R. Dechter, A. Dechter, and J. Pearl, `Optimization in constraint networks ', in Influence Diagrams, Belief Nets and Decision Analysis, eds., R.M. Oliver and J.Q. Smith, 411--425, John Wiley & Sons, (1990).  Home/Search   Document Not in Database   Summary   Related Articles   Check

....where only one agent is allowed to change variable value at a time. Since all agents must communicate with a single leader agent, the approach may not apply in situations where agents may only communicate with neighboring agents. 28 8.3 Other Work in DCOP R. Dechter, A. Dechter, and Pearl [3] 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, they outline an approach for distributed search for the optimal solution based on dynamic ....

R. Dechter, A. Dechter, and J. Pearl. Optimization in constraint networks. In In uence Diagrams, Belief Nets, and Decision Analysis. 1990.

....with probabilities. By defining Omega differently (e.g. based on addition) our relational data model can be easily extended to solve a number of apparently different but closely related problems such as dynamic programming [2] solving sparse linear equations [11] and constraint propagation [3]. ....

R. Dechter, A. Dechter and J. Pearl, "Optimization in constraint networks," in Influence Diagrams, Belief Nets, and Decision Analysis, edited by R.M. Oliver and J.Q. Smith, Wiley, 1990.

....if u has an additive decomposition over the set of maximal cliques of G. One of the reasons why this result is interesting is that the functional form generated by CA independencies is precisely the form that has often been assumed to hold of a utility function in work on computing MEU, e.g. [JJD94, DDP88, ST90]. Nevertheless, we concede that it remains far from clear to us whether CA independence is the best or most natural independence notion to use. Additional graphical models that can capture and reason with other forms of utility independencies are needed. Such work may eventually lead to models ....

R. Dechter, A. Dechter, and J. Pearl. Optimization in constraint networks. In R. M. Oliver and J. Q. Smith, editors, Influence Diagrams Belief Nets and Decision Analysis, pages 411--425. Wiley, 1988.

....constructive search algorithms: for a given variable, choose the value which appears in the most solutions. Similar proposals for counting solutions or estimating solution probabilities have been made [6, 14, 4, 11, 16, 15] Solution probabilities are orthogonal to preference over solutions (e.g. [3, 1]) or probabilistic constraints (e.g. 1] in which there is uncertainty regarding whether a constraint applies. The main purpose of this paper is to document a connection between constraint reasoning and probabilistic reasoning. We show that pAC algorithm is a generalization of the basic arc ....

....15] the propagation of solution probabilities can perform some kinds of optimization by giving domain elements nonuniform a priori weights. We are looking at using propagation techniques to provide heuristic information for optimization of different kinds of objective functions, extending both [3, 1]. Acknowledgements Thanks to Matt Vernooy for providing the data from his thesis, and to the anonymous referees for their helpful comments. ....

Rina Dechter, Avi Dechter, and Judea Pearl. Optimization in constraint networks. In Influence Diagrams, Belief Nets and Decision Analysis, pages 411-- 425. John Wiley and Sons Ltd, 1990.

....either oriented along the rows or the columns in the roster. As it will be shown later, the constraint graph of this subproblem is a hyper tree. This observation motivates a further investigation of efficient procedures for solving tree structured constraint problems [Freuder, 1982, Freuder, 1990, Dechter et al. 1990, Freuder and Wallace, 1992 ] As a consequence, this paper addresses opportunities to extend branch and bound search by the cycle cutset method [Dechter and Pearl, 1987, Dechter, 1990] This improved branch and bound is then used to perform the the repair step in Algorithm 1. For this reason, ....

.... of efficient procedures for solving tree structured constraint problems [Freuder, 1982, Freuder, 1990, Dechter et al. 1990, Freuder and Wallace, 1992 ] As a consequence, this paper addresses opportunities to extend branch and bound search by the cycle cutset method [Dechter and Pearl, 1987, Dechter, 1990] . This improved branch and bound is then used to perform the the repair step in Algorithm 1. For this reason, line 3 of Algorithm 1 returns two sets of variables to indicate a subproblem: # # and ## . The variables in ## are supposed to form a tree structured subproblem that can be solved by ....

[Article contains additional citation context not shown here]

Rina Dechter, Avi Dechter, and Judea Pearl. Optimization in constraint networks. In R. M. Oliver and J. Q. Smith, editors, Influence Diagrams, Belief Nets and Decision Analysis, pages 411--425. John Wiley & Sons, 1990.

