## Distributed partial constraint satisfaction problem (1997)

### Cached

### Download Links

- [www.edu.kobe-u.ac.jp]
- [www.kecl.ntt.co.jp]
- [lang.is.kyushu-u.ac.jp]
- [jmvidal.cse.sc.edu]
- DBLP

### Other Repositories/Bibliography

Venue: | Principles and Practice of Constraint Programming |

Citations: | 60 - 12 self |

### BibTeX

@INPROCEEDINGS{Hirayama97distributedpartial,

author = {Katsutoshi Hirayama and Makoto Yokoo},

title = {Distributed partial constraint satisfaction problem},

booktitle = {Principles and Practice of Constraint Programming},

year = {1997},

pages = {222--236}

}

### Years of Citing Articles

### OpenURL

### Abstract

Abstract. Many problems in multi-agent systems can be described as distributed Constraint Satisfaction Problems (distributed CSPs), where the goal is to nd a set of assignments to variables that satis es all constraints among agents. However, when real problems are formalized as distributed CSPs, they are often over-constrained and have no solution that satis es all constraints. This paper provides the Distributed Partial Constraint Satisfaction Problem (DPCSP) as a new framework for dealing with over-constrained situations. We also present new algorithms for solving Distributed Maximal Constraint Satisfaction Problems (DM-CSPs), which belong to an important class of DPCSP. The algorithms are called the Synchronous Branch and Bound (SBB) and the Iterative Distributed Breakout (IDB). Both algorithms were tested on hard classes of over-constrained random binary distributed CSPs. The results can be summarized as SBB is preferable when we are mainly concerned with the optimality ofasolution, while IDB is preferable when we want to get a nearly optimal solution quickly. 1

### Citations

1014 | Distributed snapshots: determining global states of distributed systems
- Chandy, Lamport
- 1985
(Show Context)
Citation Context ...at it incorporates a procedure to detect whether the algorithm nds a global solution, in contrast with previous distributed constraint satisfaction algorithms that need to invoke thesnapshot algorithm=-=[2]-=- for detection. On the other hand, one major drawback is that the distributed breakout is not complete, i.e., it may fail to nd a solution even if one exists and also cannot determine that no solution... |

422 | R.J.: Partial constraint satisfaction
- Freuder, Wallace
- 1992
(Show Context)
Citation Context ...esign so that the CSP is not over-constrained. Freuder extended the CSP framework and provided a Partial Constraint Satisfaction Problem (PCSP), which is one of the approaches to over-constrained CSPs=-=[4]-=-. In a PCSP, weare required to nd consistent assignments to an allowable relaxed problem. Although a distributed CSP and a PCSP extend the traditional CSP framework in di erent directions, they are no... |

249 | Distributed Constraint Satisfaction: foundations of cooperation in multi-agent systems
- Yokoo
- 2001
(Show Context)
Citation Context ...have found that certain kind of problems in the real world cannot be handled in the conventional CSP framework, and several studies have been made in order to extend the traditional CSP framework. In =-=[14]-=-, Yokoo et al. presented a distributed Constraint Satisfaction Problem (distributed CSP) as the general framework for dealing with problems in multi-agent systems. A distributed CSP can be considered ... |

194 |
The breakout method for escaping from local minima
- Morris
- 1993
(Show Context)
Citation Context ...developed the distributed breakout[15] for solving distributed CSPs. This method is characterized by hill-climbing in parallel while excluding neighbors' simultaneous action[6] and the breakout method=-=[10]-=- as a strategy for escaping from quasi-local-minima. In the distributed breakout, each agent rst initializes its assignment arbitrarily, sends its assignment to neighbors with ok? messages, and then r... |

144 | Constraint hierarchies
- Borning, Freeman-Benson, et al.
- 1992
(Show Context)
Citation Context ...extend the distributed CSP framework to enable handling of over-constrained distributed CSPs. In [13], Yokoo proposed a method for over-constrained distributed CSPs by introducing constraint hierarchy=-=[1]-=- and relaxing the less important constraints if there exists no solution. This method can be applied to problems where constraints are hierarchically structured. However, we recognize that constraints... |

114 | The distributed vehicle monitoring testbed: A tool for investigating distributed problem solving networks
- Lesser, Corkill
- 1983
(Show Context)
Citation Context ...ributed among multiple agents and the agents are required to satisfy all constraints by communicating with each other. Many problems in multi-agent systems, such as distributed interpretation problems=-=[9]-=-, distributed resource allocation problems[3], distributed scheduling problems[11], and multi-agent truth maintenance systems[7], can be formalized as distributed CSPs.sOn the other hand, when a probl... |

