#### DMCA

## Adopt: asynchronous distributed constraint optimization with quality guarantees (2005)

### Cached

### Download Links

Venue: | ARTIFICIAL INTELLIGENCE LABORATORY, MASSACHUSETTS INSTITUTE OF TECHNOLOGY |

Citations: | 315 - 51 self |

### Citations

1987 |
Distributed Algorithms
- Lynch
- 1996
(Show Context)
Citation Context ...In this paper, we will assume the DFS ordering is done in a preprocessing step so every agent already knows its parent and children. Several distributed algorithms for forming DFS trees already exist =-=[7, 12, 19]-=- which do not require central control but only that agents have unique ids. Variable value assignments (VALUE messages) are sent down the DFS tree while cost feedback (COST messages) percolate back up... |

1145 | Temporal constraint networks.
- Dechter, Meiri, et al.
- 1991
(Show Context)
Citation Context ...leader agent, the approach may not apply in situations where agents may only communicate with neighboring agents. 27Modi, Shen, Tambe & Yokoo 8.3 Other Work in DCOP R. Dechter, A. Dechter, and Pearl =-=[8]-=- present a theoretical analysis of the constraint optimization problem establishing complexity results in terms of the structure of the constraint graph and global optimization function. In addition, ... |

570 | Towards flexible teamwork.
- Tambe
- 1997
(Show Context)
Citation Context ...ion problems such as distributed planning, distributed scheduling, distributed resource allocation and others [13, 14, 18, 24]. Satellite constellations [2], disaster rescue [15], multiagent teamwork =-=[29]-=-, human/agent organizations [5], intelligent forces [4], distributed and reconfigurable robots [26] and sensor networks [28] are a just a few examples of multiagent applications where distributed reas... |

527 | Depth-first iterative-deepening: An optimal admissible tree search
- Korf
- 1985
(Show Context)
Citation Context ...rony of the algorithm, which enables concurrency. SynchID is an algorithm we have constructed in order to isolate the causes of speedups obtained by Adopt. SynchID simulates iterative deepening search=-=[16]-=- in a distributed environment by requiring agents to execute sequentially and synchronously. SynchID’s search strategy is similar to Adopt in that both algorithms use the lower bounds to do backtracki... |

332 | Valued Constraint Satisfaction Problems: Hard and Easy Problems,
- Schiex, Fargier, et al.
- 1995
(Show Context)
Citation Context ...esent apply only to aggregation operators that are associative, commutative, and monotonic. This class of optimization functions is described formally by Schiex, Fargier and Verfaillie as Valued CSPs =-=[25]-=- and by Bistarelli, Montanari and Rossi as Semi-Ring CSPs [3]. Monotonicity requires that the cost of a solution can only increase as more costs are aggregated. For example, summation over the natural... |

295 | Distributed Constraint Satisfaction for Formalizing Distributed Problem Solving. - Yokoo, Durfee, et al. - 1992 |

132 | An asynchronous complete method in distributed constraint optimization,” in
- Modi, Shen, et al.
- 2003
(Show Context)
Citation Context ...l technique for bounded-error approximation. A bounded-error 1. Additional details may also be found in the first author’s PhD thesis [23]. This article is an extension of an earlier conference paper =-=[21]-=-. Additional exposition, examples and experiments are presented here. 3Modi, Shen, Tambe & Yokoo approximation algorithm is guaranteed to deliver a solution whose quality is within a userspecified di... |

112 | Robocup rescue: Search and rescue in large-scale disasters as a domain for autonomous agents research
- Kitano, Tadokoro, et al.
- 1999
(Show Context)
Citation Context ...ty of multiagent coordination problems such as distributed planning, distributed scheduling, distributed resource allocation and others [13, 14, 18, 24]. Satellite constellations [2], disaster rescue =-=[15]-=-, multiagent teamwork [29], human/agent organizations [5], intelligent forces [4], distributed and reconfigurable robots [26] and sensor networks [28] are a just a few examples of multiagent applicati... |

109 | A dynamic distributed constraint satisfaction approach to resource allocation. - Modi, Jung, et al. - 2001 |

