Downloads & Free Reading Options - Results

Phylogenetic networks provide a way to describe and visualize evolutionary histories that have undergone so-called reticulate evolutionary events such as recombination, hybridization or horizontal gene transfer. The level k of a network determines how non-treelike the evolution can be, with level-0 networks being trees. We study the problem of constructing level-k phylogenetic networks from triplets, i.e. phylogenetic trees for three leaves (taxa). We give, for each k, a level-k network that is uniquely defined by its triplets. We demonstrate the applicability of this result by using it to prove that (1) for all k of at least one it is NP-hard to construct a level-k network consistent with all input triplets, and (2) for all k it is NP-hard to construct a level-k network consistent with a maximum number of input triplets, even when the input is dense. As a response to this intractability we give an exact algorithm for constructing level-1 networks consistent with a maximum number of input triplets.

Phylogenetic networks provide a way to describe and visualize evolutionary histories that have undergone so-called reticulate evolutionary events such as recombination, hybridization or horizontal gene transfer. The level k of a network determines how non-treelike the evolution can be, with level-0 networks being trees. We study the problem of constructing level-k phylogenetic networks from triplets, i.e. phylogenetic trees for three leaves (taxa). We give, for each k, a level-k network that is uniquely defined by its triplets. We demonstrate the applicability of this result by using it to prove that (1) for all k of at least one it is NP-hard to construct a level-k network consistent with all input triplets, and (2) for all k it is NP-hard to construct a level-k network consistent with a maximum number of input triplets, even when the input is dense. As a response to this intractability we give an exact algorithm for constructing level-1 networks consistent with a maximum number of input triplets.

Carrier-grade networks comprise several layers where different protocols coexist. Nowadays, most of these networks have different control planes to manage routing on different layers, leading to a suboptimal use of the network resources and additional operational costs. However, some routers are able to encapsulate, decapsulate and convert protocols and act as a liaison between these layers. A unified control plane would be useful to optimize the use of the network resources and automate the routing configurations. Software-Defined Networking (SDN) based architectures, such as OpenFlow, offer a chance to design such a control plane. One of the most important problems to deal with in this design is the path computation process. Classical path computation algorithms cannot resolve the problem as they do not take into account encapsulations and conversions of protocols. In this paper, we propose algorithms to solve this problem and study several cases: Path computation without bandwidth constraint, under bandwidth constraint and under other Quality of Service constraints. We study the complexity and the scalability of our algorithms and evaluate their performances on real topologies. The results show that they outperform the previous ones proposed in the literature.

This paper deals with the statistical signal pro- cessing over graphs for tracking infection diffusion in social networks. Infection (or Information) diffusion is modeled using the Susceptible-Infected-Susceptible (SIS) model. Mean field approximation is employed to approximate the discrete valued infected degree distribution evolution by a deterministic ordinary differential equation for obtaining a generative model for the infection diffusion. The infected degree distribution is shown to follow polynomial dynamics and is estimated using an exact non- linear Bayesian filter. We compute posterior Cramer-Rao bounds to obtain the fundamental limits of the filter which depend on the structure of the network. Considering the time-varying nature of the real world networks, the relationship between the diffusion thresholds and the degree distribution is investigated using generative models for real world networks. In addition, we validate the efficacy of our method with the diffusion data from a real-world online social system, Twitter. We find that SIS model is a good fit for the information diffusion and the non-linear filter effectively tracks the information diffusion.

This paper deals with the statistical signal pro- cessing over graphs for tracking infection diffusion in social networks. Infection (or Information) diffusion is modeled using the Susceptible-Infected-Susceptible (SIS) model. Mean field approximation is employed to approximate the discrete valued infected degree distribution evolution by a deterministic ordinary differential equation for obtaining a generative model for the infection diffusion. The infected degree distribution is shown to follow polynomial dynamics and is estimated using an exact non- linear Bayesian filter. We compute posterior Cramer-Rao bounds to obtain the fundamental limits of the filter which depend on the structure of the network. Considering the time-varying nature of the real world networks, the relationship between the diffusion thresholds and the degree distribution is investigated using generative models for real world networks. In addition, we validate the efficacy of our method with the diffusion data from a real-world online social system, Twitter. We find that SIS model is a good fit for the information diffusion and the non-linear filter effectively tracks the information diffusion.

Genetic algorithms are versatile optimization tools suitable for solving multi-disciplinary optimization problems in aerodynamics where the design parameters may exhibit multi- modal or non-smooth variations. However, the fitness evaluation phase of the algorithms casts a large overhead on the computational requirement and is particularly acute in aerodynamic problems where time-consuming CFD methods are needed for evaluating performance. Methods and strategies to improve the performance of basic genetic algorithms are important to enable the method to be useful for complicated three-dimensional or multi-disciplinary problems. Two such methods are studied in the present work: micro genetic algorithms and artificial neural networks. Both methods are applied to inverse and direct airfoil design problems and the resulting improvement in efficiency is noted and discussed.

