## Inferring State Constraints for Domain-Independent Planning (1998)

Citations: | 80 - 3 self |

### BibTeX

@INPROCEEDINGS{Gerevini98inferringstate,

author = {Alfonso Gerevini and Lenhart Schubert},

title = {Inferring State Constraints for Domain-Independent Planning},

booktitle = {},

year = {1998},

pages = {905--912},

publisher = {AAAI Press}

}

### OpenURL

### Abstract

We describe some new preprocessing techniques that enable faster domain-independent planning. The first set of techniques is aimed at inferring state constraints from the structure of planning operators and the initial state. Our methods consist of generating hypothetical state constraints by inspection of operator effects and preconditions, and checking each hypothesis against all operators and the initial conditions. Another technique extracts (supersets of) predicate domains from sets of ground literals obtained by Graphplan-like forward propagation from the initial state. Our various techniques are implemented in a package called DISCOPLAN. We show preliminary results on the effectiveness of adding computed state constraints and predicate domains to the specification of problems for SAT-based planners such as SATPLAN or MEDIC. The results suggest that large speedups in planning can be obtained by such automated methods, potentially obviating the need for adding h...

### Citations

955 | Fast planning through planning graph analysis
- Blum, Furst
- 1997
(Show Context)
Citation Context ...der constraints on predicate domains, in the form of sets of possible tuples of arguments. We infer such sets by forward propagation of ground atoms from the initial state in the manner of Graphplan (=-=Blum & Furst, 1995-=-) but without exclusion calculations. These methods have been implemented in our DISCOPLAN package (DIScovering COnstraints for PLANning), and we propose some additional, provably correct techniques t... |

521 | Pushing the envelope: planning, propositional logic and stochastic search
- Kautz, Selman
(Show Context)
Citation Context ...straints discovered by DISCOPLAN in the context of SAT-based planning, we have extended two recent systems for compiling a planning problem specification into a set of propositional clauses: SATPLAN (=-=Kautz & Selman 1996-=-) and MEDIC (Ernst, Millstein, & Weld 1997). While SATPLAN has its own planning language based on logical axiom schemas, MEDIC accepts cation in Table I is consistent since every object of the domain,... |

429 | Planning as satisfiability
- Kautz, Selman
- 1992
(Show Context)
Citation Context ...ow that the use of the inferred constraints dramatically boosted speed and success in finding solutions regardless the kind of solver used. This is surprising since Kautz and Selman's experiments in (=-=Kautz & Selman 1992) suggest -=-that systematic SAT solvers based on the Davis-Putnam procedure such as Tableau could not be expected to benefit from addition of "axioms that explicitly rule out various impossible states" ... |

411 | UCPOP: A sound, complete partial order plmzner for ndi
- Peazbezthy, Weld
- 1991
(Show Context)
Citation Context ... problem was unsolvable). The last two columns indicate the improvements due to use of inferred constraints in terms of cpu-time speedup and failure rate reduction. UCPOP-like specifications as input(=-=Penberthy & Weld 1992-=-), and offers the possibility of automatically compiling them using several representational settings (for details see (Ernst, Millstein, & Weld 1997)). 7 Our extended versions of SATPLAN and MEDIC ex... |

201 | Experimental results on the cross-over point in satisfiability problems
- Crawford, Auton
- 1993
(Show Context)
Citation Context ... et al.'s SATPLAN system, and, as we shall see, of Ernst et al.'s method of compiling UCPOP domain specifications and a particular problem into SAT clauses and then applying a solver such as Tableau (=-=Crawford & Auton 1993-=-), Walksat (Selman, Kautz, & Cohen 1994) or Relsat (Bayardo & Schrag 1997) to obtain a plan. To make effective use of state constraints and predicate domain constraints in the above types of planners,... |

168 | Extending planning graphs to an ADL subset
- Koehler, Nebel, et al.
- 1997
(Show Context)
Citation Context ...nning graph, in both time and space. Our program for constructing the forward action graph of a given planning problem is based on a simple modification of IPP's C-code for building a planning graph (=-=Koehler et al. 1997-=-), 4 in which we avoid computing and propagating mutual exclusion relations. The depth of the action graph can be determined either by explicitly setting a parameter before running the program, or by ... |

121 | Automatic SAT-compilation of planning problems - Ernst, Millstein, et al. - 1997 |

44 | Accelerating partial-order planners: Some techniques for effective search control and pruning - Gerevini, Schubert - 1996 |

40 | Using CSP look-back techniques to solve real-world SAT instances - Jr, Schrag - 1997 |

17 | Explanation closure, action closure, and the Sandewall test suite for reasoning about change
- Schubert
- 1994
(Show Context)
Citation Context ...the facts appearing as nodes of the action graph at level t. Therefore if a ground predicationsOE does not appear at level t of the action graph, then under the CWA and the action closure assumption (=-=Schubert 1994-=-) (which are commonly used in most current planning systems)sOE is false at time t. So, the absence of ground predications in the forward action graph can be used to formulate explicit constraints on ... |