Theory and Algorithms Commons^™

Open-Loop State-Space Model Identification From Closed-Loop Data, Lori Guy

Mechanical & Aerospace Engineering Theses & Dissertations

This thesis provides an investigation of a system identification algorithm which identifies an open-loop state-space model from a linear system that is operating under closed-loop conditions. In order to investigate the system identification algorithm some basic ideas of system identification theory are reviewed. Examples using simulated data are presented to characterize the effects of varying the parameters for open-loop and closed-loop system identification processes. Both noise· free and noise contaminated cases are simulated. Linear Quadratic Gaussian (LQG) control theory is reviewed and the motivation for using iterative LQG control feedback is discussed. The derivation of the proposed system identification algorithm …

Approximation Algorithms: Good Solutions To Hard Problems, Ran Libeskind-Hadas

All HMC Faculty Publications and Research

Consider a computer network represented by an undirected graph where the vertices represent computer nodes and the edges represent links between the nodes. Since some of the links in the network may become faulty, link testing devices are placed at some of the nodes. A tester at a particular node can test all links incident to that node. Since the testers are expensive, however, we wish to deploy the minimum number of these devices such that every link is incidient to at least one node containing a tester. In graph theoretic terms, a vertex cover is a subset of the …

A Linear-Time Recognition Algorithm For P4-Reducible Graphs, B. Jamison, S. Olariu

Computer Science Faculty Publications

The P₄-reducible graphs are a natural generalization of the well-known class of cographs, with applications to scheduling, computational semantics, and clustering. More precisely, the P₄-reducible graphs are exactly the graphs none of whose vertices belong to more than one chordless path with three edges. A remarkable property of P₄-reducible graphs is their unique tree representation up to isomorphism. In this paper we present a linear-time algorithm to recognize P₄-reducible graphs and to construct their corresponding tree representation.

Rule Extraction: From Neural Architecture To Symbolic Representation, Gail A. Carpenter, Ah-Hwee Tan

Research Collection School Of Computing and Information Systems

This paper shows how knowledge, in the form of fuzzy rules, can be derived from a supervised learning neural network called fuzzy ARTMAP. Rule extraction proceeds in two stages: pruning, which simplifies the network structure by removing excessive recognition categories and weights; and quantization of continuous learned weights, which allows the final system state to be translated into a usable set of descriptive rules. Three benchmark studies illustrate the rule extraction methods: (1) Pima Indian diabetes diagnosis, (2) mushroom classification and (3) DNA promoter recognition. Fuzzy ARTMAP and ART-EMAP are compared with the ADAP algorithm, the k nearest neighbor system, …

Finding Connected Components On A Scan Line Array Processor, Ronald I. Greenberg

Computer Science: Faculty Publications and Other Works

This paper provides a new approach to labeling the connected components of an n x n image on a scan line array processor (comprised of n processing elements). Variations of this approach yield an algorithm guaranteed to complete in o(n lg n) time as well as algorithms likely to approach O(n) time for all or most images. The best previous solutions require using a more complicated architecture or require Omega(n lg n) time. We also show that on a restricted version of the architecture, any algorithm requires Omega(n lg n) time in the worst case.

An Analysis Of Bayesian Networks As Classifiers, Gregory C. Ahlquist

Theses and Dissertations

An analysis of Bayesian networks as classifiers is presented. This analysis results in an algorithm and several tools related to Bayesian network classifiers. The tools calculate and display the decision regions for two level Bayesian network classifiers. They collectively provide an approach to analyze the effects of changing network parameters on the network's decision regions. The algorithm defines a Bayesian network classifier to solve traditional classification problems. The algorithm is data driven, meaning that the resulting Bayesian network classifier is uniquely tuned to the classification problem at hand. Also, the algorithm contains procedures for defining the topology of a Bayesian …

The Fat-Pyramid And Universal Parallel Computation Independent Of Wire Delay, Ronald I. Greenberg

Computer Science: Faculty Publications and Other Works

This paper shows that a fat-pyramid of area Θ(A) requires only O(log A) slowdown to simulate any competing network of area A under very general conditions. The result holds regardless of the processor size (amount of attached memory) and number of processors in the competing networks as long as the limitation on total area is met. Furthermore, the result is valid regardless of the relationship between wire length and wire delay. We especially focus on elimination of the common simplifying assumption that unit time suffices to traverse a wire regardless of its length, since the assumption becomes more and more …

Improved Algorithms For Bipartite Network Flow, Ravindra K. Ahuja, James B. B. Orlin, Clifford Stein, Robert E. Tarjan

