Mathematics, computer science and statistics.

Chair: Joseph Kolibal, University of Southern Mississippi

Vice-chair: Andrew Harrell, CEWES-GM

THURSDAY MORNING

Hunter Henry Executive Room 12

9:30 COMPILED ENGLISH

A. Louise Perkins (1*), Tom Rishel (2), Farnaz Zand (1), and Sumanth Yenduri (1), University of Southern Mississippi, Hattiesburg, MS 39406 and Hancock Bank, Gulfport, MS 39501

Chomsky's hierarchical structure for spoken languages has served as the cornerstone of Computer Programming Language theory for decades. However, typical Programming Languages are still only Context Free because we are still unclear on how to disambiguate context sensitive grammars. In ongoing research we are using a suite of Context Free Grammars coupled to data mining algorithms that define text. These algorithms ares stochastic, so that errors can still occur. None-the-less the ambiguity problem is removed from the grammar context and promoted to a hyperlevel where context defining tools already exist. This shift in compiling paradigm allows us to build simple English language to programming compilers for narrowly defined contexts such as Banking.

10:00 SUITABILITY OF ENVIRONMENTS FOR NOVICE JAVA PROGRAMMERS

Edward Dillon* and Dawn Wilkins, University of Mississippi, University, MS 38677

For programmers who are new to the language of Java, it is traditional that they start writing programs on some simple text editor like Notepad and use a separate compiler invoked at the Command Prompt. Alternatively, an environment known as an Integrated Development Environment (IDE) is available to aid users in writing code and producing programs more efficiently. However, there are some problems concerning whether IDEs are too complex for novice users to understand. This research carefully examines Java IDEs to compare and contrast the IDE vs text editor approach, with the goal of recommending an appropriate software environment for novice programmers.

10:30 NETWORK ROUTING AND THE IMPLEMENTION OF UNIT/TASK SCHEDULING SOLUTIONS

Andrew Harrell, U.S. Army Engineer Research and Development Center (ERDC), Vicksburg, MS 39180

For a long time discussion with the US Army combat engineer units and members of the US Army engineer community have resulted in the statement of a need for a rapid and accurate obstacle system and logistics haul effectiveness capability. The network analysis methodology to be presented in this talk was developed several years ago and has applications to both of these problems. It can also solve the problem of how to implement the an optimum solution (how to route it through an on and off-road network of tactical unit movement possibilities). Two different C-code versions of variations of the basic Edmonds-Karp-Dinic network maximum flow/min-cost algorithms exist. They, along with a Unix screen GUI, enable a user to solve routing and scheduling problems interactively. The programs that compute the optimum solution first compute best vehicle traverses across off and on-road digital map data. Then (after defining an overall network from a set of given tasks along with starting, intermediate, and end points) the user can compute the overall traffic throughput and display a list (ranked according to time and traffic flow rates) of possibilities to schedule units to move through the network to perform these tasks.

11:00 NETWORK FIREWALLS SECURITY

James Haywood, Jr., Mississippi Valley State University, Itta Bena, MS 38941

This project was based upon testing the security of the network firewall rule changes. There were three basic key points that was mandatory to investigate the security of the network firewall rule changes which were the collection of data, the network firewall rule change test, and the results. The collection of data consisted of four subunits of research. These subunits were the approver (process owner) of the change, the person who gained the access by the approver, reason for the firewall rule change and the recent status of the changed firewall rule. In gathering the data, a sample selection is acquired through the total amount of network firewall rule changes.

The testing part of the project consists of analyzing the verification of the necessary approvers, the person granted access to make the change, documentation for the reason of change and current status report are all acquired from samples picked. In order to test the firewalls a sample portion of the data was collected. The tool used to conduct the test was the Auditor-Type Toolkit which allows template of a testing sheet that compiles all necessary verification into a spreadsheet.

The results of the network security firewall test were base upon the finds of actual testing. If any data was missing from the acquired sample, the process owner of the firewalls intrusion detection was notified and had to find the missing data. If process owner couldn't find the missing data then the firewall rule change was reverted.

11:30 A CREDIBILITY-BASED SECURE SOURCE ROUTING PROTOCOL FOR WIRELESS AD HOC NETWORKS

Natarajan Meghanathan (1*) and Michael Williams (2), (1) Jackson State University, Jackson, MS 39217 and (2) University of North Texas, Denton, TX 76203

