PC-DPOP: A New Partial Centralization Algorithm for Distributed Optimization
نویسندگان
چکیده
Fully decentralized algorithms for distributed constraint optimization often require excessive amounts of communication when applied to complex problems. The OptAPO algorithm of [Mailler and Lesser, 2004] uses a strategy of partial centralization to mitigate this problem. We introduce PC-DPOP, a new partial centralization technique, based on the DPOP algorithm of [Petcu and Faltings, 2005]. PCDPOP provides better control over what parts of the problem are centralized and allows this centralization to be optimal with respect to the chosen communication structure. Unlike OptAPO, PC-DPOP allows for a priory, exact predictions about privacy loss, communication, memory and computational requirements on all nodes and links in the network. Upper bounds on communication and memory requirements can be specified. We also report strong efficiency gains over OptAPO in experiments on three problem domains.
منابع مشابه
PC-DPOP: A Partial Centralization Extension of DPOP
A number of algorithms have been recently proposed for distributed constraint optimization problems (DCOP). Fully decentralized examples include ADOPT [11] (search) or DPOP [12] (dynamic programming). Difficult optimization problems have been shown to require important amounts of communication when these algorithms are used (large number of small messages for ADOPT, or few but large messages fo...
متن کاملSymDPOP: Adapting DPOP to exploit partial symmetries
This work proposes a new approach to dealing with symmetries in Distributed Constraint Optimisation Problems by adapting the DPOP algorithm [1]. In contrast to an already proposed distributed preprocessing method leading to a problem redefinition [2], we present here a method that exploits the structure of DFS trees, with no explicit redefinition of the problem. We exhibit the flexibility in th...
متن کاملMB-DPOP: A New Memory-Bounded Algorithm for Distributed Optimization
In distributed combinatorial optimization problems, dynamic programming algorithms like DPOP ([Petcu and Faltings, 2005]) require only a linear number of messages, thus generating low communication overheads. However, DPOP’s memory requirements are exponential in the induced width of the constraint graph, which may be prohibitive for problems with large width. We present MB-DPOP, a new hybrid a...
متن کاملImproving DPOP with function filtering
DPOP is an algorithm for distributed constraint optimization which has, as main drawback, the exponential size of some of its messages. Recently, some algorithms for distributed cluster tree elimination have been proposed. They also suffer from exponential size messages. However, using the strategy of cost function filtering, in practice these algorithms obtain important reductions in maximum m...
متن کاملGeneralizing DPOP: Action-GDL, a new complete algorithm for DCOPs
In this paper we made three main contributions (fully detailed in [5]). Firstly, we formulate a new algorithm, the so-called Action-GDL, that takes inspiration from GDL [1], extending and applying it to Distributed Constraint Optimization Problems (DCOPs). Secondly, we show the generality of Action-GDL showing how it generalizes DPOP[4], one of the low complexity, state-of-the-art algorithm to ...
متن کامل