Distributed simulation is an enabling concept to support the networked interaction of models and real world elements that are geographically distributed. This technology has brought a new set of challenging problems to solve, such as Data Distribution Management (DDM). The aim of DDM is to limit and control the volume of the data exchanged during a distributed simulation, and reduce the processing requirements of the simulation hosts by relaying events and state information only to those applications that require them. In this thesis, we propose a new DDM scheme, which we refer to as dynamic grid-based DDM. A lightweight UNT-RTI has been developed and implemented to investigate the performance of our DDM scheme. Our results clearly indicate that our scheme is scalable and it significantly reduces both the number of multicast groups used, and the message overhead, when compared to previous grid-based allocation schemes using large-scale and real-world scenarios.
This thesis presents algorithms to utilize the wireless bandwidth efficiently and at the same time meet the quality of service (QoS) requirements of the users. In the proposed algorithms we present an adaptive frame structure based upon the airlink frame loss probability and control the admission of call requests into the system based upon the load on the system and the QoS requirements of the incoming call requests. The performance of the proposed algorithms is studied by developing analytical formulations and simulation experiments. Finally we present an admission control algorithm which uses an adaptive delay computation algorithm to compute the queuing delay for each class of traffic and adapts the service rate and the reliability in the estimates based upon the deviation in the expected and obtained performance. We study the performance of the call admission control algorithm by simulation experiments. Simulation results for the adaptive frame structure algorithm show an improvement in the number of users in the system but there is a drop in the system throughput. In spite of the lower throughput the adaptive frame structure algorithm has fewer QoS delay violations. The adaptive call admission control algorithm adapts the call dropping probability of different classes of traffic and optimizes the system performance w.r.t the number of calls dropped and the reliability in meeting the QoS promised when the call is admitted into the system.
As bandwidth constraints on LAN/WAN environments decrease, the demand for distributed services will continue to increase. In particular, the proliferation of user-level applications requiring high-capacity distributed file storage systems will demand that such services be universally available. At the same time, the advent of high-speed networks have made the deployment of application and communication solutions based upon an Intelligent Mobile Agent (IMA) framework practical. Agents have proven to present an ideal development paradigm for the creation of autonomous large-scale distributed systems, and an agent-based communication scheme would facilitate the creation of independently administered distributed file services. This thesis thus outlines an architecture for such a distributed file system based upon an IMA communication framework.
The World Wide Web is one of the largest sources of information; more and more applications are being developed daily to make use of this information. This thesis presents a multi-agent architecture that deals with some of the issues related to Internet data extraction. The primary issue addresses the reliable, efficient and quick extraction of data through the use of HTTP performance monitoring agents. A second issue focuses on how to make use of available data to take decisions and alert the user when there is change in data; this is done with the help of user agents that are equipped with a Defeasible reasoning interpreter. An additional issue is the visualization of extracted data; this is done with the aid of VRML visualization agents. The cited issues are discussed using stock portfolio management as an example application.
This thesis is a proposed solution to the problem of including an effective interrupt mechanism in the set of concurrent- processing primitives of a block-structured programming language or system. The proposed solution is presented in the form of a programming language definition and model. The language is called TRIPLE.
Two applications of a binary tree data type based on a simple pairing function (a bijection between natural numbers and pairs of natural numbers) are explored. First, the tree is used to encode natural numbers, and algorithms that perform basic arithmetic computations are presented along with formal proofs of their correctness. Second, using this "canonical" representation as a base type, algorithms for encoding and decoding additional isomorphic data types of other mathematical constructs (sets, sequences, etc.) are also developed. An experimental application to a memory management system is constructed and explored using these isomorphic types. A practical analysis of this system's runtime complexity and space savings are provided, along with a proof of concept framework for both applications of the binary tree type, in the Java programming language.
