K. Hirayama and Makoto Yokoo. Distributed partial constraint satisfaction problem. In Principles and Practice of Constraint Programming, pages 222-- 236, 1997.  Home/Search   Document Details and Download   Summary   Related Articles   Check

No context found.

K. Hirayama and M. Yokoo. Distributed partial constraint satisfaction problem. In G. Smolka, editor, Principles and Practice of Constraint Programming, pages 222--236. 1997.

No context found.

K. Hirayama and M. Yokoo. Distributed partial constraint satisfaction problem. In G. Smolka, editor, Principles and Practice of Constraint Programming, pages 222--236. 1997.

....results demonstrating that time to solution decreases as the given error bound is allowed to increase. 1. Introduction A large class of multiagent coordination and distributed resource allocation problems can be modelled via the distributed constraint optimization problem (DCOP) 11] [7]. In DCOP, a set of collaborative agents must optimize over a distributed set of constraints, i.e. nd solutions that meet some quality requirements. Role allocation and reallocation in multia 1. This paper is an extension of a earlier conference paper [15] Full proofs, detailed examples, ....

.... but has failed to apply to general DCOP problems, even rather natural ones such as minimizing the total number of constraint violations (MaxCSP) Other existing algorithms that provide quality guarantees for optimization problems, such as the Synchronous Branch and Bound (SynchBB) algorithm [7] discussed later, are prohibitively slow since they require synchronous, sequential communication. Other fast, asynchronous solutions, such as variants of local search [7] 23] cannot provide guarantees on the quality of the solutions they nd. As we can see from the above, one of the main ....

[Article contains additional citation context not shown here]

K. Hirayama and M. Yokoo. Distributed partial constraint satisfaction problem. In G. Smolka, editor, Principles and Practice of Constraint Programming, pages 222-236. 1997.

....Multiagent Systems General Terms Algorithms Keywords Constraint Satisfaction Optimization, Multi agent Coordination 1. INTRODUCTION A large class of multi agent coordination and distributed resource allocation problems can be modelled via distributed constraint optimization (DCOP) 5] [2]. In DCOP, a set of collaborative agents must optimize over a distributed set of constraints, i.e. find solutions that meet some quality requirements. Multi agent teamwork [ 10] 12] distributed scheduling [5] and distributed sensor networks [8] are some examples of these types of applications. ....

.... , but has failed to apply to general DCOP problems, even rather natural ones such as minimizing the total number of constraint violations (MaxCSP) Other existing algorithms that provide quality guarantees for optimization problems such as the Synchronous Branch and Bound (SynchBB) algorithm [2] discussed later, are prohibitively slow since they require synchronous, sequential communication. Other fast, asynchronous solutions such as variants of local search [2] 14] cannot provide guarantees on the quality of the solutions they find. As we can see from the above, one of the main ....

[Article contains additional citation context not shown here]

K. Hirayama and M. Yokoo. Distributed partial constraint satisfaction proNem. In G. Smolka, editor, Principles and Practice of Constraint Programming, pages 222 236. 1997.

....Multiagent Systems General Terms Algorithms Keywords Constraint Satisfaction Optimization, Multi agent Coordination 1. INTRODUCTION A large class of multi agent coordination and distributed resource allocation problems can be modelled via distributed constraint optimization (DCOP) 5] [2]. In DCOP, a set of collaborative agents must optimize over a distributed set of constraints, i.e. find solutions that meet some quality requirements. Multi agent teamwork [10] 12] distributed scheduling [5] and distributed sensor networks [8] are some examples of these types of applications. ....

.... , but has failed to apply to general DCOP problems, even rather natural ones such as minimizing the total number of constraint violations (MaxCSP) Other existing algorithms that provide quality guarantees for optimization problems such as the Synchronous Branch and Bound (SynchBB) algorithm [2] discussed later, are prohibitively slow since they require synchronous, sequential communication. Other fast, asynchronous solutions such as variants of local search [2] 14] cannot provide guarantees on the quality of the solutions they find. As we can see from the above, one of the main ....

[Article contains additional citation context not shown here]

K. Hirayama and M. Yokoo. Distributed partial constraint satisfaction problem. In G. Smolka, editor, Principles and Practice of Constraint Programming, pages 222--236. 1997.

....cost. Overconstrained problems [3] are an example where no satisfactory solution may be possible and rather than simply returning failure, agents must find high quality solutions, i.e, a solution that is closest to a satisfactory solution. The Distributed Constraint Optimization Problem (DCOP) 5] [2] is a way to model problems where solutions have degrees of quality or cost. It requires a set of collaborative agents to optimize a global objective function that is distributed among them as a set of valued constraints, that is, constraints that are described as functions that return a range of ....

....Despite the fact that agents are asynchronously and concurrently choosing values for their variables, Adopt is able to guarantee globally optimal solution quality. The only existing complete method for DCOP is the Synchronous Branch and Bound (SynchBB) algorithm described by Hirayama and Yokoo[2]. The search strategy in branch and bound search is to update decrease upper bounds during search. Using synchronous computation, SynchBB simulates branch and bound search in a distributed environment. It requires agents to perform computation in a sequential manner in which only one agent ....