We consider the problem of dominating set-based virtual backbone used for routing in asymmetric wireless ad-hoc networks. These networks have non-uniform transmission ranges and are modeled using the well-established disk graphs. The corresponding graph theoretic problem seeks a strongly connected dominating-absorbent set of minimum cardinality in a digraph. A subset of nodes in a digraph is a strongly connected dominating-absorbent set if the subgraph induced by these nodes is strongly connected and each node in the graph is either in the set or has both an in-neighbor and an out-neighbor in it. Distributed algorithms for this problem are of practical significance due to the dynamic nature of ad-hoc networks. We present a first distributed approximation algorithm, with a constant approximation factor and O(Diam) running time, where Diam is the diameter of the graph. Moreover we present a simple heuristic algorithm and conduct an extensive simulation study showing that our heuristic outperforms previously known approaches for the problem.

This paper considers deployment problems where a mobile robotic network must optimize its configuration in a distributed way in order to minimize a steady-state cost function that depends on the spatial distribution of certain probabilistic events of interest. Moreover, it is assumed that the event location distribution is a priori unknown, and can only be progressively inferred from the observation of the actual event occurrences. Three classes of problems are discussed in detail: coverage control problems, spatial partitioning problems, and dynamic vehicle routing problems. In each case, distributed stochastic gradient algorithms optimizing the performance objective are presented. The stochastic gradient view simplifies and generalizes previously proposed solutions, and is applicable to new complex scenarios, such as adaptive coverage involving heterogeneous agents. Remarkably, these algorithms often take the form of simple distributed rules that could be implemented on resource-limited platforms.

This paper considers deployment problems where a mobile robotic network must optimize its configuration in a distributed way in order to minimize a steady-state cost function that depends on the spatial distribution of certain probabilistic events of interest. Moreover, it is assumed that the event location distribution is a priori unknown, and can only be progressively inferred from the observation of the actual event occurrences. Three classes of problems are discussed in detail: coverage control problems, spatial partitioning problems, and dynamic vehicle routing problems. In each case, distributed stochastic gradient algorithms optimizing the performance objective are presented. The stochastic gradient view simplifies and generalizes previously proposed solutions, and is applicable to new complex scenarios, such as adaptive coverage involving heterogeneous agents. Remarkably, these algorithms often take the form of simple distributed rules that could be implemented on resource-limited platforms.

This project considers attack and defense problems on networks with respect to WMD attacks. It provides novel optimization models and solutions for network vulnerability assessment and defense measurement in the face of cascading failures and dynamic attacks. The critical infrastructures considered are complex systems which consist of multiple dynamic independent networks interacting to each other. The attacks we considered are dynamic, that is, another attack may be launched during the recovery.

Modeling tools and technologies are important for aerospace development. At the University of Illinois, we have worked on advancing the state of the art in modeling by Markov reward models in two important areas: reducing the memory necessary to numerically solve systems represented as stochastic activity networks and other stochastic Petri net extensions while still obtaining solutions in a reasonable amount of time, and finding numerically stable and memory-efficient methods to solve for the reward accumulated during a finite mission time. A long standing problem when modeling with high level formalisms such as stochastic activity networks is the so-called state space explosion, where the number of states increases exponentially with size of the high level model. Thus, the corresponding Markov model becomes prohibitively large and solution is constrained by the the size of primary memory. To reduce the memory necessary to numerically solve complex systems, we propose new methods that can tolerate such large state spaces that do not require any special structure in the model (as many other techniques do). First, we develop methods that generate row and columns of the state transition-rate-matrix on-the-fly, eliminating the need to explicitly store the matrix at all. Next, we introduce a new iterative solution method, called modified adaptive Gauss-Seidel, that exhibits locality in its use of data from the state transition-rate-matrix, permitting us to cache portions of the matrix and hence reduce the solution time. Finally, we develop a new memory and computationally efficient technique for Gauss-Seidel based solvers that avoids the need for generating rows of A in order to solve Ax = b. This is a significant performance improvement for on-the-fly methods as well as other recent solution techniques based on Kronecker operators. Taken together, these new results show that one can solve very large models without any special structure.

