#### DMCA

## Improvements to SAT-based conformant planning (2001)

Venue: | IN PROC. OF 6TH EUROPEAN CONFERENCE ON PLANNING (ECP-01) |

Citations: | 19 - 5 self |

### Citations

1155 | Fast planning through planning graph analysis
- Blum, Furst
- 1997
(Show Context)
Citation Context ... initial state and/or the effects of actions. Some preliminary experimental results reported in [4] show that the approach can be competitive w.r.t. CGP [5], a conformant planner based on plan graphs =-=[6]-=-. ? A special thank to Paolo Ferraris for many fruitful discussions on the topic of this paper. Paolo has also participated to the design of the architecture of C-PLAN, and has developed the first ver... |

1021 | Symbolic Boolean manipulation with ordered binarydecision diagrams.
- Bryant
- 1992
(Show Context)
Citation Context ... 20 they both run out of memory. Indeed, both GPT and CMBP can require huge amounts of memory: GPT for storing the set of belief states visited, and CMBP for storing (as Binary Decision Diagrams, BDD =-=[18]) the tran-=-sition relation and the set of belief states visited. We also consider the elaboration of the "bomb in the toilet" in which dunking a package clogs the toilet. There is the additional action... |

907 | Symbolic model checking without BDDs
- Biere, Cimatti, et al.
- 1999
(Show Context)
Citation Context ...ches have different range of applicability is also confirmed by -- previous work comparing BDD and DLL as SAT procedures, see [20], -- recent work in symbolic reachability in formal verification, see =-=[21, 22]-=-. Beside the already cited [15, 16], two other works on conformant planning are [5] and [24]. In [5], the authors propose an approach based on plan graphs. The underlying idea is to construct a planni... |

824 |
D.: A machine program for theorem proving
- Davis, Logemann, et al.
- 1962
(Show Context)
Citation Context ... the set of clauses obtained from ' by deleting the clauses in which L occurs as a disjunct, and eliminating L from the others. As it can be seen, C-SAT is the Davis-Logemann-Loveland (DLL) procedure =-=[10]-=-, except that, as soon as one assignment satisfying cnf(P ) is found, the procedure C-SAT TEST is invoked. Indeed, each assignment satisfying cnf(P ) corresponds to a set of possible plans, and C-SAT ... |

572 | Pushing the envelope: Planning, propositional logic, and stochastic search
- Kautz, Selman
- 1996
(Show Context)
Citation Context ...rmed the action. Then, the problem of finding a plan of length n leading from an initial state I to a goal state G corresponds to finding an assignment satisfying I 0 ^ ^ n 1 i=0 tr D i ^ Gn (1) (see =-=[2]-=- for more details). This simple idea had a lot of impact in the classical planning community, mainly because it led to very impressive results (see, e.g., [2]). In [3], the planning as satisfiability ... |

500 | Planning as satisfiability
- Kautz, Selman
- 1992
(Show Context)
Citation Context ...ntal analysis shows that the resulting procedure is competitive with other state-of-the-art conformant planners on domains with a high degree of parallelism. 1 Introduction Planning as satisfiability =-=[1]-=- is an efficient technique for classical planning. In the classical setting, the idea behind planning as satisfiability is simple: For any action description D, it is possible to compute a proposition... |

378 | Hybrid algorithms for the constraint satisfaction problem - Prosser - 1993 |

224 | Planning with incomplete information as heuristic search in belief space
- BONET, GEFFNER
(Show Context)
Citation Context ...possible only if the package has not been previously dunked. We first consider planning problems with jP j = 2; 4; 6; 8; 10; 15; 20, and jT j = 1. We compare our system with Bonet's and Geffner's GPT =-=[15]-=-, and Cimatti's and Roveri's CMBP [16, 17] planners. These are two among the most recent conformant planners, and according to the results presented in [17], also the most effective to date. We remark... |