Dartmouth Scholarship

In this paper, network flow algorithms for bipartite networks are studied. A network G = (V,E) is called bipartite if its vertex set V can be partitioned into two subsets V_1 and V_2 such that all edges have one endpoint in V_1 and the other in $V_2 $. Let $n = |V|, n_1 = |V_1 | , n_2 = |V_2 |, m = |E| and assume without loss of generality that n_1 \leqslant n_2. A bipartite network is called unbalanced if n_1 \ll n_2 $ and balanced otherwise. (This notion is necessarily imprecise.) It is shown that several maximum flow …

Defining Locally Shared Memory Constructs For Special Purpose Parallel Architectures, David J. Nielsen

Electrical & Computer Engineering Theses & Dissertations

Locally shared memory systems offer significant advantages over other parallel processing systems for specific classes of problems. Locally shared memory systems tend to be easier to program because explicit passing of messages is not necessary. The goal in defining a locally shared memory system is to allow only a small number of processors access to any single memory. If this goal is met, locally shared memory systems provide an architecture which is relatively simple to implement.

To make shared memory architectures attractive to designers of special purpose parallel architectures, the architectures must be scalable. To be scalable, the number of …

Genetic Algorithms And Artificial Life, Melanie Mitchell, Stephanie Forrest

Computer Science Faculty Publications and Presentations

Genetic algorithms are computational models of evolution that play a central role in many artificial-life models. We review the history and current scope of research on genetic algorithms in artificial life, giving illustrative examples in which the genetic algorithm is used to study how learning and evolution interact, and to model ecosystems, immune system, cognitive systems, and social systems. We also outline a number of open questions and future directions for genetic algorithms in artificial-life research

Polylog Depth Circuits For Integer Factoring And Discrete Logarithms, Jonathan P. Sorenson

Scholarship and Professional Work - LAS

AbstractIn this paper, we develop parallel algorithms for integer factoring and for computing discrete logarithms. In particular, we give polylog depth probabilistic boolean circuits of subexponential size for both of these problems, thereby solving an open problem of Adleman and Kompella.

Existing sequential algorithms for integer factoring and discrete logarithms use a prime base which is the set of all primes up to a bound B. We use a much smaller value for B for our parallel algorithms than is typical for sequential algorithms. In particular, for inputs of length n, by setting B = nlogdn with d a positive …

Dynamic Task Scheduling For The Atamm Multicomputer Operating System Using Embedded Firmware On Microcontrollers, Sudhir Sastry

Electrical & Computer Engineering Theses & Dissertations

A dynamic task scheduling strategy for the distributed processing of large grain dataflow algorithms using embedded firmware on an ATAMM testbed consisting of interconnected microcontrollers is presented in this thesis. The ODU/NASA developed Algorithm to Architecture Mapping Model, ATAMM, uses marked graph models to specify data and control flow for the execution of iterative, deterministic large grain dataflow algorithms in a multicomputing environment. The testbed consists of a bank of four 68HC11 microcontrollers that communicate over a token bus. The token bus arbitration scheme used is contention free and well suited for real-time computing applications. The execution of data flow …

Adaptive Execution Of Data Parallel Computations On Networks Of Heterogeneous Workstations, Robert Prouty, Steve Otto, Jonathan Walpole

Computer Science Faculty Publications and Presentations

Parallel environments consisting of a network of heterogeneous workstations introduce an inherently dynamic environment that differs from multicomputers. Workstations are usually considered “shared” resources while multicomputers provide dedicated processing power. The number of workstations available for use is continually changing; the parallel machine presented by the network is in effect continually reconfiguring itself. Application programs must effectively adapt to the changing number of processing nodes while maintaining computational efficiency. This paper examines methods for adapting to this dynamic environment within the framework of explicit message passing under the data parallel programming model. We present four requirements which we feel a …

Effective Parallel Algorithm Animation, Paul W. Chase

Theses and Dissertations

The AFIT Algorithm Animation Research Facility AAARF was developed by the Air Force Institute of Technology AFIT as a teaching aid for data structures and algorithm design. In particular, an extensive set of performance animations has been developed for the Intel iPSC Hypercube parallel processing system. This research focuses in part on developing animation support for discrete event simulation, mission routing, and evolutionary algorithms based on abstract representations of parallel algorithm behavior. The effort also builds extensions to the AAARF system and examines direction for further research. An innovative adaptable application-specific animation construction environment has been designed and implemented. The …

A Topology-Aware Collision Resolution Algorithm, Lewis Barnett Iii

Department of Math & Statistics Technical Report Series

A new collision resolution algorithm called the Space Division Multiple Access protocol (SDMA) is presented. SDMA gains a performance advantage over similar protocols by using information about the positions of stations on the network. The protocol can operate asynchrononsly on a broadcast bus, allowing variable sized packet traffic. Through simulation the protocol is demonstrated to have better performance than Ethernet and the Capetanakis Tree protocol, a similar collision resolution protocol, under some traffic conditions. In particular, under heavy loads, SDMA displays better average throughput and lower variance of delay than Ethernet. The protocol demonstrates a performance bias based on the …

