Hybrid cloud computing combines private clouds with geographically-distributed resources from public clouds, desktop grids or in-house gateways to provide the most flexibility of each kind of cloud platforms. Service ...
详细信息
Hybrid cloud computing combines private clouds with geographically-distributed resources from public clouds, desktop grids or in-house gateways to provide the most flexibility of each kind of cloud platforms. Service provisioning for wide-area applications such as cloud backup or cloud network games is sensitive to wide-area network metrics such as round trip time, bandwidth, or loss rates. In order to optimize the quality of the service provision in hybrid clouds, it is highly valuable for the hybrid clouds to collect detailed network metrics between participating nodes of the hybrid clouds. However, since nodes can be large-scale and dynamic, the network metrics may be diverse for different cloud services, it is challenging to increase the generality, scalability, accuracy, and the robustness of the measurement process. We propose a novel distributed level monitoring method HPM (Hierarchical Performance Measurement) satisfying these requirements. For each kind of network metric, HPM represents the degree of pairwise closeness with discrete level values inspired by the hierarchical clustering tree. HPM maps probed metric to discrete levels based on an existing distributed K-means clustering method that helps maximize the similarity of the network metric in the same level, which therefore optimizes the matching between pairwise levels and the real-world pairwise proximity. Furthermore, for scalability reasons, HPM computes the pairwise levels with decentralized coordinates. Each node independently maintains its low-dimensional coordinate based on a novel decentralized implementation of the Maximum Margin Matrix Factorization method, which optimizes the mapping between the network metrics and the level values. Simulation results for the round trip time, bandwidth, loss, and hop count metric confirm that HPM converges fast, is robust to parameter settings, scales well with increasing levels or system size, and adapts well to diverse metrics. A prototype deployment on
This paper uses a charging selection concept for plug-in electric vehicles (PEVs) to maximize user convenience levels while meeting predefined circuit-level demand limits. The optimal PEV-charging selection problem re...
详细信息
This paper uses a charging selection concept for plug-in electric vehicles (PEVs) to maximize user convenience levels while meeting predefined circuit-level demand limits. The optimal PEV-charging selection problem requires an exhaustive search for all possible combinations of PEVs in a power system, which cannot be solved for the practical number of PEVs. Inspired by the efficiency of the convex relaxation optimization tool in finding close-to-optimal results in huge search spaces, this paper proposes the application of the convex relaxation optimization method to solve the PEV-charging selection problem. Compared with the results of the uncontrolled case, the simulated results indicate that the proposed PEV-charging selection algorithm only slightly reduces user convenience levels, but significantly mitigates the impact of the PEV-charging on the power system. We also develop a distributed optimization algorithm to solve the PEV-charging selection problem in a decentralized manner, i.e., the binary charging decisions (charged or not charged) are made locally by each vehicle. Using the proposed distributed optimization algorithm, each vehicle is only required to report its power demand rather than report several of its private user state information, mitigating the security problems inherent in such problem. The proposed decentralized algorithm only requires low-speed communication capability, making it suitable for real-time implementation.
In many large network settings, such as computer networks, social networks, or hyperlinked text documents, much information can be obtained from the network's spectral properties. However. traditional centralized ...
详细信息
In many large network settings, such as computer networks, social networks, or hyperlinked text documents, much information can be obtained from the network's spectral properties. However. traditional centralized approaches for computing eigenvectors struggle with at least two obstacles: the data may be difficult to obtain (both due to technical reasons and because of privacy concerns), and the sheer size of the networks makes the computation expensive. A decentralized, distributed algorithm addresses both of these obstacles: it utilizes the computational power of all nodes in the network and their ability to communicate, thus speeding up the computation with the network size. And as each node knows its incident edges, the data collection problem is avoided as well, Our main result is a simple decentralized algorithm for computing the top k eigenvectors, of a symmetric weighted adjacency matrix. and a proof that it converges essentially in O(tau(mix), log(2) n) rounds of communication and computation, where tau(mix), is the mixing time of a random walk on the network. An additional contribution of our work is a decentralized way of actually detecting convergence, and diagnosing the current error. Our protocol scales well, in that the amount of computation performed at any node in any one round, and the sizes of messages sent, depend linearly on the degree of the node, polynomially on k, but not at all on the (typically much larger) number n of nodes. To achieve independence of n, the coordinates of the computed eigenvectors are held locally by the nodes to which they correspond, enabling many eigenanalyses without distributing complete global state. (c) 2007 Elsevier Inc. All rights reserved.
This paper proposes a new type of range-free localization method based on affine transformation. Nodes extract subgraphs with a grid topology from a sensor network and assign x-y coordinates to themselves in a decentr...
详细信息
ISBN:
(纸本)9781467324472;9781467324458
This paper proposes a new type of range-free localization method based on affine transformation. Nodes extract subgraphs with a grid topology from a sensor network and assign x-y coordinates to themselves in a decentralized manner. The nodes estimate their positions using an affine transformation based on the mapping of the physical positions and the x-y coordinates of three anchors in an extracted graph. In contrast with multilateration-based localization methods, the proposed method works well even in a non-convex hull deployment, such as a terrain with big regions without sensors. We provide a theoretical analysis and simulation results. We also present a strategy for minimizing the position estimation error and maximizing the coverage of the proposed method. In the simulation results, the position estimation error is 0.18 (normalized by the radio communication range) and the coverage is almost 100% in a non-convex hull deployment.
This paper introduces algorithms for the decentralized low-rank matrix completion problem. Assume a low-rank matrix W - [W-1, W-2, ..., W-L]. In a network, each agent l observes some entries of W-l. In order to recove...