The device optimization is a very important element in semiconductor technology advancement. Its objective is to find a design point for a semiconductor device so that the optimized design goal meets all specified constraints. As in other engineering fields, a nonlinear optimizer is often used for design optimization. One major drawback of using a nonlinear optimizer is that it can only partially explore the design space and return a local optimal solution. This dissertation provides an adaptive optimization design methodology to allow the designer to explore the design space and obtain a globally optimal solution. One key element of our method is to quickly compute the set of all feasible solutions, also called the acceptability region. We described a polytope-based representation for the acceptability region and an adaptive linearization technique for device performance model approximation. These efficiency enhancements have enabled significant speed-up in estimating acceptability regions and allow acceptability regions to be estimated for a larger class of device design tasks. Our linearization technique also provides an efficient mechanism to guarantee the global accuracy of the computed acceptability region. To visualize the acceptability region, we study the orthogonal projection of high-dimensional convex polytopes and propose an output sensitive algorithm for projecting polytopes into two dimensions.
The C Navigational System (CNS) is a proposed programming environment for the C programming language. The introduction covers the major influences of programming environments and the components of a programming environment. The system is designed to support the design, coding and maintenance phases of software development. CNS provides multiple views to both the source and documentation for a programming project. User-defined and system-defined links allow the source and documentation to be hierarchically searched. CNS also creates a history list and function interface for each function in a module. The final chapter compares CNS and several other programming environments (Microscope, Rn, Cedar, PECAN, and Marvel).
This paper presents the scan-line algorithm which has been implemented on the Lisp Machine. The scan-line algorithm resides beneath a library of primitive software routines which draw more fundamental objects: lines, triangles and rectangles. This routine, implemented in microcode, applies the A(BC)*D approach to word boundary alignments in order to create an extremely fast, efficient, and general purpose drawing primitive. The scan-line algorithm improves on previous methodologies by limiting the number of CPU intensive instructions and by minimizing the number of words referenced. This paper will describe how to draw scan-lines and the constraints imposed upon the scan-line algorithm by the Lisp Machine's hardware and software.
New peripheral devices are being developed at an ever increasing rate. Before such accessories can be used in the UNIX environment (UNIX is a trademark of Bell Laboratories), they must be able to communicate with the operating system. This involves writing a device driver for each device. In order to do this, very detailed knowledge is required of both the device to be integrated and the version of UNIX to which it will be attached. The process is long, detailed and prone to subtle problems and errors. This paper presents a menu-driven utility designed to simplify and accelerate the design and implementation of UNIX device drivers by freeing developers from many of the implementation specific low-level details.
A useful compiler has been designed that takes a high level neural network specification and constructs a low level configuration file explicitly specifying all network parameters and connections. The neural network model for which this compiler was designed is the adaptrode neuronal model, and the configuration file created can be used by the Adnet simulation engine to perform network experiments. The specification language is very flexible and provides a general framework from which almost any network wiring configuration may be created. While the compiler was created for the specialized adaptrode model, the wiring specification algorithms could also be used to specify the connections in other types of networks.
Because of an interest in object-oriented database systems, designers have created systems to store and manipulate specific sets of abstract data types that belong to the real world environment they represent. Unfortunately, the advantage of these systems is also a disadvantage since no single object-oriented database system can be used for all applications. This paper describes an object-oriented database management system called the Object-oriented Database Editor (ODE) which overcomes this disadvantage by allowing designers to create and execute an object-oriented database that represents any type of environment and then to store it and simulate that environment. As conditions within the environment change, the designer can use ODE to alter that environment without loss of data. ODE provides a flexible environment for the user; it is efficient; and it can run on a personal computer.
Efficient unification is considered within the context of logic programming. Unification is explained in terms of equivalence classes made up of terms, where there is a constraint that no equivalence class may contain more than one function term. It is demonstrated that several well-known "efficient" but nonlinear unification algorithms continually maintain the said constraint as a consequence of their choice of data structure for representing equivalence classes. The linearity of the Paterson-Wegman unification algorithm is shown largely to be a consequence of its use of unbounded lists of pointers for representing equivalences between terms, which allows it to avoid the nonlinearity of "union-find".