....graph of these instances. An interesting property of these algorithms (most are dynamic programming like algorithms) is that they can be easily extended to cope with extensions of CSP such as Valued CSP [SFV95] or Semiring CSP [BMR95] or more generally with constraint optimization problems [BB72, DDP90, SS88] Although this is our main motivation, this paper will stick to classical CSP, considering that the extension of these algorithms to optimization is not the hardest part and has already been tackled in several cases. The constraint graph of a CSP is defined by representing each variable ....

....For optimization problems (e.g. an extended version of Max CSP in which each tuple has a cost) projection must simply be extended: the cost of a tuple t in the induced constraint will be the optimum cost of solutions of the subproblem than contains t. This has been popularized by many authors [DDP90, She91, BMR95, BMR97] A nice property of these dynamic programming algorithms is that they are not restricted to optimization but can also perform counting or more generally, compute discrete integrals on solutions. See [Arn85] for an example on graph reliability problems. This property makes ....

R. Dechter, A. Dechter, and J. Pearl. Optimization in constraint networks. In R.M Oliver and J.Q. Smith, editors, Influence Diagrams, Belief Nets and Decision Analysis, chapter 18, pages 411--425. John Wiley & Sons Ltd., 1990.

....either oriented along the rows or the columns in the roster. As it will be shown later, the constraint graph of this subproblem is a hyper tree. This observation motivates a further investigation of efficient procedures for solving tree structured constraint problems [Freuder, 1982, Freuder, 1990, Dechter et al. 1990, Freuder and Wallace, 1992] As a consequence, this paper addresses opportunities to extend branch and bound search by the cycle cutset method [Dechter and Pearl, 1987, Dechter, 1990] This improved branch and bound is then used to perform the the repair step in Algorithm 1. For this reason, ....

.... of efficient procedures for solving tree structured constraint problems [Freuder, 1982, Freuder, 1990, Dechter et al. 1990, Freuder and Wallace, 1992] As a consequence, this paper addresses opportunities to extend branch and bound search by the cycle cutset method [Dechter and Pearl, 1987, Dechter, 1990] . This improved branch and bound is then used to perform the the repair step in Algorithm 1. For this reason, line 3 of Algorithm 1 returns two sets of variables to indicate a subproblem: X 0 and X T . The variables in X T are supposed to form a tree structured subproblem that can be solved by ....

[Article contains additional citation context not shown here]

Rina Dechter, Avi Dechter, and Judea Pearl. Optimization in constraint networks. In R. M. Oliver and J. Q. Smith, editors, Influence Diagrams, Belief Nets and Decision Analysis, pages 411--425. John Wiley & Sons, 1990.

....than one variable at a time can be eliminated. In fact, in our algorithm all variables in the right hand side of a rule are considered at a time. Moreover, a related algorithm which solves a constraint problem with a tree like shape in a bottom up way, as in Definition 48, has been described in [DDP90] for optimization problems and in [DD88] for belief maintenance. In those papers, the idea is to consider either the constraint graph, if it is acyclic, or the dual graph of a constraint problem (where nodes are constraints and arcs are associated with variables shared among constraints) and to ....

....if it is acyclic, or the dual graph of a constraint problem (where nodes are constraints and arcs are associated with variables shared among constraints) and to use techniques like cycle cutset [DP88a] or tree clustering [DP88b] to provide such a dual graph with a tree like shape. However, in [DDP90] constraints are combined via the usual and operator, and the value associated to each tuple is computed in a completely independent way, via a given utility function. Thus, apart from the tree structure, our approach and that in [DDP90] are very different, since we also generalize the way ....

[Article contains additional citation context not shown here]

R. Dechter, A. Dechter, and J. Pearl. Optimization in constraint networks. In R.M. Oliver and J.Q. Smith, editors, Influence Diagrams, Belief Nets and Decision Analysis. John Wiley and Sons Ltd., 1990.

....parfaitement au premier point. Il n est pas surprenant qu un nombre important de travaux aient tente d apporter des reponses aux problemes de prise en compte d incertitudes, de preferences et d imprecisions sur les donnees. Nous citerons pour memoire les travaux de [ROS 76, SHA 81, HER 88, FRE 89, DEC 90, SCH 92, MC 92, FRE 92, BOW 92, BRE 92, FAR 93a, FAR 94a, FAR 94b, RUT 94, GUE 94] qui offrent chacun : une extension du formalisme CSP permettant d evaluer une solution potentielle en terme de cou t [SHA 81, FRE 92] de possibilite [ROS 76, FAR 94a] de priorite [SCH 92] de probabilite [FAR ....

.... Une approche similaire a ete suivie, par exemple, dans [MIN 76, GON 90] pour les problemes de cheminement dans les graphes, ou dans [SHA 91] pour etendre le plus largement possible le cadre d application d une classe d algorithmes de propagation dans les hyper arbres (utilise dans les CSP [DEC 90, FRE 92] ou enfin dans [BIS 95] qui s appuie sur une structure de diode etendue pour etudier les possibles generalisations des algorithmes d etablissement de la k coherence. La section 2 definit ce qu est un CSP value, justifie rapidement la structure alge brique utilisee et enumere quelques ....

