## Algorithms for Distributed Constraint Satisfaction: A Review (2000)

### Cached

### Download Links

- [www.kecl.ntt.co.jp]
- [www.cs.bgu.ac.il]
- [lang.is.kyushu-u.ac.jp]
- [www.owlnet.rice.edu]
- [www.eecs.harvard.edu]
- [jmvidal.cse.sc.edu]
- DBLP

### Other Repositories/Bibliography

Venue: | In CP |

Citations: | 201 - 7 self |

### BibTeX

@INPROCEEDINGS{Yokoo00algorithmsfor,

author = {Makoto Yokoo and Katsutoshi Hirayama},

title = {Algorithms for Distributed Constraint Satisfaction: A Review},

booktitle = {In CP},

year = {2000},

pages = {2000}

}

### Years of Citing Articles

### OpenURL

### Abstract

. When multiple agents are in a shared environment, there usually exist constraints among the possible actions of these agents. A distributed constraint satisfaction problem (distributed CSP) is a problem to find a consistent combination of actions that satisfies these inter-agent constraints. Various application problems in multi-agent systems can be formalized as distributed CSPs. This paper gives an overview of the existing research on distributed CSPs. First, we briefly describe the problem formalization and algorithms of normal, centralized CSPs. Then, we show the problem formalization and several MAS application problems of distributed CSPs. Furthermore, we describe a series of algorithms for solving distributed CSPs, i.e., the asynchronous backtracking, the asynchronous weak-commitment search, the distributed breakout, and distributed consistency algorithms. Finally,we showtwo extensions of the basic problem formalization of distributed CSPs, i.e., handling multiple local variables, and dealing with over-constrained problems. Keywords: Constraint Satisfaction, Search, distributed AI 1.

### Citations

943 | Temporal constraint networks
- Dechter, Meiri, et al.
- 1991
(Show Context)
Citation Context ...n CSPs, Tsang’s textbook [22] on constraint satisfaction covers topics from basic concepts to recent research results. There are several concise overviews of constraint satisfaction problems, such as =-=[6, 15]-=-. 3. Problem definition of distributed CSP A distributed CSP is a CSP in which the variables and constraints are distributed among automated agents. Finding a value assignment to variables that satisf... |

813 | Foundations of Constraint Satisfaction
- Tsang
- 1993
(Show Context)
Citation Context ...computational costs, so we must find an appropriate combination of consistency algorithms and backtracking so that the total search costs are minimized. For further readings on CSPs, Tsang’s textbook =-=[22]-=- on constraint satisfaction covers topics from basic concepts to recent research results. There are several concise overviews of constraint satisfaction problems, such as [6, 15]. 3. Problem definitio... |

618 |
A Truth Maintenance System
- Doyle
- 1979
(Show Context)
Citation Context ...ributed CSPs Various application problems in MAS can be formalized as distributed CSPs. For example, a multi-agent truth maintenance system [13] is a distributed version of a truth maintenance system =-=[7]-=-. In this system, there exist multiple agents, each of which has its own truth maintenance system (Figure 2). Each agent has uncertain data that can be IN or OUT, i.e., believed or not believed, and e... |

577 | Where the really hard problems are
- Cheeseman, Kanefsky, et al.
- 1991
(Show Context)
Citation Context ...can assume that constraints among agents are rather sparse. The setting where m = n × 2�7 has been identified as a critical setting which produces particularly difficult, phase-transition problems in =-=[2]-=-. We can see that the distributed breakout outperforms the asynchronous weakcommitment search when problem instances are critically difficult. In the distributed breakout, each mode (wait ok? or wait ... |

422 | R.J.: Partial constraint satisfaction
- Freuder, Wallace
- 1992
(Show Context)
Citation Context ...roblem closer to the original problem, and in some case they want to make the relaxation minimal. A distributed partial CSP can be formalized using terms in a partial CSP, which has been presented in =-=[9]-=- for dealing with over-constrained centralized CSPs. It can be defined using the following components: — a set of agents, A =�1� 2�����m�, — ��Pi�Ui�� �PSi� ≤��Mi� for each agent i, — �G� �N� S��. For... |

397 | P.: Minimizing conflicts: A heuristic repair method for constraint satisfaction and scheduling problems
- Minton, Johnston, et al.
- 1992
(Show Context)
Citation Context ...les and values greatly affects the efficiency of the algorithm. Various heuristics have been developed during the long history of CSP studies. A value-ordering heuristic called min-conflict heuristic =-=[17]-=- is a very successful one among these heuristics. In the min-conflict backtracking, each variable has a tentative initial value. The tentative initial value is revised when the variable is added to th... |

268 | The distributed constraint satisfaction problem: formalization and algorithms - Yokoo, Durfee, et al. - 1998 |

253 |
Constraint satisfaction
- Mackworth
- 1992
(Show Context)
Citation Context ...s). Even though the definition of a CSP is very simple, a surprisingly wide variety of AI problems can be formalized as CSPs. Therefore, the research on CSP has a long and distinguished history in AI =-=[15]-=-. A distributed CSP is a CSP in which variables and constraints are distributed among multiple automated agents. Various application problems in Multi-agent Systems (MAS) that are concerned with findi... |

