal state may never be reached due to cycles. An execution of a weak plan may reach states not covered by the plan, it only guarantees that some execution exists that reaches the goal from each state covered by the plan.