194 | Admissible heuristics for optimal planning
- Haslum, Gener
- 2000
(Show Context)
Citation Context ... GPT run out of memory. This seems to point out that C-PLAN range of applicability is different from the range of applicability of CMBP and GPT. Analogous results supporting this fact are reported in =-=[19]-=- where it is shown that for classical, highly parallel domains the planning as satisfiability approaches appear to do best. The fact that SAT-based approaches and BDD-based approaches have different r... |

183 |
Using csp look-back techniques to solve real world sat instances
- Schrag
- 1997
(Show Context)
Citation Context ...ning We do not enter into the details about how backjumping and learning are implemented in SAT, and assume that the reader is familiar with the topic (see, e.g., [11--13]). Here we do the same as in =-=[12, 13]-=-, except that we have to extend the procedure there described in order to account for the rejection of assignments corresponding to possible but not valid plans. Indeed, what we can simply do --- assu... |

154 | Constructing conditional plans by a theorem prover
- Rintanen
(Show Context)
Citation Context ...nd DLL as SAT procedures, see [20], -- recent work in symbolic reachability in formal verification, see [21, 22]. Beside the already cited [15, 16], two other works on conformant planning are [5] and =-=[24]-=-. In [5], the authors propose an approach based on plan graphs. The underlying idea is to construct a planning graph for every possible deterministic version of the original planning problem. Constrai... |

153 | An action language based on causal explanation: Preliminary report
- Giunchiglia, Lifschitz
- 1998
(Show Context)
Citation Context ... Some of these optimizations have been incorporated in C-PLAN, a SAT-based system for planning in domains whose action description component is specified using the highly expressive action language C =-=[7]-=-. As a consequence, C-PLAN allows for conformant planning in domains with constraints, concurrent actions, and nondeterminism. A comparative experimental analysis shows that C-PLAN is competitive with... |

139 |
Conformant graphplan
- Smith, Weld
- 1998
(Show Context)
Citation Context ...roblems having incomplete information about the initial state and/or the effects of actions. Some preliminary experimental results reported in [4] show that the approach can be competitive w.r.t. CGP =-=[5]-=-, a conformant planner based on plan graphs [6]. ? A special thank to Paolo Ferraris for many fruitful discussions on the topic of this paper. Paolo has also participated to the design of the architec... |

135 |
A critique of pure reason
- McDermott
- 1987
(Show Context)
Citation Context ...een implemented on top of SIM, an efficient SAT checker developed by our group [13]. To evaluate C-PLAN's effectiveness we consider an elaboration of the traditional "bomb in the toilet" pro=-=blem from [14]-=-. There is a finite set P of packages and a finite set T of toilets. One of the packages is armed because it contains a bomb. Dunking a package in a toilet disarms the package and is possible only if ... |

134 | Conformant Planning via Symbolic Model Checking and Heuristic Search. AIJ
- Cimatti, Roveri, et al.
- 2004
(Show Context)
Citation Context ...een previously dunked. We first consider planning problems with jP j = 2; 4; 6; 8; 10; 15; 20, and jT j = 1. We compare our system with Bonet's and Geffner's GPT [15], and Cimatti's and Roveri's CMBP =-=[16, 17]-=- planners. These are two among the most recent conformant planners, and according to the results presented in [17], also the most effective to date. We remark that both CMBP and GPT are sequential pla... |

51 | Conformant planning via model checking
- CIMATTI, ROVERI
- 1999
(Show Context)
Citation Context ...een previously dunked. We first consider planning problems with jP j = 2; 4; 6; 8; 10; 15; 20, and jT j = 1. We compare our system with Bonet's and Geffner's GPT [15], and Cimatti's and Roveri's CMBP =-=[16, 17]-=- planners. These are two among the most recent conformant planners, and according to the results presented in [17], also the most effective to date. We remark that both CMBP and GPT are sequential pla... |