This article is from Theoretical Biology & Medical Modelling , volume 11 . Abstract Background: Extracting relevant information from microarray data is a very complex task due to the characteristics of the data sets, as they comprise a large number of features while few samples are generally available. In this sense, feature selection is a very important aspect of the analysis helping in the tasks of identifying relevant genes and also for maximizing predictive information. Methods: Due to its simplicity and speed, Stepwise Forward Selection (SFS) is a widely used feature selection technique. In this work, we carry a comparative study of SFS and Genetic Algorithms (GA) as general frameworks for the analysis of microarray data with the aim of identifying group of genes with high predictive capability and biological relevance. Six standard and machine learning-based techniques (Linear Discriminant Analysis (LDA), Support Vector Machines (SVM), Naive Bayes (NB), C-MANTEC Constructive Neural Network, K-Nearest Neighbors (kNN) and Multilayer perceptron (MLP)) are used within both frameworks using six free-public datasets for the task of predicting cancer outcome. Results: Better cancer outcome prediction results were obtained using the GA framework noting that this approach, in comparison to the SFS one, leads to a larger selection set, uses a large number of comparison between genetic profiles and thus it is computationally more intensive. Also the GA framework permitted to obtain a set of genes that can be considered to be more biologically relevant. Regarding the different classifiers used standard feedforward neural networks (MLP), LDA and SVM lead to similar and best results, while C-MANTEC and k-NN followed closely but with a lower accuracy. Further, C-MANTEC, MLP and LDA permitted to obtain a more limited set of genes in comparison to SVM, NB and kNN, and in particular C-MANTEC resulted in the most robust classifier in terms of changes in the parameter settings. Conclusions: This study shows that if prediction accuracy is the objective, the GA-based approach lead to better results respect to the SFS approach, independently of the classifier used. Regarding classifiers, even if C-MANTEC did not achieve the best overall results, the performance was competitive with a very robust behaviour in terms of the parameters of the algorithm, and thus it can be considered as a candidate technique for future studies.

Phases of matter are sharply defined in the thermodynamic limit. One major challenge of accurately simulating quantum phase diagrams of interacting quantum systems is due to the fact that numerical simulations usually deal with the energy density, a local property of quantum wavefunctions, while identifying different quantum phases generally rely on long-range physics. In this paper we construct generic fully symmetric quantum wavefunctions under certain assumptions using a type of tensor networks: projected entangled pair states, and provide practical simulation algorithms based on them. We find that quantum phases can be organized into crude classes distinguished by short-range physics, which is related to the fractionalization of both on-site symmetries and space-group symmetries. Consequently, our simulation algorithms, which are useful to study long-range physics as well, are expected to be able to sharply determine crude classes in interacting quantum systems efficiently. Examples of these crude classes are demonstrated in half-integer quantum spin systems on the kagome lattice. Limitations and generalizations of our results are discussed.

Book Source: Digital Library of India Item 2015.199379 dc.contributor.author: Saleem, Mohammad .S dc.date.accessioned: 2015-07-08T13:12:20Z dc.date.available: 2015-07-08T13:12:20Z dc.date.digitalpublicationdate: 2005-08-27 dc.identifier.barcode: 5990010101759 dc.identifier.origpath: /rawdataupload/upload/0101/761 dc.identifier.copyno: 1 dc.identifier.uri: http://www.new.dli.ernet.in/handle/2015/199379 dc.description.scannerno: 14 dc.description.scanningcentre: IIIT, Allahabad dc.description.main: 1 dc.description.tagged: 0 dc.description.totalpages: 98 dc.format.mimetype: application/pdf dc.language.iso: English dc.publisher: Indian Institute Of Technology Kanpur dc.rights: Out_of_copyright dc.source.library: Indian Institute Of Technology Kanpur dc.subject.classification: Technology dc.subject.classification: Electrical Engineering dc.subject.keywords: Algorithms dc.subject.keywords: Self Growing dc.title: Algorithms For Self Growing And Higher Order Neural Networks dc.type: Print-Paper dc.type: Book

Emerging developments in the speed, size, and power requirements of processors, coupled with networking advances, enable new applications for networks of sensors, cameras, and robots. However, we live in a world filled with uncertainty and noise, which affects the sensors we use, the environments we model, and the objects we observe. In this dissertation, we define Structured Tracking, where we apply novel machine learning and inference techniques to leverage environmental and tracked object structure. This approach improves accuracy and robustness while reducing computation. We focus on three application areas of societal benefit: safety, security, and privacy. We apply Belief Propagation (BP)[148] algorithms to sensor networks, and describe our modular framework for the more general Reweighted-BP formulation[203].

We address four problems related to multi-agent optimization, filtering and agreement. First, we investigate collaborative optimization of an objective function expressed as a sum of local convex functions, when the agents make decisions in a distributed manner using local information, while the communication topology used to exchange messages and information is modeled by a graph-valued random process, assumed independent and identically distributed. Specifically, we study the performance of the consensusbased multi-agent distributed subgradient method and show how it depends on the probability distribution of the random graph. For the case of a constant stepsize, we first give an upper bound on the difference between the objective function, evaluated at the agents--estimates of the optimal decision vector, and the optimal value. In addition, for a particular class of convex functions, we give an upper bound on the distances between the agents-- estimates of the optimal decision vector and the minimizer and we provide the rate of convergence to zero of the time varying component of the aforementioned upper bound. The addressed metrics are evaluated via their expected values. As an application we show how the distributed optimization algorithm can be used to perform collaborative system identification and provide numerical experiments under the randomized and broadcast gossip protocols. Second, we generalize the asymptotic consensus problem to convex metric spaces. Under minimal connectivity assumptions, we show that if at each iteration an agent updates its state by choosing a point from a particular subset of the generalized convex hull generated by the agents current state and the states of its neighbors, then agreement is achieved asymptotically. In addition, we give bounds on the distance between the consensus point(s) and the initial values of the agents.