[Article contains additional citation context not shown here]

K. Hirayama and M. Yokoo. Distributed partial constraint satisfaction problem. In G. Smolka, editor, Principles and Practice of Constraint Programming -- CP97, pages 222--

....agent tries to satisfy as many of its own constraints as possible, this solution criterion can be 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, 2 A universe is used for relaxing a problem by enlarging a variable domain. As noted in [9] all kinds of relaxation of a CSP can be expressed in terms of relaxing a constraint (enlarging permitted values for variables) byintroducing a ....

Hirayama, K. and M. Yokoo: 1997, `Distributed partial constraint satisfaction problem'. In: Proceedings of the Third International Conference on Principles and Practice of Constraint Programming (CP-97). pp. 222--236. Lecture Notes in Computer Science 1330.

.... application problems, we would rather want to have a partial solution which achieves some partiality criterion (e.g. a partial solution that satisfies as many constraints as possible) We have presented a distributed partial CSP as a general model for handling an over constrained distributed CSP [4]. Intuitively, in a distributed partial CSP, agents search for a solvable distributed CSP and its solution by relaxing an over constrained distributed CSP. By determining the way of relaxing a distributed CSP, we can introduce various partial solutions to a distributed CSP. A distributed maximal ....

....in a distributed partial CSP, agents search for a solvable distributed CSP and its solution by relaxing an over constrained distributed CSP. By determining the way of relaxing a distributed CSP, we can introduce various partial solutions to a distributed CSP. A distributed maximal CSP [4] can be seen as a subclass of a distributed partial CSP. In a distributed maximal CSP, agents search for variable values that minimize the maximal number of violated constraints over agents. As discussed in [4] a distributed maximal CSP has promising application problems in MAS. However, not all ....

[Article contains additional citation context not shown here]

K. Hirayama and M. Yokoo. Distributed partial constraint satisfaction problem. In G. Smolka, editor, Principles and Practice of Constraint Programming --CP97, volume 1330 of Lecture Notes in Computer Science, pages 222--236. Springer-Verlag, 1997.

....never terminates. However, in many application problems in MAS, we sometimes face a situation where we would like to get some partial solution instead of such useless information. We have presented the distributed partial CSP as a general model for handling over constrained distributed CSPs [4]. Intuitively, in the distributed partial CSP, agents search for a solvable distributed CSP and its solution by relaxing an over constrained distributed CSP. By determining the way of relaxing a distributed CSP, we can introduce various partial solutions to a distributed CSP. The distributed ....

....in the distributed partial CSP, agents search for a solvable distributed CSP and its solution by relaxing an over constrained distributed CSP. By determining the way of relaxing a distributed CSP, we can introduce various partial solutions to a distributed CSP. The distributed maximal CSP [4] can be seen as a subclass of distributed partial CSP. In the distributed maximal CSP, agents search for variable values that minimize the maximal number of violated constraints over agents. As discussed in [4] the distributed maximal CSP is one of promising approaches to over constrained ....

[Article contains additional citation context not shown here]

K. Hirayama and M. Yokoo. Distributed partial constraint satisfaction problem. In G. Smolka, editor, Principles and Practice of Constraint Programming --CP97, volume 1330 of Lecture Notes in Computer Science, pages 222--236. Springer-Verlag, 1997.

....do well because it can make better use of the global knowledge of the entire problem. However, considering other aspects like privacy or security for example, we believe such an algorithm is not suitable for MAS application problems. Therefore, we have developed a series of distributed algorithms [13, 21, 22, 23, 24, 25, 26], where agents knowledge of the entire problem stays limited throughout the execution of the algorithms. Among these algorithms, the AWC is basically designed for a distributed CSP where an agent has a CSP with one variable. In the AWC, a priority is defined for each variable. An agent starts ....

Hirayama, K. and Yokoo, M.: Distributed Partial Constraint Satisfaction Problem. In: Smolka, G. (ed.): Principles and Practice of Constraint Programming --CP97. Lecture Notes in Computer Science, Vol.1330. Springer-Verlag (1997) 222--236

No context found.

K. Hirayama and Makoto Yokoo. Distributed partial constraint satisfaction problem. In Principles and Practice of Constraint Programming, pages 222-- 236, 1997.

No context found.

Katsutoshi Hirayama and Makoto Yokoo. Distributed partial constraint satisfaction problem. In Principles and Practice of Constraint Programming, pages 222--236, 1997.

No context found.

Katsutoshi Hirayama and Makato Yokoo. Distributed Partial Constraint Satisfaction Problem. In Proceedings of the Third International Conference on Principles and Practice of Constraint Programming (CP'97), pages 222--236, 1997.

No context found.

Hirayama, K., and Yokoo, M. 1997. Distributed Partial Constraint Satisfaction Problem. In Proceedings of the Third International Conference on Principles and Practice of Constraint Programming (CP'97), 222--236.

No context found.

K. Hirayama and M. Yokoo. Distributed partial constraint satisfaction problem. In G. Smolka, editor, Principles and Practice of Constraint Programming, pages 222--236. 1997.

No context found.

Hirayama, K., and Yokoo, M. 1997. Distributed partial constraint satisfaction problem. In Smolka, G., ed., Proceedings of the Conference on Constraint Processing (CP-97), volume 1330 of LNCS, 222--236.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC