Results 1  10
of
13
W.C.: On modeling multiagent task scheduling as a distributed constraint optimization problem
 In: IJCAI
"... This paper investigates how to represent and solve multiagent task scheduling as a Distributed Constraint Optimization Problem (DCOP). Recently multiagent researchers have adopted the C TÆMS language as a standard for multiagent task scheduling. We contribute an automated mapping that transforms C T ..."
Abstract

Cited by 17 (3 self)
 Add to MetaCart
This paper investigates how to represent and solve multiagent task scheduling as a Distributed Constraint Optimization Problem (DCOP). Recently multiagent researchers have adopted the C TÆMS language as a standard for multiagent task scheduling. We contribute an automated mapping that transforms C TÆMSinto a DCOP. Further, we propose a set of representational compromises for C TÆMS that allow existing distributed algorithms for DCOP to be immediately brought to bear on C TÆMS problems. Next, we demonstrate a key advantage of a constraint based representation is the ability to leverage the representation to do efficient solving. We contribute a set of preprocessing algorithms that leverage existing constraint propagation techniques to do variable domain pruning on the DCOP. We show that these algorithms can result in 96 % reduction in state space size for a given set of C TÆMS problems. Finally, we demonstrate up to a 60 % increase in the ability to optimally solve C TÆMS problems in a reasonable amount of time and in a distributed manner as a result of applying our mapping and domain pruning algorithms. 1
Dynamic Configuration of Agent Organizations
"... It is useful to impose organizational structure over multiagent coalitions. Hierarchies, for instance, allow for compartmentalization of tasks: if organized correctly, tasks in disjoint subtrees of the hierarchy may be performed in parallel. Given a notion of the way in which a group of agents need ..."
Abstract

Cited by 10 (0 self)
 Add to MetaCart
It is useful to impose organizational structure over multiagent coalitions. Hierarchies, for instance, allow for compartmentalization of tasks: if organized correctly, tasks in disjoint subtrees of the hierarchy may be performed in parallel. Given a notion of the way in which a group of agents need to interact, the Dynamic Distributed Multiagent Hierarchy Generation (DynDisMHG) problem is to determine the best hierarchy that might expedite the process of coordination. This paper introduces a distributed algorithm, called Mobed, for both constructing and maintaining organizational agent hierarchies, enabling exploitation of parallelism in distributed problem solving. The algorithm is proved correct and it is shown that individual additions of agents to the hierarchy will run in an amortized linear number of rounds. The hierarchies resulting after perturbations to the agent coalition have constantbounded edit distance, making Mobed very well suited to highly dynamic problems. 1
Hierarchical variable ordering for multiagent agreement problems
 In AAMAS
, 2006
"... The Multiagent Agreement Problem (MAP) is a special form of Distributed Constraint Optimization (DCOP) that requires agents to choose values for variables to satisfy not only their own constraints, but also equality constraints with other agents. For solving MAPs, we introduce the AdoptMVA algorithm ..."
Abstract

Cited by 8 (0 self)
 Add to MetaCart
(Show Context)
The Multiagent Agreement Problem (MAP) is a special form of Distributed Constraint Optimization (DCOP) that requires agents to choose values for variables to satisfy not only their own constraints, but also equality constraints with other agents. For solving MAPs, we introduce the AdoptMVA algorithm which is an extension of the existing Adopt algorithm designed to take advantage of the partial centralization that exists in MAP domains where agents control multiple variables. Second, while existing solution approaches to DCOP require variables to be prioritized in some fashion in order to guarantee optimality, it is unclear how to order variables effectively when agents own multiple variables. We investigate a hierarchical approach which leverages known ordering techniques from the sequential constraint satisfaction literature by combining ordering at the agent level with orderings at the variable level to obtain efficient global orderings. Finally, we identify a promising technique for converting known effective variable orderings into effective agent orderings and identify an intraagent variable ordering heuristic for MAP that is the most efficient of the ones tested. While the contributions presented in this paper are applicable to general DCOPs, we focus our discussion on MAPs because we feel it is a significant problem class worthy of specific attention. 1.
Evaluation of CBR on Live Networks
"... Abstract. A large class of problems in multiagent systems can be solved by distributed constraint optimization (DCOP). Several algorithms have been created to solve these problems, however, no extensive evaluation of current DCOP algorithms on live networks exists in the literature. This paper uses ..."
Abstract

Cited by 6 (4 self)
 Add to MetaCart
(Show Context)
Abstract. A large class of problems in multiagent systems can be solved by distributed constraint optimization (DCOP). Several algorithms have been created to solve these problems, however, no extensive evaluation of current DCOP algorithms on live networks exists in the literature. This paper uses DCOPolis—a framework for comparing and deploying DCOP software in heterogeneous environments—to contribute an analysis of two stateoftheart DCOP algorithms run in various network environments solving a number of different problem types. Then, we use this empirical validation to evaluate the use of both cyclebased runtime and concurrent constraint checks. 1
NogoodBased Asynchronous ForwardChecking Algorithms
, 2012
"... We propose two new algorithms for solving Distributed Constraint Satisfaction Problems (DisCSPs). The first algorithm, AFCng, is a nogoodbased version of Asynchronous Forward Checking (AFC). Besides its use of nogoods as justification of value removals, AFCng allows simultaneous backtracks going ..."
Abstract

Cited by 5 (4 self)
 Add to MetaCart
(Show Context)
We propose two new algorithms for solving Distributed Constraint Satisfaction Problems (DisCSPs). The first algorithm, AFCng, is a nogoodbased version of Asynchronous Forward Checking (AFC). Besides its use of nogoods as justification of value removals, AFCng allows simultaneous backtracks going from different agents to different destinations. The second algorithm, Asynchronous Forward Checking Tree (AFCtree), is based on the AFCng algorithm and is performed on a pseudotree ordering of the constraint graph. AFCtree runs simultaneous search processes in disjoint problem subtrees and exploits the parallelism inherent in the problem. We prove that AFCng and AFCtree only need polynomial space. We compare the performance of these algorithms with other DisCSP algorithms on random DisCSPs and instances from real benchmarks: sensor networks and distributed meeting scheduling. Our experiments show that AFCng improves on AFC and that AFCtree outperforms all compared algorithms, particularly on sparse problems. 1
ADOPTing: Unifying Asynchronous Distributed Optimization with Asynchronous Backtracking
, 2007
"... This article presents an asynchronous algorithm for solving Distributed Constraint Optimization problems (DCOPs). The proposed technique unifies asynchronous backtracking (ABT) and asynchronous distributed optimization (ADOPT) where valued nogoods enable more flexible reasoning and more opportunitie ..."
Abstract

Cited by 4 (1 self)
 Add to MetaCart
(Show Context)
This article presents an asynchronous algorithm for solving Distributed Constraint Optimization problems (DCOPs). The proposed technique unifies asynchronous backtracking (ABT) and asynchronous distributed optimization (ADOPT) where valued nogoods enable more flexible reasoning and more opportunities for communication, leading to an important speedup. While feedback can be sent in ADOPT by COST messages only to one predefined predecessor, our extension allows for sending such information to any relevant agent. The concept of valued nogood is an extension by Dago and Verfaille of the concept of classic nogood that associates the list of conflicting assignments with a cost and, optionally, with a set of references to culprit constraints. DCOPs have been shown to have very elegant distributed solutions, such as ADOPT, distributed asynchronous overlay (DisAO), or DPOP. These algorithms are typically tuned to minimize the longest causal chain of messages as a measure of how the algorithms will scale for systems with remote agents (with large latency in communication). ADOPT has the property of maintaining the initial distribution of the problem. To be efficient, ADOPT needs a preprocessing step consisting of computing a DepthFirst Search (DFS) tree on the constraint graph. Valued nogoods allow for automatically detecting and exploiting the best DFS tree compatible with the current ordering. To exploit such DFS trees it is now sufficient to ensure that they exist. Also, the inference rules available for valued nogoods help to exploit schemes of communication where more feedback is sent to higher priority agents. Together they result in an order of magnitude improvement.
General Terms Algorithms
"... ABSTRACT Many different multiagent problems, such as distributed scheduling, can be formalized as distributed constraint optimization problems (DCOP [1]). Ordering the constraint variables is an important preprocessing step of the ADOPT algorithm [1], the state of the art method of solving DCOP. Cu ..."
Abstract
 Add to MetaCart
(Show Context)
ABSTRACT Many different multiagent problems, such as distributed scheduling, can be formalized as distributed constraint optimization problems (DCOP [1]). Ordering the constraint variables is an important preprocessing step of the ADOPT algorithm [1], the state of the art method of solving DCOP. Currently ADOPT uses depthfirst search (DFS) trees for that purpose. For certain classes of tasks DFS ordering does not exploit the problem structure as compared to pseudotree ordering [3]. Also the variables are currently ordered by using a centralized scheme, which requires global information about the problem structure. We present a variable ordering algorithm, which is both decentralized and makes use of pseudotrees, thus exploiting the problem structure when possible. This allows to apply ADOPT to domains, where global information is unavailable, and find solutions more efficiently. The worstcase pseudotree depth resulting from our algorithm is p2kV , where V is the set of variables, and k is maximum cluster size in constraint graph. The algorithm has space and time complexity polynomial in size of the constraint graph.
A Decentralized Variable Ordering Method for Distributed Constraint Optimization
"... Many different multiagent problems, such as distributed scheduling, can be formalized as distributed constraint optimization problems (DCOP [1]). Ordering the constraint variables is an important preprocessing step of the ADOPT algorithm [1], the state of the art method of solving DCOP. Currently A ..."
Abstract
 Add to MetaCart
(Show Context)
Many different multiagent problems, such as distributed scheduling, can be formalized as distributed constraint optimization problems (DCOP [1]). Ordering the constraint variables is an important preprocessing step of the ADOPT algorithm [1], the state of the art method of solving DCOP. Currently ADOPT uses depthfirst search (DFS) trees for that purpose. For certain classes of tasks DFS ordering does not exploit the problem structure as compared to pseudotree ordering [3]. Also the variables are currently ordered by using a centralized scheme, which requires global information about the problem structure. We present a variable ordering algorithm, which is both decentralized and makes use of pseudotrees, thus exploiting the problem structure when possible. This allows to apply ADOPT to domains, where global information is unavailable, and find solutions more efficiently. The worstcase pseudotree depth resulting from our algorithm is p 2kV , where V is the set of variables, and k is maximum cluster size in constraint graph. The algorithm has space and time complexity polynomial in size of the constraint graph.