We consider flow control algorithms consisting of two parts: quasi- static flow control and dynamic flow control. The quasi-static part uses short term average information on network utilization to allocate maximum data rates and to determine routes for each user. The rates are allocated to achieve an optimal trade-off between assigned priority cost functions for each user and the cost of congestion in the network. This optimization can be done by a distributed algorithm and is essentially no more complicated than optimizing routing alone. The dynamic flow control has the function of admitting or rejecting individual units of traffic into the network so as to enforce the maximum allocated rates and to prevent congestion by smoothing out the fluctuations in buffer occupancy. (Author)

Three kinds of numerical optimization algorithms have been investigated for use in estimating the electron momentum transfer and excitation cross sections for atoms and molecules based on measured electron transport, or swarm data. The methods investigated are the downhill or creeping simplex; simulated annealing; and neural networks. These methods have been used to obtain the cross section for momentum transfer for a model system from E/N (Electric field, E, divided by the total gas density, N) dependent drift velocities and characteristic energies. In addition the creeping simplex has been used to obtain momentum transfer cross sections for the He and Ar and the momentum transfer cross section and a vibrational excitation cross section for methane from measured drift velocity and characteristic energy data. A neural network has been used to obtain an estimate of the momentum transfer cross section of xenon in the vicinity of the Ramsauer minimum from swarm data. These results serve as examples of what may be possible using these and, perhaps other optimization algorithms. (jhd)

We propose algorithms for nonparametric sample-based spacial change detection and estimation in large scale sensor networks. We collect random samples containing the location of sensors and their local decisions, and assume that the local decisions can be stimulated or normal , reflecting the local strength of some stimulating agent. Then change in the location of the agent manifests itself by a change in the distribution of stimulated sensors. In this paper, we are aiming at developing a test that, given two collections of samples, can decide whether the distribution generating the samples has changed or not, and give an estimated changed area if a change is indeed detected. The focus of this paper is to reduce the complexity of the detection and estimation algorithm. We propose two fast algorithms with almost linear complexity and analyze their completeness, flexibility and robustness.

In this paper, we introduce algorithms for constructing broadcasting and multicasting trees. These algorithms are general because they may be used for tree cost functions that are of arbitrary form. Thus, essentially the same algorithmic procedures are used for different tree cost functions. We evaluate the effectiveness of the general algorithms by applying them to different cost functions that are often used to model wired and wireless networks. Besides providing a unifying framework for dealing with many present and future tree-construction applications, these algorithms typically outperform some existing algorithms that are specifically designed for energy-aware wireless networks. These general algorithms perform well at the expense of higher computational complexity. They are centralized algorithms, requiring the full network information for tree construction. Thus, we also present variations of these general algorithms to yield other algorithms that have lower complexity and distributed implementation.

This work presents joint iterative power allocation and interference suppression algorithms for DS-CDMA networks which employ multiple relays and the amplify and forward cooperation strategy. We propose a joint constrained optimization framework that considers the allocation of power levels across the relays subject to individual and global power constraints and the design of linear receivers for interference suppression. We derive constrained minimum mean-squared error (MMSE) expressions for the parameter vectors that determine the optimal power levels across the relays and the parameters of the linear receivers. In order to solve the proposed optimization problems efficiently, we develop recursive least squares (RLS) algorithms for adaptive joint iterative power allocation, and receiver and channel parameter estimation. Simulation results show that the proposed algorithms obtain significant gains in performance and capacity over existing schemes.

Emerging Information-Centric Networking (ICN) architectures seek to optimally utilize both bandwidth and storage for efficient content distribution over the network. The Virtual Interest Packet (VIP) framework has been proposed to enable joint design of forwarding and caching within the Named Data Networking (NDN) architecture. The virtual plane of the VIP framework captures the measured demand for content objects, but does not reflect interest collapse and suppression in the NDN network. We aim to further improve the performance of the existing VIP algorithms by using a modified virtual plane where VIP counts are appropriately scaled to reflect interest suppression effects. We characterize the stability region of the modified virtual plane with VIP scaling, develop a new distributed forwarding and caching algorithm operating on the scaled VIPs, and demonstrate the throughput optimality of the scaled VIP algorithm in the virtual plane. Numerical experiments demonstrate significantly enhanced performance relative to the existing VIP algorithm, as well as a number of other baseline algorithms.

