## Planning as satisfiability (1992)

### Cached

### Download Links

- [www.cs.rochester.edu]
- [www.cs.rochester.edu]
- [www.cs.rochester.edu]
- [www.cs.cornell.edu]
- [www.aaai.org]
- DBLP

### Other Repositories/Bibliography

Venue: | IN ECAI-92 |

Citations: | 456 - 26 self |

### BibTeX

@INPROCEEDINGS{Kautz92planningas,

author = {Henry Kautz and Bart Selman},

title = {Planning as satisfiability},

booktitle = {IN ECAI-92},

year = {1992},

pages = {359--363},

publisher = {Wiley}

}

### Years of Citing Articles

### OpenURL

### Abstract

We develop a formal model of planning based on satisfiability rather than deduction. The satis ability approach not only provides a more flexible framework for stating di erent kinds of constraints on plans, but also more accurately reflects the theory behind modern constraint-based planning systems. Finally, we consider the computational characteristics of the resulting formulas, by solving them with two very different satisfiability testing procedures.

### Citations

1567 | Some philosophical problems from the standpoint of artificial intelligence - McCarthy, Hayes - 1969 |

1192 | Chaff: engineering an efficient sat solver
- Moskewicz, Madigan, et al.
- 2001
(Show Context)
Citation Context ...T solver walksat (Selman, Kautz, & Cohen 1994), the forwardchecking DPLL-based solver satz (Li & Anbulagan 1997), and the clause-learning DPLL-based solvers relsat (Bayardo & Schrag 1997) and zChaff (=-=Moskewicz et al. 2001-=-). By contrast, SATPLAN04 uses a single highly optimized DPLL-based solver called “siege”, that was developed by Lawrence Ryan as part of his research at Simon Fraiser University under the direction o... |

1113 | A computing procedure for quantification theory - Davis, Putnam - 1960 |

1014 | Fast planning through planning graph analysis
- Blum, Furst
- 1995
(Show Context)
Citation Context ... actions may occur in parallel at each time step, and the total number of time steps in guaranteed to be as small as possible. Also like blackbox, SATPLAN works by: 1. Constructing a graphplan-style (=-=Blum & Furst 1995-=-) style planning graph up to some length k; 2. Translating the constraints implied by the graph into a set of clauses, where each specific instance of an action or fact at a point in time is a proposi... |

700 | A new method for solving hard satisfiability problems - Selman, Levesque, et al. - 1992 |

663 | Planning for Conjunctive Goals - Chapman - 1987 |

545 | Pushing the envelope: Planning, propositional logic, and stochastic search
- Kautz, Selman
- 1996
(Show Context)
Citation Context ...l approach to planning (Kautz & Selman 1992). This changed in 1996, when Satplan was shown to be competitive with current planning technology, leading to a mini-explosion of interest in the approach (=-=Kautz & Selman 1996-=-). Within a few years, however, heuristic search planning appeared to be vastly superior to planning as satisfiability, and many researchers wrote off the earlier success of the approach as a fluke. I... |

532 | Optimization by simulated annealing: An experimental evaluation. Part 1, graph partitioning - Johnson, Aragon, et al. - 1989 |

503 | An extensible SAT-solver - Eén, Sörensson - 2003 |

290 | and easy distributions of SAT problems - Hard - 1992 |

272 | Unifying SAT-based and Graphbased planning
- Kautz, Selman
- 1999
(Show Context)
Citation Context ...updated version of the planning as satisfiability approach originally proposed in (Kautz & Selman 1992; 1996) using hand-generated translations, and implemented for PDDL input in the blackbox system (=-=Kautz & Selman 1999-=-). Like blackbox, SATPLAN04 accepts the STRIPS subset of PDDL and finds solutions with minimal parallel length: that is, many (non-interferring) actions may occur in parallel at each time step, and th... |

236 | Application of theorem proving to problem solving - Green - 1969 |

217 | Using csp look-back techniques to solve real-world sat instances
- Bayardo, Schrag
- 1997
(Show Context)
Citation Context ... search engine, the local-search SAT solver walksat (Selman, Kautz, & Cohen 1994), the forwardchecking DPLL-based solver satz (Li & Anbulagan 1997), and the clause-learning DPLL-based solvers relsat (=-=Bayardo & Schrag 1997-=-) and zChaff (Moskewicz et al. 2001). By contrast, SATPLAN04 uses a single highly optimized DPLL-based solver called “siege”, that was developed by Lawrence Ryan as part of his research at Simon Frais... |