94 |
Multistage negotiation for distributed constraint satisfaction
- Conry, Kuwabara, et al.
- 1991
(Show Context)
Citation Context ...are required to satisfy all constraints by communicating with each other. Many problems in multi-agent systems, such as distributed interpretation problems[9], distributed resource allocation problems=-=[3]-=-, distributed scheduling problems[11], and multi-agent truth maintenance systems[7], can be formalized as distributed CSPs.sOn the other hand, when a problem designer tries to describe a real problem ... |

90 | Distributed constrained heuristic search
- Sycara, Roth, et al.
- 1991
(Show Context)
Citation Context ...nts by communicating with each other. Many problems in multi-agent systems, such as distributed interpretation problems[9], distributed resource allocation problems[3], distributed scheduling problems=-=[11]-=-, and multi-agent truth maintenance systems[7], can be formalized as distributed CSPs.sOn the other hand, when a problem designer tries to describe a real problem as a CSP, the resulting CSP is often ... |

86 | Distributed breakout algorithm for solving distributed constraint satisfaction problems
- Yokoo, Hirayama
- 1996
(Show Context)
Citation Context ... allow agents to assign or change their variable values in parallel, and thus SBB cannot take advantage of parallelism. 5.3 Iterative Distributed Breakout Outline We developed the distributed breakout=-=[15]-=- for solving distributed CSPs. This method is characterized by hill-climbing in parallel while excluding neighbors' simultaneous action[6] and the breakout method[10] as a strategy for escaping from q... |

85 |
Distributed Truth Maintenance
- Bridgeland, Huhns
- 1990
(Show Context)
Citation Context ...lems in multi-agent systems, such as distributed interpretation problems[9], distributed resource allocation problems[3], distributed scheduling problems[11], and multi-agent truth maintenance systems=-=[7]-=-, can be formalized as distributed CSPs.sOn the other hand, when a problem designer tries to describe a real problem as a CSP, the resulting CSP is often over-constrained and has no solutions. For suc... |

20 |
Phase transition in max-csp
- Larrosa, Meseguer
- 1996
(Show Context)
Citation Context ...e classes of random binary CSPs with n = m = 10, p1 taking values from f18=45� 27=45� 36=45� 45=45g, and p2 from f0:8� 0:9g. These classes of problems are known to be relatively hard ones for Max-CSPs=-=[8]-=-. Accordingly, we believe that they are suitable for the problems used to evaluate the methods. Both SBB and IDB are implemented on a discrete event simulator that simulates concurrent activities of m... |

18 | Constraint relaxation in distributed constraint satisfaction problem
- Yokoo
- 1993
(Show Context)
Citation Context ...ot giveany indication of howwe should handle over-constrained distributed CSPs. Thus it makes sense to extend the distributed CSP framework to enable handling of over-constrained distributed CSPs. In =-=[13]-=-, Yokoo proposed a method for over-constrained distributed CSPs by introducing constraint hierarchy[1] and relaxing the less important constraints if there exists no solution. This method can be appli... |

17 |
Conjunctive width heuristics for maximal constraint satisfaction
- Wallace, Freuder
- 1993
(Show Context)
Citation Context ...e used conjunctive widthsGlobal distance 8 7 6 5 4 3 2 SBB IDB 0 1000 2000 3000 Cycle 4000 5000 6000 Optimal Fig. 5. Anytime curve for an instance of h10� 10� 27=45� 0:9i heuristics(width/domain-size)=-=[12]-=- for variable ordering and lexical order for value ordering. Also note that the initial value of 8iNi was set to the value of the maximum degree of a constraint graph minus one, and that of 8iSi was z... |

10 |
Forming Coalitions for Breaking Deadlocks
- Hirayama, Toyoda
- 1995
(Show Context)
Citation Context ...ibuted Breakout Outline We developed the distributed breakout[15] for solving distributed CSPs. This method is characterized by hill-climbing in parallel while excluding neighbors' simultaneous action=-=[6]-=- and the breakout method[10] as a strategy for escaping from quasi-local-minima. In the distributed breakout, each agent rst initializes its assignment arbitrarily, sends its assignment to neighbors w... |