Our investigation consists of three inter-related thrusts: traffic modeling and analysis, network control, and information assurance in wireline and wireless networks. We model multi-scale behavior in network systems, where traffic and system behavior can be highly correlated over multiple time scales (e.g., LRD). We investigate the causes of LRD traffic in network systems, which may result from traffic correlation, protocol behavior (e.g., retransmissions), and network congestion; and statistically analyzed the properties of LRD traffic from empirical data sets. We develop a unifying theory for network control that exploits the interactions across network functionalities, operates at appropriate time-scales, and is effective in the presence of LRD. We formulate optimization and distributed control problems for providing network services, and study the impact of LRD traffic on network control, performance, and security. We also develop an integrative approach that combines the LRD modeling and network control to obtain non-parametric or semi-parametric techniques for the distributed detection of information flow and flow changes needed for preventing security attacks. We characterize flow detectability as a function of flow rate, delay and memory constraints, and developed distributed detection schemes that guarantee vanishingly low detection error probabilities.

[CITATION] Solving combinatorial optimization problems by simulated annealing, genetic algorithms, and neural networks Y Lu - 1991 - University of Minnesota Cited by 4

The challenges of data processing, transmission scheduling and routing within a space network present a multi-criteria optimization problem. Long delays, intermittent connectivity, asymmetric data rates and potentially high error rates make traditional networking approaches unsuitable. The delay tolerant networking architecture and protocols attempt to mitigate many of these issues, yet transmission scheduling is largely manually configured and routes are determined by a static contact routing graph. A high level of variability exists among the requirements and environmental characteristics of different missions, some of which may allow for the use of more opportunistic routing methods. In all cases, resource allocation and constraints must be balanced with the optimization of data throughput and quality of service. Much work has been done researching routing techniques for terrestrial-based challenged networks in an attempt to optimize contact opportunities and resource usage. This paper examines several popular methods to determine their potential applicability to space networks.

The long term goals of the project are: (1) To establish systems and algorithms for controlled Lagrangian particle tracking that will be used to improve the accuracy of model based prediction of trajectories of controlled underwater vehicles subjected to ocean current. (2) To achieve a mission planning system for robotic underwater sensor networks that are able to perform automatic or semiautomatic adaptation to extreme ocean conditions and platform failure, deployment, and recovery. OBJECTIVES We develop a set of automation middleware that implement a set of novel algorithms for robotic underwater sensor networks serving applications of ocean sampling and ocean model improvement. We design novel model adjustment, cooperative control, and distributed sensing algorithms that will be implemented through the automation middleware.

Three kinds of numerical optimization algorithms have been investigated for use in estimating the electron momentum transfer and excitation cross sections for atoms and molecules based on measured electron transport, or swarm data. The methods investigated are the downhill or creeping simplex; simulated annealing; and neural networks. These methods have been used to obtain the cross section for momentum transfer for a model system from E/N (Electric field, E, divided by the total gas density, N) dependent drift velocities and characteristic energies. In addition the creeping simplex has been used to obtain momentum transfer cross sections for He and Ar and the momentum transfer cross section and a vibrational excitation cross section for methane from measured drift velocity and characteristic energy data. A neural network has been used to obtain an estimate of the momentum transfer cross section of xenon in the vicinity of the Ramsauer minimum from swarm data. These results serve as examples of what may be possible using these and, perhaps other optimization algorithms. Keywords: Draft velocity, Cross section, Numerical optimization, Simplex method.

The design of algorithm plays an important role in operations research in general and dynamic programming in particular. In most dynamic programming algorithms, formalism of computing, data structure and complexity analysis does not appear. One of the objectives of this research is to provide such a formalism. Converging branch, diverging branch, feed-forward loop and feedback loop systems are considered. In each case, first, a high level algorithm followed by the detailed computer algorithm is described. Formulas for storage and computational complexities for each computer algorithm are derived. Finally algorithms are implemented on VAX-11/780 computers using UCSD PASCAL. (Author)

Energy  besides  Other  factors  production  is  considered  the  main  factor  in  the growth  and  economic  development  and  in  the  performance  of  different  sectors economic  can  play  beneficial  roles.  Hence,  the  country  authorities  should  try  to predict  anything  more  precise  energy  consumption  in  the  proper  planning  and guidance  consumption,  to  control  the  way  they  desired  energy  demand  and  supply parameters. The purpose of this paper  is Evaluation Hybrid model of artificial neural networks  and  genetic  algorithms  in  the  forecast  demand  energy.  for  Prediction  energy  consumption  in  the  country.  Case  study  is  energy consumption in transportation sector of Iran.  So for this review, were used the annual data energy consumption of transport as a variable output of forecast models and data from the  entire  country's annual population, GDP and the  number of  vehicle as the input variables. Evaluation results showed that   the hybrid model of neural networks and  genetic  algorithm  (ANN-GA),  compared  to  other models  with  the  highest accuracy in predicting energy demand in the transportation sector.