157 | Encoding plans in propositional logic - Kautz, McAllester, et al. - 1996 |

156 | Solving large-scale constraint satisfaction and scheduling problems using a heuristic repair method - Minton, Johnston, et al. - 1990 |

150 | Constructing conditional plans by a theorem-prover
- Rintanen
- 1999
(Show Context)
Citation Context ...that are higher in the computational complexity hierarchy. For example, conditional planning in nondeterministic domains can be encoding as evaluating the truth of a quantified Boolean formula (QBF) (=-=Rintanen 1999-=-). Probabilistic planning can be encoded as stochastic satisfiability, an extension of QBF that includes a “random choice” quantifier (Majercik & Littman 1999). Many planning problems involve reasonin... |

129 | Automatic SATcompilation of planning problems - Ernst, Millstein, et al. - 1997 |

128 | BLACKBOX: A new Approach to the application of theorem proving to problem solving
- Kautz, Selman
- 1998
(Show Context)
Citation Context ...tz, McAllester, & Selman 1996). The first complete implementation of Satplan that took STRIPS notation as input was the MEDIC system of Ernst et al. (1997). The next year saw the release of Blackbox (=-=Kautz & Selman 1998-=-), which also performed “mutex propagation”, a form of local-consistency reasoning introduced by Graphplan (Blum& Furst 1995), before generating each encoding. In the 1998 International Planning Compe... |

127 | Heuristics based on unit propagation for satisfiability problems
- Li, Anbulagan
- 1997
(Show Context)
Citation Context ...rs used. Blackbox included the original graphplan (non-translation based) search engine, the local-search SAT solver walksat (Selman, Kautz, & Cohen 1994), the forwardchecking DPLL-based solver satz (=-=Li & Anbulagan 1997-=-), and the clause-learning DPLL-based solvers relsat (Bayardo & Schrag 1997) and zChaff (Moskewicz et al. 2001). By contrast, SATPLAN04 uses a single highly optimized DPLL-based solver called “siege”,... |

99 | Synthesizing plans that contain actions with context dependent effects - Pednault - 1988 |

74 | The lpsat engine and its application to resource planning
- Wolfman, Weld
- 1999
(Show Context)
Citation Context ...ne way to handle metric information is to compile a planning problem to both a Boolean formula and a set of linear inequalities, where the each inequality is associated with a particular proposition (=-=Wolfman & Weld 1999-=-; Shin & Davis 2005). An alternative approach would be to encode metric constraints as Boolean arithmetic in clausal form. Although such an approach may seem prima facie impractical, compiling metric ... |

61 | Contingent planning under uncertainty via stochastic satisfiability
- Majercik, Littman
- 1999
(Show Context)
Citation Context ...ng the truth of a quantified Boolean formula (QBF) (Rintanen 1999). Probabilistic planning can be encoded as stochastic satisfiability, an extension of QBF that includes a “random choice” quantifier (=-=Majercik & Littman 1999-=-). Many planning problems involve reasoning about metric quantities, such as resources and metric time. One way to handle metric information is to compile a planning problem to both a Boolean formula ... |

58 | A new method for solving hard satis ability problems - Selman, Levesque, et al. - 1992 |

56 | Local search topology in planning benchmarks: An empirical analysis - Hoffmann - 2001 |

55 | The detection and exploitation of symmetry in planning problems
- Fox, Long
- 1999
(Show Context)
Citation Context ... of symmetries in the encoded problem. Symmetry detection has been an important theme of much work in constraint satisfaction, satisfiability testing and planning (Joslin & Roy 1997; Shlyakhter 2001; =-=Fox & Long 1999-=-). While much of this work has concentrated on automatically detecting symmetries, it is often more straightforward and practical to provide the author of a planning domain with an easy way to state h... |