This paper reviews numerous theoretical results on control structures and demonstrates their practical examples. This study deals with the design of run-time support routines by using top-down structured programming technique. A number of examples are given as illustration of this method. In conclusion, structured programming has proved to be an important methodology for systematic program design and development.
This paper discusses FORTRAN optimizations that the user can perform manually at the source code level to improve object code performance. It makes use of descriptive examples within the text of the paper for explanatory purposes. The paper defines key areas in writing a FORTRAN program and recommends ways to improve efficiency in these areas.
The analysis of an executing program and the isolation of critical code has been a problem since the first program was written. This thesis examines the process of program analysis through the use of a software monitoring system. Since there is a trend toward structured languages a subset of PL/I was developed t~o exhibit source statement monitoring and costing techniques. By filtering a PL/W program through a preorocessor which determines the cost of source statements and inserts monitoring code, a post-execution analysis of the program can be obtained. This analysis displays an estimated time cost for each source statements the number of times the statement w3s executed, and the product of these values. Additionally, a bar graph is printed in order to quickly locate very active code.
The purpose of this research was to investigate the generation of machine code from high-level programming language. The following steps were undertaken: 1) Choose a high-level programming language as the source language and a computer as the target computer. 2) Examine all stages during the compiling of a high-level programming language and all data sets involved in the compilation. 3) Discover the mechanism for generating machine code and the mechanism to generate more efficient machine code from the language. 3) Construct an algorithm for generating machine code for the target computer. The results suggest that compiler is best implemented in a high-level programming language, and that SCANNER and PARSER should be independent of target representations, if possible.
Synthetic seismograms are a computer-generated aid in the search for hydrocarbons. Heretofore the solution has been done by z-transforms. This thesis presents a solution based on the method of finite differences. The resulting algorithm is fast and compact. The method is applied to three variations of the problem, all three are reduced to the same approximating equation, which is shown to be optimal, in that grid refinement does not change it. Two types of algorithms are derived from the equation. The number of obvious multiplications, additions and subtractions of each is analyzed. Critical section of each requires one multiplication, two additions and two subtractions. Four sample synthetic seismograms are shown. Implementation of the new algorithm runs twice as fast as previous computer program.
In this thesis, the first chapter provides the general description of this interpreter. The second chapter contains a formal definition of the syntax of BASIC along with an introduction to the semantics. The third chapter contains the design of data structure. The fourth chapter contains the description of algorithms along with stages for testing the interpreter and the design of debug output. The stages and actions-are represented internally to the computer in tabular forms. For statement parsing working syntax equations are established. They serve as standards for the conversion of source statements into object pseudocodes. As the statement is parsed for legal form, pseudocodes for this statement are created. For pseudocode execution, pseudocodes are represented internally to the computer in tabular forms.
This thesis describes the implementation of a general purpose personal information storage and retrieval system. Chapter one contains an introduction to information storage and retrieval. Chapter two contains a description of the features a useful personal information retrieval system should contain. This description forms the basis for the implementation of the personal information storage and retrieval system described in chapter three. The system is implemented in UCSD Pascal on an Apple II microcomputer.
The purpose of this project was to describe a novel design approach for a digital computer peripheral controller, then design and construct a case study controller. This document consists of three chapters and an appendix. Chapter II presents the design approach chosen; a variation to a design presented by Charles R. Richards in an article published in Electronics magazine. Richards' approach consists of a finite state machine circuitry controlling all the functions of a controller. The variation to Richards' approach consists of considering the various logically independent processes which a controller carries out and assigning control of each process to a separate finite state machine. The appendix contains the documentation of the design and construction of the controller.