Energy  besides  Other  factors  production  is  considered  the  main  factor  in  the growth  and  economic  development  and  in  the  performance  of  different  sectors economic  can  play  beneficial  roles.  Hence,  the  country  authorities  should  try  to predict  anything  more  precise  energy  consumption  in  the  proper  planning  and guidance  consumption,  to  control  the  way  they  desired  energy  demand  and  supply parameters. The purpose of this paper  is Evaluation Hybrid model of artificial neural networks  and  genetic  algorithms  in  the  forecast  demand  energy.  for  Prediction  energy  consumption  in  the  country.  Case  study  is  energy consumption in transportation sector of Iran.  So for this review, were used the annual data energy consumption of transport as a variable output of forecast models and data from the  entire  country's annual population, GDP and the  number of  vehicle as the input variables. Evaluation results showed that   the hybrid model of neural networks and  genetic  algorithm  (ANN-GA),  compared  to  other models  with  the  highest accuracy in predicting energy demand in the transportation sector.

This paper develops a mesh network protection scheme that guarantees a quantifiable minimum grade of service upon a failure within a network. The scheme guarantees that a fraction q of each demand remains after any single link failure. A linear program is developed to find the minimum cost capacity allocation to meet both demand and protection requirements. For q - 1/2, an exact algorithmic solution for the optimal routing and allocation is developed using multiple shortest paths. For q 1/2, a heuristic algorithm based on disjoint path routing is developed that performs, on average, within 1.4% of optimal, and runs four orders of magnitude faster than the minimum-cost solution achieved via the linear program. Moreover, the partial protection strategies developed Achieve reductions of up to 82% over traditional full protection schemes.

The goal of this project was to develop efficient synchronization schemes to ensure robust operation of wireless airborne networks in the absence of GPS (Global Positioning Systems), and in the presence of arbitrary network delay distributions. To cope with the Gaussian or non-Gaussian nature of the random network delays, a novel method, referred to as the Gaussian Mixture Kalman Particle Filter (GMKPF), is proposed to estimate the clock offset and shown to be robust to arbitrary network delays. GMKPF represents a better and more flexible alternative to the Gaussian Maximum Likelihood (GML), and Exponential Maximum Likelihood (EML) estimators for clock offset estimation in non-Gaussian or non-exponential random delay models. The computer simulations illustrate that GMKPF yields much more accurate results relative to GML and EML when the network delays are modeled in terms of a single non-Gaussian/non-exponential distribution or as a mixture of several distributions. As deliverables, the set of Matlab programs used to implement GMKPF and validate its performance are uploaded separately into Jiffy.

Several deterministic and stochastic multi-variable global optimization algorithms (Conjugate Gradient, Nelder-Mead, Quasi-Newton, and Global) are investigated in conjunction with energy minimization principle to resolve the pressure and volumetric flow rate fields in single ducts and networks of interconnected ducts. The algorithms are tested with seven types of fluid: Newtonian, power law, Bingham, Herschel-Bulkley, Ellis, Ree-Eyring and Casson. The results obtained from all those algorithms for all these types of fluid agree very well with the analytically derived solutions as obtained from the traditional methods which are based on the conservation principles and fluid constitutive relations. The results confirm and generalize the findings of our previous investigations that the energy minimization principle is at the heart of the flow dynamics systems. The investigation also enriches the methods of Computational Fluid Dynamics for solving the flow fields in tubes and networks for various types of Newtonian and non-Newtonian fluids.

Several deterministic and stochastic multi-variable global optimization algorithms (Conjugate Gradient, Nelder-Mead, Quasi-Newton, and Global) are investigated in conjunction with energy minimization principle to resolve the pressure and volumetric flow rate fields in single ducts and networks of interconnected ducts. The algorithms are tested with seven types of fluid: Newtonian, power law, Bingham, Herschel-Bulkley, Ellis, Ree-Eyring and Casson. The results obtained from all those algorithms for all these types of fluid agree very well with the analytically derived solutions as obtained from the traditional methods which are based on the conservation principles and fluid constitutive relations. The results confirm and generalize the findings of our previous investigations that the energy minimization principle is at the heart of the flow dynamics systems. The investigation also enriches the methods of Computational Fluid Dynamics for solving the flow fields in tubes and networks for various types of Newtonian and non-Newtonian fluids.