108 | Constraint Solving over Semirings
- Bistarelli, Montanari, et al.
- 1995
(Show Context)
Citation Context ..., commutative, and monotonic. This class of optimization functions is described formally by Schiex, Fargier and Verfaillie as Valued CSPs [25] and by Bistarelli, Montanari and Rossi as Semi-Ring CSPs =-=[3]-=-. Monotonicity requires that the cost of a solution can only increase as more costs are aggregated. For example, summation over the natural numbers is monotonic but summation over the integers is not.... |

90 | Backtracking in distributed constraint networks.
- Hamadi, Bessiere, et al.
- 1998
(Show Context)
Citation Context ...In this paper, we will assume the DFS ordering is done in a preprocessing step so every agent already knows its parent and children. Several distributed algorithms for forming DFS trees already exist =-=[7, 12, 19]-=- which do not require central control but only that agents have unique ids. Variable value assignments (VALUE messages) are sent down the DFS tree while cost feedback (COST messages) percolate back up... |

79 | Electric elves: Applying agent technology to support human organizations.
- Chalupsky, Gil, et al.
- 2001
(Show Context)
Citation Context ... planning, distributed scheduling, distributed resource allocation and others [13, 14, 18, 24]. Satellite constellations [2], disaster rescue [15], multiagent teamwork [29], human/agent organizations =-=[5]-=-, intelligent forces [4], distributed and reconfigurable robots [26] and sensor networks [28] are a just a few examples of multiagent applications where distributed reasoning problems arise. DCOP prov... |

79 | Distributed constraint satisfaction algorithm for complex local problems. - Yokoo, Hirayama - 1998 |

77 | Taking advantage of stable sets of variables in constraint satisfaction problems.
- Freuder, Quinn
- 1985
(Show Context)
Citation Context ...ized in a Depth-First Search (DFS) tree which defines parent/child relationships between the agents. The DFS tree ordering is equivalent to the pseudo-tree arrangements described by Freuder and Quinn =-=[11]-=-. The use of DFS trees has been proposed by Collins, Dechter and Katz in the context of DisCSP [6]. For a given constraint graph, a DFS tree is valid if there are no constraints between agents in diff... |

72 | Distributed partial constraint satisfaction problem.
- Hirayama, Yokoo
- 1997
(Show Context)
Citation Context ...Constraint Optimization Problem (DCOP) for modeling a wide variety of multiagent coordination problems such as distributed planning, distributed scheduling, distributed resource allocation and others =-=[13, 14, 18, 24]-=-. Satellite constellations [2], disaster rescue [15], multiagent teamwork [29], human/agent organizations [5], intelligent forces [4], distributed and reconfigurable robots [26] and sensor networks [2... |

69 | On the feasibility of distributed constraint satisfaction
- Collin, Dechter, et al.
- 1991
(Show Context)
Citation Context .... The DFS tree ordering is equivalent to the pseudo-tree arrangements described by Freuder and Quinn [11]. The use of DFS trees has been proposed by Collins, Dechter and Katz in the context of DisCSP =-=[6]-=-. For a given constraint graph, a DFS tree is valid if there are no constraints between agents in different subtrees of the DFS tree. Constraints are only allowed between an agent and its ancestors or... |

60 |
Modsaf behavior simulation and control.
- Calder, Smith, et al.
- 1993
(Show Context)
Citation Context ...cheduling, distributed resource allocation and others [13, 14, 18, 24]. Satellite constellations [2], disaster rescue [15], multiagent teamwork [29], human/agent organizations [5], intelligent forces =-=[4]-=-, distributed and reconfigurable robots [26] and sensor networks [28] are a just a few examples of multiagent applications where distributed reasoning problems arise. DCOP provides a useful framework ... |

60 | Hormones for self-reconfigurable robots”.
- Shen, Salemi, et al.
- 2000
(Show Context)
Citation Context ...and others [13, 14, 18, 24]. Satellite constellations [2], disaster rescue [15], multiagent teamwork [29], human/agent organizations [5], intelligent forces [4], distributed and reconfigurable robots =-=[26]-=- and sensor networks [28] are a just a few examples of multiagent applications where distributed reasoning problems arise. DCOP provides a useful framework for investigating how agents can coordinate ... |