Resources are said to be fragmented in the network when they are available in non-contiguous blocks, and calls are dropped as they may not end sufficient resources. Hence, available resources may remain unutilized. In this thesis, the effect of resource fragmentation (RF) on RSVP-controlled networks was studied and new algorithms were proposed to reduce the effect of RF. In order to minimize the effect of RF, resources in the network are dynamically redistributed on different paths to make them available in contiguous blocks. Extra protocol messages are introduced to facilitate resource redistribution in the network. The Dynamic Resource Redistribution (DRR) algorithm when used in conjunction with RSVP, not only increased the number of calls accommodated into the network but also increased the overall resource utilization of the network. Issues such as how many resources need to be redistributed and of which call(s), and how these choices affect the redistribution process were investigated. Further, various simulation experiments were conducted to study the performance of the DRR algorithm on different network topologies with varying traffic characteristics.
Many activities are comprised of temporally dependent events that must be executed in a specific chronological order. Supportive software applications must preserve these temporal dependencies. Whenever the processing of this type of an application includes transactions submitted to a database that is shared with other such applications, the transaction concurrency control mechanisms within the database must also preserve the temporal dependencies. A basis for preserving temporal dependencies is established by using (within the applications and databases) real-time timestamps to identify and order events and transactions. The use of optimistic approaches to transaction concurrency control can be undesirable in such situations, as they allow incorrect results for database read operations. Although the incorrectness is detected prior to transaction committal and the corresponding transaction(s) restarted, the impact on the application or entity that submitted the transaction can be too costly. Three transaction concurrency control algorithms are proposed in this dissertation. These algorithms are based on timestamp ordering, and are designed to preserve temporal dependencies existing among data-dependent transactions. The algorithms produce execution schedules that are equivalent to temporally ordered serial schedules, where the temporal order is established by the transactions' start times. The algorithms provide this equivalence while supporting currency to the extent out-of-order commits and reads. With respect to the stated concern with optimistic approaches, two of the proposed algorithms are risk-free and return to read operations only committed data-item values. Risk with the third algorithm is greatly reduced by its conservative bias. All three algorithms avoid deadlock while providing risk-free or reduced-risk operation. The performance of the algorithms is determined analytically and with experimentation. Experiments are performed using functional database management system models that implement the proposed algorithms and the well-known Conservative Multiversion Timestamp Ordering algorithm.
This research focused on the off-line cursive script recognition application. The problem is very large and difficult and there is much room for improvement in every aspect of the problem. Many different aspects of this problem were explored in pursuit of solutions to create a more practical and usable off-line cursive script recognizer than is currently available.
Surface fitting methods play an important role in many scientific fields as well as in computer aided geometric design. The problem treated here is that of constructing a smooth surface that interpolates data values associated with scattered nodes in the plane. The data is said to be convex if there exists a convex interpolant. The problem of convexity-preserving interpolation is to determine if the data is convex, and construct a convex interpolant if it exists.
DNA sequence analysis involves precise discrimination of two of the sequence's most important components: exons and introns. Exons encode the proteins that are responsible for almost all the functions in a living organism. Introns interrupt the sequence coding for a protein and must be removed from primary RNA transcripts before translation to protein can occur. A pattern recognition technique called Finite Induction (FI) is utilized to study the language of exons and introns. FI is especially suited for analyzing and classifying large amounts of data representing sequences of interest. It requires no biological information and employs no statistical functions. Finite Induction is applied to the exon and intron components of DNA by building a collection of rules based upon what it finds in the sequences it examines. It then attempts to match the known rule patterns with new rules formed as a result of analyzing a new sequence. A high number of matches predict a probable close relationship between the two sequences; a low number of matches signifies a large amount of difference between the two. This research demonstrates FI to be a viable tool for measurement when known patterns are available for the formation of rule sets.
We extend the Jinni mobile agent architecture with a multicast network transport layer, an agent-to-agent delegation mechanism and a reflection based Prolog-to-Java interface. To ensure that our agent infrastructure runs efficiently, independently of router-level multicast support, we describe a blackboard based algorithm for locating a randomly roaming agent. As part of the agent-to-agent delegation mechanism, we describe an alternative to code-fetching mechanism for stronger mobility of mobile agents with less network overhead. In the context of direct and reflection based extension mechanisms for Jinni, we describe the design and the implementation of a reflection based Prolog-to-Java interface. The presence of subtyping and method overloading makes finding the most specific method corresponding to a Prolog call pattern fairly difficult. We describe a run-time algorithm which provides accurate handling of overloaded methods beyond Java's reflection package's limitations.