38 p. 28 cm

38 p. 28 cm

With the advent of the Web, large scale social andinformation networks containing detailed traces of human activity havebecome available. This offers great opportunities to measure, model andpredict actions of millions of people. For example, we had anopportunity to analyze a “planetary scale” Microsoft InstantMessenger network that contains 240 million people, with more than 1billion conversations per day (4.5TB of data), which makes it thelargest social network analyzed to date. In this talk I will focus on two aspects of the dynamics of large real-world networks: (a) dynamics of information diffusion and cascadingbehavior in networks, and (b) dynamics of time evolving networks.First, I will consider network cascades that are created by thediffusion process where behavior spreads from node to node like anepidemic. We study two related scenarios: information diffusion amongblogs, and a viral marketing setting of 16 million productrecommendations among four million people. Motivated by our empiricalobservations we develop algorithms for finding influential bloggers anddetecting disease outbreaks. We exploit the ”submodularity” principleto develop an efficient algorithm that achieves near optimal solutions,while scaling to large problems and being 700 times faster than asimple greedy algorithm. Second, in our recent work we foundinteresting and counter intuitive patterns, which change some of thebasic assumptions about fundamental structural properties of networksvarying over time. Leveraging our observations we developed a Kroneckergraph generator model that explains processes governing networkevolution. Moreover, we can fit the model to large networks, and thenuse it to generate realistic graphs and give formal statements abouttheir properties. Estimating the model naively takes O(N!N 2 ) while wedevelop a linear time O(E) algorithm. ©2008 Microsoft Corporation. All rights reserved.

With the advent of the Web, large scale social andinformation networks containing detailed traces of human activity havebecome available. This offers great opportunities to measure, model andpredict actions of millions of people. For example, we had anopportunity to analyze a “planetary scale” Microsoft InstantMessenger network that contains 240 million people, with more than 1billion conversations per day (4.5TB of data), which makes it thelargest social network analyzed to date. In this talk I will focus on two aspects of the dynamics of large real-world networks: (a) dynamics of information diffusion and cascadingbehavior in networks, and (b) dynamics of time evolving networks.First, I will consider network cascades that are created by thediffusion process where behavior spreads from node to node like anepidemic. We study two related scenarios: information diffusion amongblogs, and a viral marketing setting of 16 million productrecommendations among four million people. Motivated by our empiricalobservations we develop algorithms for finding influential bloggers anddetecting disease outbreaks. We exploit the ”submodularity” principleto develop an efficient algorithm that achieves near optimal solutions,while scaling to large problems and being 700 times faster than asimple greedy algorithm. Second, in our recent work we foundinteresting and counter intuitive patterns, which change some of thebasic assumptions about fundamental structural properties of networksvarying over time. Leveraging our observations we developed a Kroneckergraph generator model that explains processes governing networkevolution. Moreover, we can fit the model to large networks, and thenuse it to generate realistic graphs and give formal statements abouttheir properties. Estimating the model naively takes O(N!N 2 ) while wedevelop a linear time O(E) algorithm. ©2008 Microsoft Corporation. All rights reserved.

We consider the joint design of transmit beamforming and receive signal-splitting ratios in the downlink of a wireless network with simultaneous radio-frequency (RF) information and energy transfer. Under constraints on the signal-to-interference-plus-noise ratio (SINR) at each user and the total transmit power at the base station, the design objective is to maximize either the sum harvested energy or the minimum harvested energy. We develop a computationally efficient path-following method to solve these challenging nonconvex optimization problems. We mathematically show that the proposed algorithms iteratively progress and converge to locally optimal solutions. Simulation results further show that these locally optimal solutions are the same as the globally optimal solutions for the considered practical network settings.

We develop distributed algorithms to allocate resources in multi-hop wireless networks with the aim of minimizing total cost. In order to observe the fundamental duplexing constraint that co-located transmitters and receivers cannot operate simultaneously on the same frequency band, we first devise a spectrum allocation scheme that divides the whole spectrum into multiple sub-bands and activates conflict-free links on each sub-band. We show that the minimum number of required sub-bands grows asymptotically at a logarithmic rate with the chromatic number of network connectivity graph. A simple distributed and asynchronous algorithm is developed to feasibly activate links on the available sub-bands. Given a feasible spectrum allocation, we then design node-based distributed algorithms for optimally controlling the transmission powers on active links for each sub-band, jointly with traffic routes and user input rates in response to channel states and traffic demands. We show that under specified conditions, the algorithms asymptotically converge to the optimal operating point.

Wireless sensor networks monitor dynamic environments that change rapidly over time. This dynamic behavior is either caused by external factors or initiated by the system designers themselves. To adapt to such conditions, sensor networks often adopt machine learning techniques to eliminate the need for unnecessary redesign. Machine learning also inspires many practical solutions that maximize resource utilization and prolong the lifespan of the network. In this paper, we present an extensive literature review over the period 2002-2013 of machine learning methods that were used to address common issues in wireless sensor networks (WSNs). The advantages and disadvantages of each proposed algorithm are evaluated against the corresponding problem. We also provide a comparative guide to aid WSN designers in developing suitable machine learning solutions for their specific application challenges.

The implementation of network-centric warfare technologies is an abiding, critical interest of Air Force Science and Technology efforts for the Warfighter. Wireless communications, strategic signaling are areas of critical Air Force Mission need. Autonomous networks of multiple, heterogeneous Throughput enhancement and robust connectivity in communications and sensor networks are critical factors in net-centric USAF operations. This research directly supports the Air Force vision of information dominance and the development of anywhere, anytime operational readiness.