A Greedy Hypercube-Labeling Algorithm, D. Bhagavathi, C. E. Grosch, S. Olariu

Computer Science Faculty Publications

Due to its attractive topological properties, the hypercube multiprocessor has emerged as one of the architectures of choice when it comes to implementing a large number of computational problems. In many such applications, Gray-code labelings of the hypercube are a crucial prerequisite for obtaining efficient algorithms. We propose a greedy algorithm that, given an n-dimensional hypercube H with N=2² nodes, returns a Gray-code labeling of H, that is, a labeling of the nodes with binary strings of length n such that two nodes are neighbors in the hypercube if, and only if, their labels differ in exactly …

Universal Wormhole Routing, Ronald I. Greenberg, Hyeong-Cheol Oh

Computer Science: Faculty Publications and Other Works

We examine the wormhole routing problem in terms of the "congestion" c and "dilation" d for a set of packet paths. We show, with mild restrictions, that there is a simple randomized algorithm for routing any set of P packets in O(cdη + cLηlog P) time, where L is the number of flits in a packet, and η = min {d,L]; only a constant number of flits are stored in each queue at any time. Using this result, we show that a fat-tree network of area Θ(A) can simulate wormhole routing on any network of comparable area with O(log 3 …

Parallel Algorithms For Single-Layer Channel Routing, Ronald I. Greenberg, Shih-Chuan Hung, Jau-Der Shih

Computer Science: Faculty Publications and Other Works

We provide efficient parallel algorithms for the minimum separation, offset range, and optimal offset problems for single-layer channel routing. We consider all the variations of these problems that have linear-time sequential solutions rather than limiting attention to the ``river-routing'' context, where single-sided connections are disallowed. For the minimum separation problem, we obtain O(lgN) time on a CREW PRAM or O(lgN/lglgN) time on a CRCW PRAM, both with optimal work (processor-time product) of O(N), where N is the number of terminals. For the offset range problem, we obtain the same time and processor bounds as long as only one side of …

Cyclo-Static Scheduling Of Large Grain Dataflow Algorithms On A Local Area Atamm Multicomputing Testbed, Sudeepto Roy

Electrical & Computer Engineering Theses & Dissertations

A strategy for cyclo-statically scheduling deterministic large grain dataflow (LGDF) algorithms for distributed execution on loosely coupled multicomputer architectures is presented in this research. The computational paradigm used is the ODU/NASA developed Algorithm To Architecture Mapping Model (ATAMM), which consists of marked graphs and Gantt chart representations that model the iterative execution of deterministic LGDF algorithms for different values of throughput and computation time. It is postulated that the behavior of these algorithms could be represented by the aggregate execution of an ensemble of cyclically shifted threads of a specific node sequence. Assuming the existence of one or more such …

Feasible Offset And Optimal Offset For Single-Layer Channel Routing, Ronald I. Greenberg, Jau-Der Shih

Computer Science: Faculty Publications and Other Works

The paper provides an efficient method to find all feasible offsets for a given separation in a VLSI channel routing problem in one layer. The prior literature considers this task only for problems with no single-sided nets. When single-sided nets are included, the worst-case solution time increases from Theta(n) to Omega(n^2), where n is the number of nets. But, if the number of columns c is O(n), one can solve the problem in time O(n^{1.5}lg n ), which improves upon a `naive' O(cn) approach. As a corollary of this result, the same time bound suffices to find the optimal offset …

Minimizing Channel Density With Movable Terminals, Ronald I. Greenberg, Jau-Der Shih

Computer Science: Faculty Publications and Other Works

We give algorithms to minimize density for channels with terminals that are movable subject to certain constraints. The main cases considered are channels with linear order constraints, channels with linear order constraints and separation constraints, channels with movable modules containing fixed terminals, and channels with movable modules and terminals. In each case, previous results for running time and space are improved by a factor of L/lg n and L , respectively, where L is the channel length and n is the number of terminals.

Matching Points To Lines: Sonar-Based Localization For The Psubot, Kevin Blythe Stanton

Dissertations and Theses

The PSUBOT (pronounced pea-es-you-bought) is an autonomous wheelchair robot for persons with certain disabilities. Its use of voice recognition and autonomous navigation enable it to carry out high level commands with little or no user assistance. We first describe the goals, constraints, and capabilities of the overall system including path planning and obstacle avoidance. We then focus on localization-the ability of the robot to locate itself in space. Odometry, a compass, and an algorithm which matches points to lines are each employed to accomplish this task. The matching algorithm (which matches "points" to "lines") is the main contribution to this …

A Fast Serial Algorithm For The Finite Temperature Quenched Potts Model, Gregory N. Hassold, Elizabeth A. Holm