57 | Dynamic prioritization of complex agents in distributed constraint satisfaction problems.
- Armstrong, Durfee
- 1997
(Show Context)
Citation Context ...n an important advance and provides key insights that influence the work presented here. However, existing distributed search methods for DisCSP do not generalize easily to DCOP. Armstrong and Durfee =-=[1]-=- investigate the effect of agent priority orderings on efficiency in DisCSP. They show that variable ordering heuristics from CSP can be reused as priority orderings in DisCSP and that dynamic reorder... |

48 | Exploiting Problem Structure for Distributed Constraint Optimization.
- Liu, Sycara
- 1995
(Show Context)
Citation Context ...Constraint Optimization Problem (DCOP) for modeling a wide variety of multiagent coordination problems such as distributed planning, distributed scheduling, distributed resource allocation and others =-=[13, 14, 18, 24]-=-. Satellite constellations [2], disaster rescue [15], multiagent teamwork [29], human/agent organizations [5], intelligent forces [4], distributed and reconfigurable robots [26] and sensor networks [2... |

45 | Distributed Forward Checking - Brito, Meseguer - 2003 |

37 |
Distributed breakout revisited
- Zhang, Wittenburg
- 2002
(Show Context)
Citation Context ...and Bound (SynchBB) algorithm [13] discussed later, is prohibitively slow since it requires synchronous, sequential communication. Other fast, asynchronous solutions, such as variants of local search =-=[13, 32]-=-, cannot provide guarantees on the quality of the solutions they find. As we can see from the above, one of the main obstacles for solving DCOP is combining quality guarantees with asynchrony. Previou... |

32 | Self-stabilizing depth-first search
- Collin, Dolev
- 1994
(Show Context)
Citation Context ...In this paper, we will assume the DFS ordering is done in a preprocessing step so every agent already knows its parent and children. Several distributed algorithms for forming DFS trees already exist =-=[7, 12, 19]-=- which do not require central control but only that agents have unique ids. Variable value assignments (VALUE messages) are sent down the DFS tree while cost feedback (COST messages) percolate back up... |

27 | An approach to overconstrained distributed constraint satisfaction problems: Distributed hierarchical constraint satisfaction.
- Hirayama, Yokoo
- 2000
(Show Context)
Citation Context ...nces where the obtained cost was at the given distance from optimal. Table 1: Characteristics of Distributed Constraint Optimization Methods Method Asynch? Optimal? Dist? Satisfaction-Based Search[18]=-=[14]-=- N N Y Local Search [13][9] Y N Y Synchronous Search [13] N Y Y Greedy Repair [17] N N N Asynchronous Best-First Search (Adopt) Y Y Y [27] present an alternative representation of DisCSP in which cons... |

25 | An experimental assessment of a stochastic, anytime, decentralized, soft colourer for sparse graphs,”
- Fitzpatrick, Meertens
- 2001
(Show Context)
Citation Context ... was at the given distance from optimal. Table 1: Characteristics of Distributed Constraint Optimization Methods Method Asynch? Optimal? Dist? Satisfaction-Based Search[18][14] N N Y Local Search [13]=-=[9]-=- Y N Y Synchronous Search [13] N Y Y Greedy Repair [17] N N N Asynchronous Best-First Search (Adopt) Y Y Y [27] present an alternative representation of DisCSP in which constraints are assigned to age... |

23 | Distributed Sensor Nets: A Multiagent Perspective. - Lesser, Ortiz, et al. - 2001 |

20 | An Incomplete Method for Solving Distributed Valued ConstraintSatisfaction Problems
- Lemaitre, Verfaillie
- 1997
(Show Context)
Citation Context ...acteristics of Distributed Constraint Optimization Methods Method Asynch? Optimal? Dist? Satisfaction-Based Search[18][14] N N Y Local Search [13][9] Y N Y Synchronous Search [13] N Y Y Greedy Repair =-=[17]-=- N N N Asynchronous Best-First Search (Adopt) Y Y Y [27] present an alternative representation of DisCSP in which constraints are assigned to agents while variables are shared between agents. This app... |

17 | Resource Allocation in Networks Using Abstraction and Constraint Satisfaction Techniques
- Frei, Faltings
- 1999
(Show Context)
Citation Context ...ates the unique difficulties of interdependencies between sub-problems in distributed problem solving and illustrates the applicability of the distributed constraint representation. Frei and Faltings =-=[10]-=- focus on modelling bandwidth resource allocation as a CSP. Although they do not deal with distributed systems, they show how the use of abstraction techniques in the constraint modelling of real prob... |