This research examines energy and error tradeoffs in Anchor-Free Range-Aware Wireless Sensor Network (WSN) Localization algorithms. A concurrent and an incremental algorithm (Anchor Free Localization (AFL) and Map Growing) are examined under varying network sizes, densities, deployments, and range errors. Despite current expectations, even the most expensive configurations do not expend significant battery life (at most 0.4%), implying little energy can be conserved during localization. Due to refinement, AFL is twice as accurate, using 6 times the communication. For both, node degree affects communication most. As degree increases, Map Growing communication increases, while AFL transmissions drop. Nodes with more neighbors refine quicker with fewer messages. At high degree, many nodes receive the same message, overpowering the previous effect, and raising AFL received bits. Built from simulation data, the Energy Consumption Model predicts energy usage of incremental and concurrent algorithms used in networks with varying size, density, and deployments. It is applied to current wireless sensor nodes. Military WSNs should be flexible, cheap, and long lasting. Anchor-Free, Range-Aware algorithms best fit this need.

Thesis advisor, Kevin Wood

1 P.Ezhumalai 2 Dr.A.Chilambuchelvann 1 , Dept of Computer Science Engineering, Ralalakshmi Engineering College, Thandalam-602 105, Chennai, India 2. Professor, Dept Of Computer Science Engineering, R.M.K Engineering College,Chennai. India.

1 P.Ezhumalai 2 Dr.A.Chilambuchelvann 1 , Dept of Computer Science Engineering, Ralalakshmi Engineering College, Thandalam-602 105, Chennai, India 2. Professor, Dept Of Computer Science Engineering, R.M.K Engineering College,Chennai. India.

Mobile Ad Hoc networks (MANET), distinct from traditional distributed systems, are dynamic and selforganizing networks. MANET requires a leader to coordinate and organize tasks. The challenge is to have the right election algorithm that chooses the right leader based on various factors in MANET. In this paper, we analyze four leader election algorithms used in mobile Ad Hoc Networks. Factors considered in our analysis are time complexity, message complexity, assumptions considered, fault tolerance and timing model. Our proposed enhancements include recovered nodes inquiring about the current leader and the use of candidates during election to reduce the overhead of starting a new election session. In addition, better election criteria specific to MANET, such as battery life and signal strength, are proposed. Our evaluation and discussion shows that the proposed enhancements are effective. The analysis can be used as a reference for system designers in choosing the right election algorithm for MANET.

Mobile Ad Hoc networks (MANET), distinct from traditional distributed systems, are dynamic and selforganizing networks. MANET requires a leader to coordinate and organize tasks. The challenge is to have the right election algorithm that chooses the right leader based on various factors in MANET. In this paper, we analyze four leader election algorithms used in mobile Ad Hoc Networks. Factors considered in our analysis are time complexity, message complexity, assumptions considered, fault tolerance and timing model. Our proposed enhancements include recovered nodes inquiring about the current leader and the use of candidates during election to reduce the overhead of starting a new election session. In addition, better election criteria specific to MANET, such as battery life and signal strength, are proposed. Our evaluation and discussion shows that the proposed enhancements are effective. The analysis can be used as a reference for system designers in choosing the right election algorithm for MANET.

The proliferation of different wireless access technologies, together with the growing number of multi-radio wireless devices suggest that the opportunistic utilization of multiple connections at the users can be an effective solution to the phenomenal growth of traffic demand in wireless networks. In this paper we consider the downlink of a wireless network with $N$ Access Points (AP's) and $M$ clients, where each client is connected to several out-of-band AP's, and requests delay-sensitive traffic (e.g., real-time video). We adopt the framework of Hou, Borkar, and Kumar, and study the maximum total timely throughput of the network, denoted by $C_{T^3}$, which is the maximum average number of packets delivered successfully before their deadline. Solving this problem is challenging since even the number of different ways of assigning packets to the AP's is $N^M$. We overcome the challenge by proposing a deterministic relaxation of the problem, which converts the problem to a network with deterministic delays in each link. We show that the additive gap between the capacity of the relaxed problem, denoted by $C_{det}$, and $C_{T^3}$ is bounded by $2\sqrt{N(C_{det}+N/4)}$, which is asymptotically negligible compared to $C_{det}$, when the network is operating at high-throughput regime. In addition, our numerical results show that the actual gap between $C_{T^3}$ and $C_{det}$ is in most cases much less than the worst-case gap proven analytically. Moreover, using LP rounding methods we prove that the relaxed problem can be approximated within additive gap of $N$. We extend the analytical results to the case of time-varying channel states, real-time traffic, prioritized traffic, and optimal online policies. Finally, we generalize the model for deterministic relaxation to consider fading, rate adaptation, and multiple simultaneous transmissions.