50 | Ordered binary decision diagrams and the Davis-Putnam procedure
- Uribe, Stickel
- 1994
(Show Context)
Citation Context ...appear to do best. The fact that SAT-based approaches and BDD-based approaches have different range of applicability is also confirmed by -- previous work comparing BDD and DLL as SAT procedures, see =-=[20]-=-, -- recent work in symbolic reachability in formal verification, see [21, 22]. Beside the already cited [15, 16], two other works on conformant planning are [5] and [24]. In [5], the authors propose ... |

38 | H.: Satisfiability planning with causal theories
- McCain, Turner
- 1998
(Show Context)
Citation Context ...; : : : ; n is possible for if, starting from an arbitrarily chosen initial state, the consecutive execution of the actions 1 ; : : : ; n can lead to a goal state. According to the results in [9, =-=7]-=-, possible plans of length n correspond to the assignment satisfying (1), as in the classical setting. However, if we have incomplete information about the initial state and/or actions are non determi... |

36 | Planning as satisfiability in nondeterministic domains
- Ferraris, Giunchiglia
(Show Context)
Citation Context ... is no valid plan of lengthsn). Thus, if we are looking for optimal plans, we have to consider n = 0; 1; 2; 3; 4; : : :, and for each value of n, call C-SAT. This is the idea behind the system C-PLAN =-=[4, 8]-=-. 1 2 3 4 5 1 2 3 4 5 Fig. 2. A robot navigation problem 3 Improvements to SAT-Based conformant planning In order to understand how the basic procedure above described works consider the following rob... |

31 | SAT-based Planning in Complex Domains: Concurrency, Constraints and Nondeterminism
- Castellini, Giunchiglia, et al.
(Show Context)
Citation Context ...of and optimizations to the basic procedure. In Section 4 we perform the experimental comparative analysis. We end the paper in Section 5 with some conclusions. The material here presented is part of =-=[8]-=-. In [8], we give the precise definitions, statements, proofs, and we also present some additional experimental analysis. 2 Conformant planning via SAT This Section introduces some terminology and not... |

26 |
Planning as satisfiability with expressive action languages: Concurrency, constraints and nondeterminism
- Giunchiglia
- 2000
(Show Context)
Citation Context ...^ ^ n 1 i=0 tr D i ^ Gn (1) (see [2] for more details). This simple idea had a lot of impact in the classical planning community, mainly because it led to very impressive results (see, e.g., [2]). In =-=[3]-=-, the planning as satisfiability approach has been extended to conformant planning, that is, to planning problems having incomplete information about the initial state and/or the effects of actions. S... |

26 | Evaluating search heuristics and optimization techniques in propositional satisfiability
- Giunchiglia, Maratea, et al.
- 2001
(Show Context)
Citation Context ...ning We do not enter into the details about how backjumping and learning are implemented in SAT, and assume that the reader is familiar with the topic (see, e.g., [11--13]). Here we do the same as in =-=[12, 13]-=-, except that we have to extend the procedure there described in order to account for the rejection of assignments corresponding to possible but not valid plans. Indeed, what we can simply do --- assu... |

3 |
Gila Kamhi, Armando Tacchella, and Moshe Vardi, â€˜Benefits of bounded model checking at an industrial setting
- Copty, Fix, et al.
- 2001
(Show Context)
Citation Context ...ches have different range of applicability is also confirmed by -- previous work comparing BDD and DLL as SAT procedures, see [20], -- recent work in symbolic reachability in formal verification, see =-=[21, 22]-=-. Beside the already cited [15, 16], two other works on conformant planning are [5] and [24]. In [5], the authors propose an approach based on plan graphs. The underlying idea is to construct a planni... |

2 | NuSMV Version 2: BDD-based + SAT-based symbolic model checking
- Cimatti, Giunchiglia, et al.
- 2001
(Show Context)
Citation Context ...r hand, by dealing with the original planning problem, we are able to introduce optimizations ---like the ones described in Section 3--- which take into account the nature of the original problem. In =-=[23], a new al-=-gorithm for conformant planning based on "heuristic-symbolic search", is proposed and the experimental results are impressive. A detailed analysis of the paper and the results is in our agen... |