[Article contains additional citation context not shown here]

DECHTER R., DECHTER A. et PEARL J., Optimization in constraint networks. In : Influence Diagrams, Belief Nets and Decision Analysis, ed. par OLIVER R. et SMITH J., chap. 18, p. 411--425, John Wiley & Sons Ltd., 1990.

....the problem is altered because a relationship that was not allowed in the original problem is now acceptable. 2 There has been some work on the selective non satisfaction of constraints [Freuder 92] and other work that can be interpreted as making contributions toward relaxation [Rosenfeld 76] Dechter 90] A well grounded theory is lacking. Such a theory has wide applications to any problem expressible in the constraint satisfaction paradigm. 1.2 Motivation Constraint relaxation is applicable in two general areas. The first is in guiding search for a solution to an original problem. If we relax ....

....one of the variables to another value in its domain. If all values in the domain of a variable have been unsuccessfully tried, the variable is left unassigned and backtracking moves to another variable. 6. Other conceptualizations exist, including the dual constraint graph and join graph [Dechter 90] In general, a constraint graph is a hypergraph with arcs connecting all relevant nodes. 9 The simplest retrospective technique is chronological backtracking. If the compatibility check fails, another value is tried for the most recently assigned variable. If all values in the domain of the ....

[Article contains additional citation context not shown here]

Dechter, R., Dechter, A., and Pearl, J. Optimization in Constraint Networks. Influence Diagrams, Belief Nets, and Decision Analysis. In Oliver, R.M. and Smith, J.Q., John Wiley and Sons, Ltd, Chicester, England, 1990.

....methods, an observation that was made earlier in the context of constraint satisfaction tasks [20] The observation that a variety of tasks allow efficient algorithms of hypertrees and therefore can benefit from a tree clustering approach was recognized by several works in the last decade. In [38] the connection between optimization and constraint satisfaction and its relationship to dynamic programming is explicated. In the work of [33, 47] and later in [6] an axiomatic framework that characterize tasks that can be solved polynomially over hyper trees, is introduced. Such tasks can be ....

A. Dechter R. Dechter and J. Pearl. Optimization in constraint networks. In Influence Diagrams, Belief Nets and Decision Analysis, pages 411--425. John Wiley & Sons, 1990.

....the same scope of a constraint or in the same functional component of the criterion function. Since constraint optimization can be performed in linear time when the augmented constraint graph is a tree, both join tree clustering and cutsetconditioning can extend the method to non tree structures [32] in the usual manner. We can conclude: Theorem 8: 32] Time space of constraint optimization] Given a constraint optimization problem over n variables whose augmented constraint graph has a cycle cutset of size c, and whose augmented graph can be embedded in a clique tree having tree width r and ....

....component of the criterion function. Since constraint optimization can be performed in linear time when the augmented constraint graph is a tree, both join tree clustering and cutsetconditioning can extend the method to non tree structures [32] in the usual manner. We can conclude: Theorem 8: [32][Time space of constraint optimization] Given a constraint optimization problem over n variables whose augmented constraint graph has a cycle cutset of size c, and whose augmented graph can be embedded in a clique tree having tree width r and separator size s, the time complexity of finding an ....

A. Dechter R. Dechter and J. Pearl. Optimization in constraint networks. In Influence Diagrams, Belief Nets and Decision Analysis, pages 411-- 425. John Wiley & Sons, 1990. 36

....This work was supported in part by the NSF (grant IRI9157636) the Air Force Office of Scientific Research (grant AFOSR 900136) Toshiba of America, Xerox, Northrop and Rockwell. particular, it was shown that when a constraint network is acyclic, an optimal solution can be found in linear time [ 13 ] . Tailoring this algorithm to diagnosis results in an algorithm called structure based abduction (SAB) 12 ] which will be empirically investigated here. The performance of SAB is compared with two model based diagnosis algorithms called here MBD1 and MBD2. MBD1 uses the same strategy for ....

R. Dechter, A. Dechter, and J. Pearl. Optimization in constraint networks. In R.M. Olivier and J.Q. Smith, editors, Influence Diagrams, Belief Nets and Decision Analysis, pages 411--425. J. Wiley, New York, 1990.

....the same scope of a constraint or in the same functional component of the criterion function. Since constraint optimization can be performed in linear time when the augmented constraint graph is a tree, both join tree clustering and cutset conditioning can extend the method to non tree structures [22] in the usual manner. We can conclude: Theorem 24 (Time space of constraint optimization [22] Given a constraint optimization problem whose augmented constraint graph can be embedded in a clique tree having tree width r and separator width s and a cyclecutset size c, the time complexity of ....