17 |
Distributed component-centered design as agent-based distributed constraint optimization
- Parunak, Ward, et al.
- 1997
(Show Context)
Citation Context ...Constraint Optimization Problem (DCOP) for modeling a wide variety of multiagent coordination problems such as distributed planning, distributed scheduling, distributed resource allocation and others =-=[13, 14, 18, 24]-=-. Satellite constellations [2], disaster rescue [15], multiagent teamwork [29], human/agent organizations [5], intelligent forces [4], distributed and reconfigurable robots [26] and sensor networks [2... |

16 | Distributed intelligent backtracking - Hamadi, Bessière, et al. - 1998 |

15 |
Distributed Constraint Satisfaction:Foundation of Cooperation in Multi-agent Systems
- Yokoo
- 2001
(Show Context)
Citation Context ...ed as functions that return a range of values, rather than predicates that return only true or false. DCOP significantly generalizes the Distributed Constraint Satisfaction Problem (DisCSP) framework =-=[20, 27, 30]-=- in which problem solutions are characterized with a designation of “satisfactory or unsatisfactory” and so do not model problems where solutions have degrees of quality or cost. DCOP demands techniqu... |

14 |
Distributed Constraint Optimization for Multiagent Systems
- Modi
- 2003
(Show Context)
Citation Context ...s and built-in termination detection naturally leads to a practical technique for bounded-error approximation. A bounded-error 1. Additional details may also be found in the first author’s PhD thesis =-=[23]-=-. This article is an extension of an earlier conference paper [21]. Additional exposition, examples and experiments are presented here. 3Modi, Shen, Tambe & Yokoo approximation algorithm is guarantee... |

12 |
Faltings, Asynchronous search with aggregations, in
- Silaghi, Sam-Haroud, et al.
(Show Context)
Citation Context ...ed as functions that return a range of values, rather than predicates that return only true or false. DCOP significantly generalizes the Distributed Constraint Satisfaction Problem (DisCSP) framework =-=[20, 27, 30]-=- in which problem solutions are characterized with a designation of “satisfactory or unsatisfactory” and so do not model problems where solutions have degrees of quality or cost. DCOP demands techniqu... |

11 |
Autonomy architectures for a constellation of spacecraft
- Barrett
- 1999
(Show Context)
Citation Context ...modeling a wide variety of multiagent coordination problems such as distributed planning, distributed scheduling, distributed resource allocation and others [13, 14, 18, 24]. Satellite constellations =-=[2]-=-, disaster rescue [15], multiagent teamwork [29], human/agent organizations [5], intelligent forces [4], distributed and reconfigurable robots [26] and sensor networks [28] are a just a few examples o... |

11 | Distributed constraint reasoning under unreliable communication
- Modi, Ali, et al.
- 2003
(Show Context)
Citation Context ...g lower bounds to perform distributed optimization. In this section, we discuss some possible algorithmic modifications to Adopt. Algorithm modifications for unreliable communication are discussed in =-=[22]-=-. In addition, we are aware that the ordering of variables has a dramatic effect on the efficiency of the DCOP algorithm. In future work, we will develop distributed methods for discovering efficient ... |

6 |
Distributed forward checking
- Mesequer, Jiménez
- 2000
(Show Context)
Citation Context ...ed as functions that return a range of values, rather than predicates that return only true or false. DCOP significantly generalizes the Distributed Constraint Satisfaction Problem (DisCSP) framework =-=[20, 27, 30]-=- in which problem solutions are characterized with a designation of “satisfactory or unsatisfactory” and so do not model problems where solutions have degrees of quality or cost. DCOP demands techniqu... |

3 |
Ecm challenge problem
- Systems
- 1998
(Show Context)
Citation Context ...4]. Satellite constellations [2], disaster rescue [15], multiagent teamwork [29], human/agent organizations [5], intelligent forces [4], distributed and reconfigurable robots [26] and sensor networks =-=[28]-=- are a just a few examples of multiagent applications where distributed reasoning problems arise. DCOP provides a useful framework for investigating how agents can coordinate their decision-making in ... |

1 | Self-stabilizing depth first - Collin, Dolev - 1994 |