This dissertation addresses the general problem of recognition of acoustic signals which may be derived from speech, sonar, or acoustic phenomena. The specific problem of recognizing speech is the main focus of this research. The intention is to design a recognition system for a definite number of discrete words. For this purpose specifically, eight isolated words from the T1MIT database are selected. Four medium length words "greasy," "dark," "wash," and "water" are used. In addition, four short words are considered "she," "had," "in," and "all." The recognition system addresses the following issues: filtering or preprocessing, training, and decision-making. The preprocessing phase uses linear predictive coding of order 12. Following the filtering process, a vector quantization method is used to further reduce the input data and generate a finite inductive sequence of symbols representative of each input signal. The sequences generated by the vector quantization process of the same word are factored, and a single ruling or reference template is generated and stored in a codebook. This system introduces a new modeling technique which relies heavily on the basic concept that all finite sequences are finitely inductive. This technique is used in the training stage. In order to accommodate the variabilities in speech, the training is performed casualty, and a large number of training speakers is used from eight different dialect regions. Hence, a speaker independent recognition system is realized. The matching process compares the incoming speech with each of the templates stored, and a closeness ration is computed. A ratio table is generated anH the matching word that corresponds to the smallest ratio (i.e. indicating that the ruling has removed most of the symbols) is selected. Promising results were obtained for isolated words, and the recognition rates ranged between 50% and 100%.
The study of multicomputer interconnection networks is an important area of research in parallel processing. We introduce vertex-symmetric Hamming-group graphs as a model to design a wide variety of network topologies including the hypercube network.
There have been a number of different versioning models proposed. The research in this area can be divided into two categories: object versioning and schema versioning. In this dissertation, both problem domains are considered as a single unit. This dissertation describes a unifying version model (UVM) for maintaining changes to both objects and schema. UVM handles schema versioning operations by using object versioning techniques. The result is that the UVM allows the OODBMS to be much smaller than previous systems. Also, programmers need know only one set of versioning operations; thus, reducing the learning time by half. This dissertation shows that UVM is a simple but semantically sound and powerful version model for both objects and schema.
Genetic algorithm and artificial life techniques are applied to the development of challenging and interesting opponents in a combat-based computer game. Computer simulations are carried out against an idealized human player to gather data on the effectiveness of the computer generated opponents.
There are three main results in this dissertation. They are PLS-completeness of discrete Hopfield network convergence with eight different restrictions, (degree 3, bipartite and degree 3, 8-neighbor mesh, dual of the knight's graph, hypercube, butterfly, cube-connected cycles and shuffle-exchange), exponential convergence behavior of discrete Hopfield network, and simulation of Turing machines by discrete Hopfield Network.
The goal of a parallel algorithm is to solve a single problem using multiple processors working together and to do so in an efficient manner. In this regard, there is a need to categorize strategies in order to solve broad classes of problems with similar structures and requirements. In this dissertation, two parallel algorithm design strategies are considered: linked list ranking and parentheses matching.
This investigation proposes a powerful formalization of common sense knowledge based on function-free normal deduction graphs (NDGs) which form a powerful tool for deriving Horn and non-Horn clauses without functions. Such formalization allows common sense reasoning since it has the ability to handle not only negative but also incomplete information.
Numerical methods are usually necessary in solving Hamiltonian systems since there is often no closed-form solution. By utilizing a general property of Hamiltonians, namely the symplectic property, all of the qualities of the system may be preserved for indefinitely long integration times because all of the integral (Poincare) invariants are conserved. This allows for more reliable results and frequently leads to significantly shorter execution times as compared to conventional methods. The resonant triad Hamiltonian with one degree of freedom will be focused upon for most of the numerical tests because of its difficult nature and, moreover, analytical results exist whereby useful comparisons can be made.