详细信息
ISBN:
(纸本)9781467300469
This paper introduces algorithms for the decentralized low-rank matrix completion problem. Assume a low-rank matrix W - [W-1, W-2, ..., W-L]. In a network, each agent l observes some entries of W-l. In order to recover the unobserved entries of W via decentralized computation, we factorize the unknown matrixWas the product of a public matrix X, common to all agents, and a private matrix Y = [Y-1, Y-2, ..., Y-L], where Y-l is held by agent l. Each agent l alternatively updates Y-l and its local estimate of X while communicating with its neighbors toward a consensus on the estimate. Once this consensus is ( nearly) reached throughout the network, each agent l recoversW(l) = XYl and thus W is recovered. The communication cost is scalable to the number of agents, and W-l and Y-l are kept private to agent l to a certain extent. The algorithm is accelerated by extrapolation and compares favorably to the centralized code in terms of recovery quality and robustness to rank over-estimate.
In many large network settings, such as computer networks, social networks, or hyperlinked text documents, much information can be obtained from the network's spectral properties. However. traditional centralized ...
详细信息
In many large network settings, such as computer networks, social networks, or hyperlinked text documents, much information can be obtained from the network's spectral properties. However. traditional centralized approaches for computing eigenvectors struggle with at least two obstacles: the data may be difficult to obtain (both due to technical reasons and because of privacy concerns), and the sheer size of the networks makes the computation expensive. A decentralized, distributed algorithm addresses both of these obstacles: it utilizes the computational power of all nodes in the network and their ability to communicate, thus speeding up the computation with the network size. And as each node knows its incident edges, the data collection problem is avoided as well, Our main result is a simple decentralized algorithm for computing the top k eigenvectors, of a symmetric weighted adjacency matrix. and a proof that it converges essentially in O(tau(mix), log(2) n) rounds of communication and computation, where tau(mix), is the mixing time of a random walk on the network. An additional contribution of our work is a decentralized way of actually detecting convergence, and diagnosing the current error. Our protocol scales well, in that the amount of computation performed at any node in any one round, and the sizes of messages sent, depend linearly on the degree of the node, polynomially on k, but not at all on the (typically much larger) number n of nodes. To achieve independence of n, the coordinates of the computed eigenvectors are held locally by the nodes to which they correspond, enabling many eigenanalyses without distributing complete global state. (c) 2007 Elsevier Inc. All rights reserved.
Conventional algorithms for autonomous trajectory planning of multiple aircraft tend to require prohibitive computational time as the number of concerned aircraft increases. To overcome this drawback, this paper prese...
详细信息
ISBN:
(纸本)9788995003879
Conventional algorithms for autonomous trajectory planning of multiple aircraft tend to require prohibitive computational time as the number of concerned aircraft increases. To overcome this drawback, this paper presents a fast and decentralized trajectory planning algorithm which can be executed in parallel. The developed algorithm combines force field method for conflict reduction and quadratic programming for trajectory optimization. Due to the parallel nature, the computational time in the developed algorithm is not as sensitive to the number of concerned aircraft as the conventional algorithms. Several results of numerical simulations are presented to demonstrate the performance of the developed algorithm.
In this paper, we introduce a novel approach to improve overall lifetime in mobile ad hoc networks. Given the energy constraint on each node, this problem is formulated as an energy-controlled load balancing problem. ...
详细信息
In this paper, we introduce a novel approach to improve overall lifetime in mobile ad hoc networks. Given the energy constraint on each node, this problem is formulated as an energy-controlled load balancing problem. Thus, our approach is quite different from usual energy-efficient routing or topology control methods. The proposed algorithm is fully distributed and ensures that each node will cooperate in proportion to its remaining energy, increasing the network lifetime. The relevance of the algorithm is evaluated through both theoretical analysis and simulations. (C) 2010 Elsevier B.V. All rights reserved.
In this paper we address the trajectory tracking problem for groups of mobile robots. We consider trajectories described by completely arbitrary shaped closed curves. The proposed control strategy is a completely dece...
详细信息
ISBN:
(纸本)9781424466757
In this paper we address the trajectory tracking problem for groups of mobile robots. We consider trajectories described by completely arbitrary shaped closed curves. The proposed control strategy is a completely decentralized algorithm, and does not require any global synchronization. The desired behavior is obtained by means of some properly designed artificial potential functions.
The rapidly growing demand for wireless communication makes efficient power allocation a critical factor in the network's efficient operation. Power allocation in decentralized wireless systems, where the transmis...
详细信息
The rapidly growing demand for wireless communication makes efficient power allocation a critical factor in the network's efficient operation. Power allocation in decentralized wireless systems, where the transmission of a user creates interference to other users and directly affects their utilities, has been recently studied by pricing methods. However, pricing methods do not result in efficient/optimal power allocations for such systems for the following reason. Systems where a user's actions directly affect the utilities of other users are known to have externalities. It is well known from Mas-Colell et al. that in systems with externalities, standard efficiency theorems on market equilibrium do not apply and pricing methods do not result in Pareto optimal outcomes. In this paper, we formulate the power allocation problem for a wireless network as an allocation problem with "externalities." We consider a system where each user knows only its own utility and the channel gains from the transmitters of other users to its own receiver. The system has multiple interference temperature constraints to control interference. We present a decentralized algorithm to allocate transmission powers to the users. The algorithm takes into account the externality generated to the other users by the transmission of each user, satisfies the informational constraints of the system, overcomes the inefficiency of pricing mechanisms and guarantees convergence to globally optimal power allocations.
暂无评论