53 | Planning as temporal reasoning, in - Allen - 1991 |

50 | Plan synthesis: A logical perspective - Rosenschein - 1981 |

46 |
Generating effective symmetrybreaking predicates for search problems
- Shlyakhter
- 2001
(Show Context)
Citation Context ... as various kinds of symmetries in the encoded problem. Symmetry detection has been an important theme of much work in constraint satisfaction, satisfiability testing and planning (Joslin & Roy 1997; =-=Shlyakhter 2001-=-; Fox & Long 1999). While much of this work has concentrated on automatically detecting symmetries, it is often more straightforward and practical to provide the author of a planning domain with an ea... |

41 | A Computing Procedure for Quanti cation Theory - Davis, Putnam - 1960 |

40 | Complexity Results for Blocks-World Planning - Gupta, Nau - 1991 |

29 |
Exploiting symmetry in lifted CSPs
- Joslin, Roy
- 1997
(Show Context)
Citation Context ...cture can be viewed as various kinds of symmetries in the encoded problem. Symmetry detection has been an important theme of much work in constraint satisfaction, satisfiability testing and planning (=-=Joslin & Roy 1997-=-; Shlyakhter 2001; Fox & Long 1999). While much of this work has concentrated on automatically detecting symmetries, it is often more straightforward and practical to provide the author of a planning ... |

26 | Backtrack search algorithms for propositional logic satisfiability: Review and innovations - Nadel |

18 |
Processes and continuous change in a sat-based planner
- Shin, Davis
- 2005
(Show Context)
Citation Context ...ic information is to compile a planning problem to both a Boolean formula and a set of linear inequalities, where the each inequality is associated with a particular proposition (Wolfman & Weld 1999; =-=Shin & Davis 2005-=-). An alternative approach would be to encode metric constraints as Boolean arithmetic in clausal form. Although such an approach may seem prima facie impractical, compiling metric constraints as SAT ... |

16 |
Efficient algorithms for clause learning sat solvers
- Ryan
- 2004
(Show Context)
Citation Context ...anching and backtracking strategies, and won the the 2004 SAT competition industrial benchmark category. It is remarkable that the two most recent systems, MiniSat (Eén & Sörensson 2003) and Siege (=-=Ryan 2004-=-), can solve problems with with 0.25 million variables in less than 30 seconds. Furthermore, while Siege is a proprietary, highly optimized system, the comparably fast system, MiniSat, is a straightfo... |

15 | Towards realistic benchmarks for planning: the domains used in the classical part of IPC–4 - Hoffmann, Edelkamp, et al. - 2004 |

9 | Using problem structure for efficient clause learning - Beame, Kautz, et al. - 2003 |

9 | Symchaff: A structure-aware satisfiability solver - Sabharwal |

9 | Adaptive eager boolean encoding for arithmetic reasoning in verification - Seshia - 2005 |

8 | The Complexity of Automated Reasoning - Vellino - 1989 |

4 | Planning with constraints (MOLGEN: parts 1 and 2 - Stefik - 1981 |

4 |
The 2004 international planning competition. http://ls5-www.cs.uni-dortmund.de/ edelkamp/ipc4/main.html
- Edelkamp, Hoffmann, et al.
- 2004
(Show Context)
Citation Context ...searchers wrote off the earlier success of the approach as a fluke. It was therefore rather surprising when Satplan won first place for optimal STRIPS planning in the 2004 ICAPS planning competition (=-=Edelkamp et al. 2004-=-). This talk will attempt to deconstruct the reasons for Satplan’s successes and failures, and discuss ways the approach might be extended to handle “open” domains, metric constraints, and domain symm... |

2 | Cultural support for improivisation - Agre, Horswill - 1992 |

2 | Learning control knowledge for planning
- Huang
- 2002
(Show Context)
Citation Context ...t obviously) PTIME-solvable. It is also possible to transform planning domains where feasible (non-optimal) planning is easy but optimal planning is hard to ones where feasible planning is also hard (=-=Huang 2002-=-). Such instances are useful for comparing planning algorithms without the complexity of taking solution quality into consideration. Leveraging Plan Structure What, then, distinguishes planning from p... |

1 | Ste k. Planning with constraints (molgen - Mark - 1981 |