The efficacy of a machine learning technique is domain dependent. Some machine learning techniques work very well for certain domains but are ill-suited for other domains. One area that is of real-world concern is the flexibility with which machine learning techniques can adapt to dynamic domains. Currently, there are no known reports of any system that can learn dynamic domains, short of starting over (i.e., re-running the program). Starting over is neither time nor cost efficient for real-world production environments. This dissertation studied a method, referred to as Experience Based Learning (EBL), that attempts to deal with conditions related to learning dynamic domains. EBL is an extension of Instance Based Learning methods. The hypothesis of the study related to this research was that the EBL method would automatically adjust to domain changes and still provide classification accuracy similar to methods that require starting over. To test this hypothesis, twelve widely studied machine learning datasets were used. A dynamic domain was simulated by presenting these datasets in an uninterrupted cycle of train, test, and retrain. The order of the twelve datasets and the order of records within each dataset were randomized to control for order biases in each of ten runs. As a result, these methods provided datasets that represent extreme levels of domain change. Using the above datasets, EBL's mean classification accuracies for each dataset were compared to the published static domain results of other machine learning systems. The results indicated that the EBL's system performance was not statistically different (p>0.30) from the other machine learning methods. These results indicate that the EBL system is able to adjust to an extreme level of domain change and yet produce satisfactory results. This finding supports the use of the EBL method in real-world environments that incur rapid changes to both variables and ...
The scope of the present dissertation is a deep lossy compression of still and moving grayscale pictures while maintaining their fidelity, with a specific goal of creating a working prototype of a software system for use in low bandwidth transmission of still satellite imagery and weather briefings with the best preservation of features considered important by the end user.
Geometric packing problems are NP-complete problems that arise in VLSI design. In this thesis, we present two novel algorithms using dynamic programming to compute exactly the maximum number of k x k squares of unit size that can be packed without overlap into a given n x m grid. The first algorithm was implemented and ran successfully on problems of large input up to 1,000,000 nodes for different values. A heuristic based on the second algorithm is implemented. This heuristic is fast in practice, but may not always be giving optimal times in theory. However, over a wide range of random data this version of the algorithm is giving very good solutions very fast and runs on problems of up to 100,000,000 nodes in a grid and different ranges for the variables. It is also shown that this version of algorithm is clearly superior to the first algorithm and has shown to be very efficient in practice.
Guiding navigation in virtual environments (VEs) is a challenging task. A key issue in the navigation of a virtual environment is to be able to strike a balance between the user's need to explore the environment freely and the designer's need to ensure that the user experiences all the important events in the VE. This thesis reports on a study aimed at comparing the effectiveness of various navigation cues that are used to motivate users towards a specific target location. The results of this study indicate some significant differences in how users responded to the various cues.
With the growing popularity of e-commerce applications that use software agents, the protection of mobile agent data has become imperative. To that end, the performance of four methods that protect the data integrity of mobile agents is evaluated. The methods investigated include existing approaches known as the Partial Result Authentication Codes, Hash Chaining, and Set Authentication Code methods, and a technique of our own design, called the Modified Set Authentication Code method, which addresses the limitations of the Set Authentication Code method. The experiments were run using the DADS agent system (developed at the Network Research Laboratory at UNT), for which a Data Integrity Module was designed. The experimental results show that our Modified Set Authentication Code technique performed comparably to the Set Authentication Code method.
Peer-to-Peer (P2P) networks have seen tremendous growth in development and usage in recent times. This attention has brought many developments as well as new challenges to these networks. We will show that agent extensions to P2P networks offer solutions to many problems faced by P2P networks. In this research, an attempt is made to bring together JXTA P2P infrastructure and Jinni, a Prolog based agent engine to form an agent based P2P network. On top of the JXTA, we define simple Java API providing P2P services for agent programming constructs. Jinni is deployed on this JXTA network using an automated code update mechanism. Experiments are conducted on this Jinni/JXTA platform to implement a simple agent communication and data exchange protocol.