249 | Distributed Constraint Satisfaction: foundations of cooperation in multi-agent systems - Yokoo - 2001 |

245 |
Synthesizing Constraint Expressions
- Freuder
- 1978
(Show Context)
Citation Context ...ation. 2.2.3. Consistency algorithms. Consistency algorithms [15] are preprocessing algorithms that reduce futile backtracking. Consistency algorithms can be classified by the notion of k-consistency =-=[8]-=-. A CSP is k-consistent iff given any instantiation of any k − 1 variables satisfying all the constraints among those variables, it is possible to find an instantiation of any kth variable such that t... |

194 |
The breakout method for escaping from local minima
- Morris
- 1993
(Show Context)
Citation Context ...the number of constraint violations cannot be decreased by changing any single variable value. Various methods have been proposed for escaping from localminima. For example, in the breakout algorithm =-=[18]-=-, a weight is defined for each constraint (the initial weight is 1). The summation of the weights of violated constraints is used as an evaluation value. When trapped in a local-minimum, the breakout ... |

94 |
Multistage negotiation for distributed constraint satisfaction
- Conry, Kuwabara, et al.
- 1991
(Show Context)
Citation Context ...e automated agents. Various application problems in Multi-agent Systems (MAS) that are concerned with finding a consistent combination of agent actions (e.g., distributed resource allocation problems =-=[4]-=-, distributed scheduling problems [21], distributed interpretation tasks [16], and multi-agent truth maintenance tasks [13]) can be formalized as distributed CSPs. Therefore, we can consider distribut... |

90 | Distributed constrained heuristic search
- Sycara, Roth, et al.
- 1991
(Show Context)
Citation Context ...on problems in Multi-agent Systems (MAS) that are concerned with finding a consistent combination of agent actions (e.g., distributed resource allocation problems [4], distributed scheduling problems =-=[21]-=-, distributed interpretation tasks [16], and multi-agent truth maintenance tasks [13]) can be formalized as distributed CSPs. Therefore, we can consider distributed algorithms for solving distributed ... |

85 |
Distributed Truth Maintenance
- Bridgeland, Huhns
- 1990
(Show Context)
Citation Context ... combination of agent actions (e.g., distributed resource allocation problems [4], distributed scheduling problems [21], distributed interpretation tasks [16], and multi-agent truth maintenance tasks =-=[13]-=-) can be formalized as distributed CSPs. Therefore, we can consider distributed algorithms for solving distributed CSPs as an important infrastructure in MAS.186 yokoo and hirayama Figure 1. Example ... |

74 |
Asynchronous weak-commitment search for solving distributed constraint satisfaction problems
- Yokoo
- 1995
(Show Context)
Citation Context ... If the value selection of a higher priority agent is bad, the lower priority agents need to perform an exhaustive search to revise the bad decision. The asynchronous weak-commitment search algorithm =-=[25, 27]-=- introduces the min-conflict heuristic to reduce the risk of making bad decisions. Furthermore, the agent ordering is dynamically changed so that a bad decision can be revised without performing an ex... |

70 | Distributed Constraint Satisfaction Algorithm for Complex Local Problems - Yokoo, Durfee - 1998 |

69 | On the feasibility of distributed constraint satisfaction, in
- Collin, Dechter, et al.
- 1991
(Show Context)
Citation Context ...own(p k�l), p k is true under the assignment x j = d j. It must be noted that although algorithms for solving distributed CSPs seem similar to parallel/distributed processing methods for solving CSPs =-=[3, 30]-=-, the research motivations are fundamentally different. 1 The primary concern in parallel/distributed processing is efficiency, and we can choose any type of parallel/distributed computer architecture... |

60 | Distributed partial constraint satisfaction problem
- Hirayama, Yokoo
- 1997
(Show Context)
Citation Context ...e considered a reasonable compromise among agents, since the number of the constraint violations in the worst agent is minimized. Two algorithms for solving distributed maximal CSPs were presented in =-=[12]-=-. One is the synchronous branch and bound algorithm, and the other is the iterative distributed breakout algorithm. The synchronous branch and bound algorithm is a very simple algorithm that simulates... |

57 | Weak-commitment search for solving constraint satisfaction problems
- Yokoo
- 1994
(Show Context)
Citation Context ... can be efficient, but their completeness cannot be guaranteed. There exist several hybrid-type algorithms of backtracking and iterative improvement. For example, the weak-commitment search algorithm =-=[24]-=- is based on the min-conflict backtracking. However, in this algorithm, when for one variable no value satisfies all of the constraints with the partial solution, instead of changing one variable valu... |