Physics Publications

An efficient serial algorithm for finite temperature, quenched Potts model simulations of domain evolution has been developed. This ''n‐fold way'' algorithm eliminates unsuccessful spin flip attempts a priori by flipping sites with a frequency proportional to their site activity, defined as the sum of the probability of success for every possible spin flip at that site. Finite temperature efficiency for high‐spin degeneracy systems is achieved by utilizing a new, analytical expression for the portion of the site activity due to flips to non-neighbor spin values. Hence, to determine the activity of a site, only flips to the …

On The Difficulty Of Manhattan Channel Routing, Ronald I. Greenberg, Joseph Jaja, Sridhar Krishnamurthy

Computer Science: Faculty Publications and Other Works

We show that channel routing in the Manhattan model remains difficult even when all nets are single-sided. Given a set of n single-sided nets, we consider the problem of determining the minimum number of tracks required to obtain a dogleg-free routing. In addition to showing that the decision version of the problem isNP-complete, we show that there are problems requiring at least d+Omega(sqrt(n)) tracks, where d is the density. This existential lower bound does not follow from any of the known lower bounds in the literature.

Generalization And Parallelization Of Messy Genetic Algorithms And Communication In Parallel Genetic Algorithms, Laurence D. Merkle

Theses and Dissertations

Genetic algorithms (GA) are highly parallelizable, robust semi- optimization algorithms of polynomial complexity. The most commonly implemented GAs are 'simple' GAs (SGAs). Reproduction, crossover, and mutation operate on solution populations. Deceptive and GA-hard problems are provably difficult for simple GAs. Messy GAs (MGA) are designed to overcome these limitations. The MGA is generalized to solve permutation type optimization problems. Its performance is compared to another MGA's, an SGA's, and a permutation SGA's. Against a fully deceptive problem the generalized MGA (GMGA) consistently performs better than the simple GA. Against an NP-complete permutation problem, the GMGA performs better than the other …

Packet Routing In Networks With Long Wires, Ronald I. Greenberg, H.-C. Oh

Computer Science: Faculty Publications and Other Works

In this paper, we examine the packet routing problem for networks with wires of differing length. We consider this problem in a network independent context, in which routing time is expressed in terms of “congestion” and “dilation” measures for a set of packet paths. We give, for any constant ε > 0, a randomized on-line algorithm for routing any set of N packets in O((Clg^ε(Nd)+Dlg(Nd))/lglg(Nd)) time, where C is the maximum congestion and D is the length of the longest path, both taking wire delays into account, and d is the longest path in terms of number of wires. We also …

Autonomous Face Segmentation, Kevin P. Gay

Theses and Dissertations

The purpose of this study was to implement an autonomous face segmentor as the front end to a face recognition system on a Sun SPARCStation2. Face recognition performance criteria, specifically, the capabilities to isolate and resize faces in an image to a consistent scale, were analyzed to determine current practical limitations. Face images were acquired using a S-VHS camcorder. Segmentation was accomplished using motion detection and predefined rules. Tests were run to determine the suitable of the autonomous segmentor as the front-end to a face recognition system. The segmentation system developed consistently located faces and rescaled those faces to a …

Multiple Learner Systems Using Resampling Methods, Binyun Xie

Computer Science Theses & Dissertations

The N-Learners Problem deals with combining a number of learners such that the resultant system is "better", under some criterion, than the best of the individual learners. We consider a system of probably approximately correct concept learners. Depending on the available information, there are several methods to make the composite system better than the best of the individual learners. If a sample and an oracle that generates data points (but, not their classification) is available, then we show that we can achieve arbitrary levels of the normalized confidence of the composite system if (a) a robust learning algorithm is available, …

Formant Estimation From Dctc's Using A Feedforward Neural Network, Shubhangi U. Kelkar

Electrical & Computer Engineering Theses & Dissertations

Formants are the natural frequencies of the human vocal tract. Existing methods for estimating formants from speech signals are computationally complex and subject to errors for certain type of speech sounds. This thesis describes a method for estimating vowel formant frequencies from Discrete Cosine Transform Coefficients (DCTC's), a form of cepstral coefficients, using a feedforward neural network with back-propagation training. Experimental results are based on a large multispeaker data base. The results are obtained for both a linear transformation and a feedforward neural network with a nonlinear hidden layer. In general, the neural network transformation is superior to the linear …

Implementation And Analysis Of Np-Complete Algorithms On A Distributed Memory Computer, Joel S. Garmon

Theses and Dissertations

The purpose of this research is to explore methods used to parallelize NP-complete problems and the degree of improvement that can be realized using different methods of load balancing. A serial and four parallel A* branch and bound algorithms were implemented and executed on an Intel iPSC/2 hypercube computer. One parallel algorithm used a global, or centralized, list to store unfinished work and the other three parallel algorithms used a distributed list to store unfinished work locally on each processor. the three distributed list algorithms are: without load balancing, with load balancing, and with load balancing and work distribution. The …