Automatic memory management is crucial in implementation of runtime systems even though it induces a significant computational overhead. In this thesis I explore the use of statistical properties of the directed graph describing the set of live data to decide between garbage collection and heap expansion in a memory management algorithm combining the dynamic array represented heaps with a mark and sweep garbage collector to enhance its performance. The sampling method predicting the density and the distribution of useful data is implemented as a partial marking algorithm. The algorithm randomly marks the nodes of the directed graph representing the live data at different depths with a variable probability factor p. Using the information gathered by the partial marking algorithm in the current step and the knowledge gathered in the previous iterations, the proposed empirical formula predicts with reasonable accuracy the density of live nodes on the heap, to decide between garbage collection and heap expansion. The resulting heuristics are tested empirically and shown to improve overall execution performance significantly in the context of the Jinni Prolog compiler's runtime system.
Agent-oriented software engineering (AOSE) covers issues on developing systems with software agents. There are many techniques, mostly agent-oriented and object-oriented, ready to be chosen as building blocks to create agent-based systems. There have been several AOSE methodologies proposed intending to show engineers guidelines on how these elements are constituted in having agents achieve the overall system goals. Although these solutions are promising, most of them are designed in ad-hoc manner without truly obeying software developing life-cycle fully, as well as lacking of examinations on agent-oriented features. To address these issues, we investigated state-of-the-art techniques and AOSE methodologies. By examining them in different respects, we commented on the strength and weakness of them. Toward a formal study, a comparison framework has been set up regarding four aspects, including concepts and properties, notations and modeling techniques, process, and pragmatics. Under these criteria, we conducted the comparison in both overview and detailed level. The comparison helped us with empirical and analytical study, to inspect the issues on how an ideal agent-based system will be formed.
The widely used vector model maintains its popularity because of its simplicity, fast speed, and the appeal of using spatial proximity for semantic proximity. However, this model faces a disadvantage that is associated with the vagueness from keywords overlapping. Efforts have been made to improve the vector model. The research on improving document representation has been focused on four areas, namely, statistical co-occurrence of related items, forming term phrases, grouping of related words, and representing the content of documents. In this thesis, we propose the idea-indexing model to improve document representation for the filtering task in IR. The idea-indexing model matches document terms with the ideas they express and indexes the document with these ideas. This indexing scheme represents the document with its semantics instead of sets of independent terms. We show in this thesis that indexing with ideas leads to better performance.
Natural language understanding has been a persistent challenge to researchers in various computer science fields, in a number of applications ranging from user support systems to entertainment and online teaching. A long term goal of the Artificial Intelligence field is to implement mechanisms that enable computers to emulate human dialogue. The recently developed ALICEbots, virtual agents with underlying AIML scripts, by A.L.I.C.E. foundation, use AIML scripts - a subset of XML - as the underlying pattern database for question answering. Their goal is to enable pattern-based, stimulus-response knowledge content to be served, received and processed over the Web, or offline, in the manner similar to HTML and XML. In this thesis, we describe a system that converts the AIML scripts to Prolog clauses and reuses them as part of a knowledge processor. The inference mechanism developed in this thesis is able to successfully match the input pattern with our clauses database even if words are missing. We also emulate the pattern deduction algorithm of the original logic deduction mechanism. Our rules, compatible with Semantic Web standards, bring structure to the meaningful content of Web pages and support interactive content retrieval using natural language.
Effective stewardship of forest ecosystems make it imperative to measure, monitor, and predict the dynamic changes of forest ecology. Measuring and monitoring provides us a picture of a forest's current state and the necessary data to formulate models for prediction. However, societal and natural events alter the course of a forest's development. A simulation environment that takes into account these events will facilitate forest management. In this thesis, we describe an efficient parallel implementation of a land cover use model, Mosaic, and discuss the development efforts to incorporate spatial interaction and succession dynamics into the model. To evaluate the performance of our implementation, an extensive set of simulation experiments was carried out using a dataset representing the H.J. Andrews Forest in the Oregon Cascades. Results indicate that a significant reduction in the simulation execution time of our parallel model can be achieved as compared to uni-processor simulations.
