OBDD-Based Universal Planning for Multiple Synchronized Agents in Non-Deterministic Domains

Model checking representation and search techniques were recently shown to be efficiently applicable to planning, in particular to non-deterministic planning. Ordered Binary Decision Diagrams (OBDDs) encode a planning domain as a non-deterministic finite automaton (NFA) and fast algorithms from model checking search for a solution plan. With proper encodings. OBDDs can effectively scale and can provide universal plans for complex planning domains. We are particularly interested in addressing the complexities arising in non-deterministic, multi-agent domains. In this paper, we present UMOP, a new universal OBDD-based planning framework applicable to non-deterministic and multi-agent domains. We introduce a new planning domain description language, NADL, to include the specification or such non-deterministic, multi-agent domains. The language contributes the explicit definition of controllable agents and uncontrollable environment agents. We describe the syntax and semantics of NADL and show how to build an efficient OBDD-based representation of an NADL description. The UMOP planning systems uses NADL and it includes the previously developed strong and strong cyclic planning algorithms (Cimatti et al., 1998a, 1998b). In addition, we introduce a new optimistic planning algorithm, which relaxes optimality guarantees and generates plausible universal plans in some domains where no strong or strong cyclic solution exists. We present empirical results in a previously tested non-deterministic domains. We also introduce three new multi-agent domains with complex environment actions. UMOP is shown to be a rich and efficient planning system.

[1]  Daniel S. Weld.  An Introduction to Least Commitment Planning , 1994, AI Mag..

[2]  Matthew L. Ginsberg,et al.  Universal Planning: An (Almost) Universally Bad Idea , 1989, AI Mag..

[3]  Richard S. Sutton,et al.  Introduction to Reinforcement Learning , 1998 .

[4]  Randal E. Bryant,et al.  Graph-Based Algorithms for Boolean Function Manipulation , 1986, IEEE Transactions on Computers.

[5]  Enrico Giunchiglia,et al.  Representing Action: Indeterminacy and Ramifications , 1997, Artif. Intell..

[6]  Manuela M. Veloso,et al.  Rationale-Based Monitoring for Planning in Dynamic Environments , 1998, AIPS.

[7]  E BryantRandal.  Graph-Based Algorithms for Boolean Function Manipulation , 1986 .

[8]  Marcel Schoppers,et al.  Universal Plans for Reactive Robots in Unpredictable Environments , 1987, IJCAI.

[9]  Edmund M. Clarke,et al.  Symbolic Model Checking with Partitioned Transistion Relations , 1991, VLSI.

[10]  Manuela M. Veloso,et al.  Task Decomposition, Dynamic Role Assignment, and Low-Bandwidth Communication for Real-Time Strategic Teamwork , 1999, Artif. Intell..

[11]  Michel Barbeau,et al.  Planning Control Rules for Reactive Agents , 1997, Artif. Intell..

[12]  Kenneth L. McMillan,et al.  Symbolic model checking , 1992 .

[13]  Daniel S. Weld.  Recent Advances in AI Planning , 1999, AI Mag..

[14]  Paolo Traverso,et al.  Automatic OBDD-Based Generation of Universal Plans in Non-Deterministic Domains , 1998, AAAI/IAAI.

[15]  Paolo Traverso,et al.  Strong Planning in Non-Deterministic Domains Via Model Checking , 1998, AIPS.