....Since constraint optimization can be performed in linear time when the augmented constraint graph is a tree, both join tree clustering and cutset conditioning can extend the method to non tree structures [22] in the usual manner. We can conclude: Theorem 24 (Time space of constraint optimization [22]) Given a constraint optimization problem whose augmented constraint graph can be embedded in a clique tree having tree width r and separator width s and a cyclecutset size c, the time complexity of finding an optimal consistent solution using tree clustering is O(n Delta exp(r) and the space ....

R. Dechter, A. Dechter, and J. Pearl. Optimization in constraint networks. In Influence Diagrams, Belief Nets and Decision Analysis, pages 411--425. John Wiley & Sons, 1990.

....time. We then extend it to general topologies by dividing the network into fictitious tree like subnetworks using the cycle cutset scheme. The algorithm is based on the method of nonserial dynamic programming methods (Bertel e Brioschi, 1972) which was also used for constraint optimization (Dechter, Dechter, Pearl, 1990). There the task was divided between a precompilation into a tree structure via a tree clustering algorithm and a run time optimization over the tree. Our adaptation is connectionist in style; i.e. the algorithm can be stated as a simple, uniform activation function (Rumelhart, Hinton, ....

....this section we discuss generalizations of our algorithm to nodes that are part of cycles, that will work well for near tree networks. A full account of this extension is deferred for future work. A well known scheme for extending tree algorithms to non tree networks, is cycle cutset decomposition (Dechter, 1990), used in Bayes networks and constraint networks. Cyclecutset decomposition is based on the fact that an instantiated variable cuts the flow of information on any path on which it lies and therefore it changes the effective connectivity of the network. Consequently, when the group of instantiated ....

Dechter, R., Dechter, A., & Pearl, J. (1990). Optimization in constraint networks. In in R.M. Oliver, J. S. (Ed.), Influence diagrams, belief nets and decision analysis. John Wiley and Sons.

No context found.

R. Dechter, A. Dechter, and J. Pearl, `Optimization in constraint networks ', in Influence Diagrams, Belief Nets and Decision Analysis, eds., R.M. Oliver and J.Q. Smith, 411--425, John Wiley & Sons, (1990).

No context found.

R. Dechter, A. Dechter, and J. Pearl, `Optimization in constraint networks ', in Influence Diagrams, Belief Nets and Decision Analysis, eds., R.M. Oliver and J.Q. Smith, 411--425, John Wiley & Sons, (1990).

No context found.

R. Dechter, A. Dechter, and J. Pearl, `Optimization in constraint networks ', in Influence Diagrams, Belief Nets and Decision Analysis, eds., R.M. Oliver and J.Q. Smith, 411--425, John Wiley & Sons, (1990).

No context found.

R. Dechter, A. Dechter, and J. Pearl, `Optimization in constraint networks ', in Influence Diagrams, Belief Nets and Decision Analysis, eds., R.M. Oliver and J.Q. Smith, 411--425, John Wiley & Sons, (1990).

No context found.

R. Dechter, A. Dechter, and J. Pearl, `Optimization in constraint networks ', in Influence Diagrams, Belief Nets and Decision Analysis, eds., R.M. Oliver and J.Q. Smith, 411--425, John Wiley & Sons, (1990).

No context found.

R. Dechter, A. Dechter, and J. Pearl. Optimization in constraint networks. In Influence Diagrams, Belief Nets, and Decision Analysis. 1990.

No context found.

R. Dechter, A. Dechter, and J. Pearl (1990). Optimization in constraint networks. In R. M. Oliver and J. Q. Smith, editors, Influence Diagrams, Belief Nets and Decision Analysis, chapter 18, pages 411--425. John Wiley & Sons Ltd.

No context found.

R. Dechter, A. Dechter, and J. Pearl. Optimization in constraint networks. In Influence Diagrams, Belief Nets, and Decision Analysis. 1990.

No context found.

R. Dechter, A. Dechter, and J. Pearl. Optimization in constraint networks. In R.M. Oliver and J.Q. Smith, editors, In uence Diagrams, Belief Nets and Decision Analysis, pages 411-425. John Wiley and Sons, 1990.

No context found.

Dechter, R. ; Dechter, A. ; et Pearl, J. 1990. Optimization in constraint networks. Dans Oliver, R., et Smith, J., eds., Influence Diagrams, Belief Nets and Decision Analysis. John Wiley & Sons Ltd. chapitre 18, 411--425. 13, 20

First 50 documents

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