There are two main approaches for intrusion detection: signature-based and anomaly-based. Signature-based detection employs pattern matching to match attack signatures with observed data making it ideal for detecting known attacks. However, it cannot detect unknown attacks for which there is no signature available. Anomaly-based detection builds a profile of normal system behavior to detect known and unknown attacks as behavioral deviations. However, it has a drawback of a high false alarm rate. In this thesis, we describe our anomaly-based IDS designed for detecting intrusions in cryptographic and application-level protocols. Our system has several unique characteristics, such as the ability to monitor cryptographic protocols and application-level protocols embedded in encrypted sessions, a very lightweight monitoring process, and the ability to react to protocol misuse by modifying protocol response directly.
In order to provide improved service quality to applications, networks need to address the need for reliability of data delivery. Reliability can be improved by incorporating fault tolerance into network routing, wherein a set of multiple routes are used for routing between a given source and destination. This thesis proposes a new fault-tolerant protocol, called the Multipath Fault Tolerant Protocol for Routing (MFTPR), to improve the reliability of network routing services. The protocol is based on a multipath discovery algorithm, the Quasi-Shortest Multipath (QSMP), and is designed to work in conjunction with the routing protocol employed by the network. MFTPR improves upon the QSMP algorithm by finding more routes than QSMP, and also provides for maintenance of these routes in the event of failure of network components. In order to evaluate the resilience of a pair of paths to failure, this thesis proposes metrics that evaluate the non-disjointness of a pair of paths and measure the probability of simultaneous failure of these paths. The performance of MFTPR to find alternate routes based on these metrics is analyzed through simulation.
The resources (memory, power and bandwidth) are limited in wireless and mobile networks. Previous research has shown that the quality of service (QoS) of the mobile client can be improved through efficient resources management. This thesis contains two areas of research that are strongly interrelated. In the first area of research, we extended the MoSync Algorithm, a network application layer media synchronization algorithm, to allow play-out of multimedia packets by the base station upon the mobile client in a First-In-First-Out (FIFO), Highest-Priority-First (PQ), Weighted Fair-Queuing (WFQ) and Round-Robin (RR) order. In the second area of research, we make modifications to the DSR and TORA routing algorithms to make them energy aware routing protocols. Our research shows that the QoS of the mobile client can be drastically improved through effective resource allocation.
Wireless ad hoc networks are expected to play an important role in civilian and military settings where wireless access to wired backbone is either ineffective or impossible. Wireless sensor networks are effective in remote data acquisition. Congestion control and power consumption in wireless ad hoc networks have received a lot of attention in recent research. Several algorithms have been proposed to reduce congestion and power consumption in wireless ad hoc and sensor networks. In this thesis, we focus upon two schemes, which deal with congestion control and power consumption issues. This thesis consists of two parts. In the first part, we describe a randomization scheme for congestion control in dynamic source routing protocol, which we refer to as RDSR. We also study a randomization scheme for GDSR protocol, a GPS optimized variant of DSR. We discuss RDSR and RGDSR implementations and present extensive simulation experiments to study their performance. Our results indicate that both RGDSR and RDSR protocols outperform their non-randomized counterparts by decreasing the number of route query packets. Furthermore, a probabilistic congestion control scheme based on local tuning of routing protocol parameters is shown to be feasible. In the second part we present a simulation based performance study of energy aware data centric routing protocol, EAD, proposed by X. Cheng and A. Boukerche. EAD reduces power consumption by requiring only a small percentage of the network to stay awake. Our experiments show that EAD outperforms the well-known LEACH scheme.
This dialog allows you to filter your current search.
Each of the Years listed note their name and the number of records that will be limited down to if you choose that option.
The list can be sorted by name or the count.
Having trouble finding an option within the list of Years? Start typing and we'll update the list to show only those items that match your needs.