56 | A comparison of ATMS and CSP techniques - Kleer - 1989 |

51 | Dynamic prioritization of complex agents in distributed constraint satisfaction problems
- Armstrong, Durfee
- 1997
(Show Context)
Citation Context ...teful to simulate the activities of multiple virtual agents without distinguishing the communications between virtual agents within a single real agent, and the communications between real agents. In =-=[1]-=-, prioritization among agents was introduced to handle multiple local variables (we call this algorithm agent-ordering AWS). In this algorithm, each agent tries to find a local solution that is consis... |

47 |
DATMS: a framework for distributed assumption based reasoning, in
- Mason, Johnson
- 1989
(Show Context)
Citation Context ...) that are concerned with finding a consistent combination of agent actions (e.g., distributed resource allocation problems [4], distributed scheduling problems [21], distributed interpretation tasks =-=[16]-=-, and multi-agent truth maintenance tasks [13]) can be formalized as distributed CSPs. Therefore, we can consider distributed algorithms for solving distributed CSPs as an important infrastructure in ... |

39 | Parallel and distributed algorithms for finite constraint satisfaction problems, in - Zhang, Mackworth - 1991 |

31 | Multiagent coordination in tightly coupled task scheduling
- Liu, Sycara
- 1996
(Show Context)
Citation Context ...ignments was solved using distributed CSP techniques. Many other application problems that are concerned with finding a consistent combination of agent actions/decisions (e.g., distributed scheduling =-=[14, 21]-=- and distributed interpretation problems [16]) can be formalized as distributed CSPs. 5. Algorithms for solving distributed CSPs In this section, we make the following assumptions for simplicity in de... |

24 | Distributed constraint satisfaction for dai problems - Yokoo - 1990 |

18 | Constraint relaxation in distributed constraint satisfaction problem
- Yokoo
- 1993
(Show Context)
Citation Context ... can be considered another reasonable compromise among agents. A simple algorithm for solving distributed hierarchical CSPs, called the asynchronous incremental relaxation algorithm, was presented in =-=[23]-=-. This algorithm repeatedly applies the asynchronous backtracking algorithm to distributed hierarchical CSPs in the following way. In the first stage, agents try to solve an original distributed CSP, ... |

16 |
A constraint maintenance system for the distributed resource allocation problem
- Prosser, Conway, et al.
- 1992
(Show Context)
Citation Context ...breakout). 5.4. Distributed consistency algorithm Achieving 2-consistency by multiple agents is relatively straightforward, since the algorithm can be achieved by the iteration of local processes. In =-=[19]-=-, a distributed system that achieves arc-consistency for resource allocation tasks was developed. This system also maintains arc-consistency, i.e., it can re-achieve arc-consistency after dynamic chan... |

7 |
A multi-agent model for the resource allocation problem: A reactive approach
- Ghedira, Verfaillie
- 1992
(Show Context)
Citation Context ...-1 and L-12. Such a problem can be easily formalized as distributed CSPs, namely, each agent has a variable that represents each goal, and possible values of the variable are plan fragments. Also, in =-=[11]-=-, a multi-agent model for resource allocation problems was developed. In this formalization, there are task agents and resource agents, and these agents cooperatively allocate shared resources with a ... |

7 | Solving a real-life time tabling and transportation problem using distributed CSP techniques
- Solotorevsky, Gudes
- 1996
(Show Context)
Citation Context ...elated constraints are satisfied or not. Time-tabling tasks are another class of application problems that can be formalized as distributed CSPs. For example, the nurse time-tabling task described in =-=[20]-=-algorithms for distributed constraint satisfaction 191 B L-12 B-1 L-11 B-2 L-10 L-9 D-1 L-8 D A-1 A-2 L-1 L-2 C-1 C L-3 C-3 L-4 C-2 L-13 L-5 D-3 L-7 D-2 L-6 A E-2 E L-14 E-1 Figure 3. Distributed res... |

6 |
Kleer, A comparison of ATMS and CSP techniques
- de
- 1989
(Show Context)
Citation Context ...es/values/constraints with a small amount of computational effort by utilizing dependencies. Also, a higher degree of consistency can be achieved using the hyper-resolutionbased consistency algorithm =-=[5]-=-. In [29], a distributed consistency algorithm that achieves k-consistency is described. In this algorithm, agents communicate nogoods among themselves, and generate new nogoods whose length are less ... |

2 | Sycara: 1996, `Multiagent Coordination in Tightly Coupled Task Scheduling - Liu, P |

1 | A Distributed ApproachtoPartial Constraint Satisfaction Problems - Ghedira - 1994 |

1 | Bridgeland: 1991, `MultiagentTruth Maintenance - Huhns, M |

1 |
A distributed approach to partial constraint satisfaction problems
- Ghedira
- 1994
(Show Context)
Citation Context ...ete, i.e., it may fail to find an optimal solution because the distributed breakout algorithm at a certain stage may fail to find a solution.algorithms for distributed constraint satisfaction 203 In =-=[10]-=-, a multi-agent model for solving maximal CSPs was developed. This model is an extension of the multi-agent model for the resource allocation problem described in Section 4. In this model, each variab... |