We propose a credibility-based secure source routing (CSSR) protocol for wireless ad hoc networks. We consider a realistic model of node behavior in which data and control packets are dropped arbitrarily. In order to increase chances of packet delivery, packets have to be sent through intermediate nodes that have a good history of forwarding packets for their peers. Each node in the network maintains a credibility table that stores the credibility points (proportional to the number of packets forwarded) for each other node in the network. Credibility points for a node are incremented by a constant value upon successful packet forward, and halved when the node drops a packet. If the credibility points for a node are not updated for a certain time period, then the value is reset to 0. The credibility of a route is defined as the lowest of the credibility points of the constituent nodes (excluding the source and destination) of the route. CSSR chooses the route with the maximum credibility. Our approach achieves two purposes: relieves congestion in the network and bypasses around arbitrarily packet dropping nodes. We simulate CSSR and investigate the adaptation of the protocol to the actual node behavior by using different increment values for the credibility points. We study the performance of CSSR and Dynamic Source Routing (DSR) protocol for different values of the percentage of arbitrarily packet dropping nodes in the network. The performance metrics studied include the packet delivery ratio, delay per packet and the number of route transitions.

THURSDAY AFTERNOON

Hunter Henry Executive Room 12

1:00 THE USE OF LIMITERS IN STOCHASTIC INTERPOLATION

Joseph Kolibal*, University of Southern Mississippi, Hattiesburg, MS 39406

Stochastic interpolation primarily requires that the function generating the row space of the de-convolution and re-convolution matrices be non-negative. This means that a range of functions can be utilized, including those that limit the local growth of the generating function to pre-determined bounds, i.e., limiters. The application of limiters can substantially improve interpolation performance for extremely noisy and irregularly spaced data, and this study examines performance improvements that have successfully enhanced the robustness of the method.

1:30 EFFICIENT SUPPORT VECTOR MACHINES THROUGH PARALLEL COMPUTING

Arun Rajendran*, Peng Li, Chaoyang Zhang, and Youping Deng, University of Southern Mississippi, Hattiesburg, MS 39406

In this study the application of Support Vector Machines for classification of large datasets is investigated. Support Vector Machines is a machine learning algorithm with good generalization properties. But the time complexity of the algorithm depends on the number of examples on the dataset, which makes the problem highly inefficient for large datasets. In this study different methods to improve the time efficiency of Support Vector Machines are analyzed and a design to implement it in parallel is proposed. Also the parallel algorithm\'s performance in different architecture is analyzed and optimized for performance. Parallel implementation improves time efficiency of the learning process, while maintaining the accuracy.

2:00 THE SECOND DERIVATIVES OF THE CHEBYSHEV POLYNOMIALS, [T.sub.N], IN TERMS OF [T.sub.N]

Joseph Kolibal*, University of Southern Mississippi, Hattiesburg, MS 39406

Expressing the second derivative of the Chebyshev polynomials in terms of the Chebyshev polynomials is required in the development of particular solutions associated with the solution of partial differential equations using radial basis functions. There are several approaches for evaluting this expansion, however its form is perhaps most elegant when expressing the expansion in matrix form obtained through the use of nearly orothogonal functions. This matrix is shown to be nonnegative and thus the coefficients in the required sum are nonnegative integers.

2:30 OBSERVING STATISTICS FROM A SCIENTIFIC DATA STANDPOINT

Richard W. Swinney, University of Mississippi, University, MS 38677

The purpose of this research is to express a scientific data environment implementing a particular study, such as Economics, using Statistics. A mapping process method has been designed that executes n-dimensions of data, implemented via a developed Java computer program. The mapped values indicate where an individual subject, such as a state, lies among its co-subjects based on combined factors, such as Education and Commerce. Using the fifty states in the U.S., the results shows where each state lies among other states, representing the combined actual statistics of subjects for each of the fifty states.

3:30 SOLUTION OF MAXWELL'S' EQUATIONS WITHIN A 3 DIMENSIONAL MULTI-BLOCKED CURVILINEAR DOMAIN

Terry Gerald (1*) and Atef Z. Elsherbeni (2), (1) SpecPro LLC, Vicksburg, MS 39180 and (2) University of Mississippi, University, MS 38677

We present a 3D solution technique for solving Maxwell's electromagnetic equations within a composite multi-blocked grid composed of sub-grids defined by curvilinear coordinate systems. The exterior of the composite grid is wrapped by an enclosing shell grid with an embedded convolutional perfectly matched layer which performs the role of an absorbing boundary condition. Sample results for practical applications will be presented showing the advantage of this gridding procedure over the traditional rectangular cells.

4:00 EXISTENCE THEOREM OF ELLIPTIC SYSTEMS WITH SINGULARITIES INVOLVING SOBOLEV OR SOBOLEV-HARDY CRITICAL EXPONENTS

Xiaoqin Wu*, Mississippi Valley State University, Itta Bena, MS 38941

In this research, we study the existence of positive solutions of the system of nonlinear elliptic equations with sigularities involving Sobolev or Sobolev-Hardy critical exponents. This problem is interesting since the Sobolev and Sobolev-Hardy embeddings are not compact for critical exponents. The main tool to prove the existence theorem is Mountain Pass Theorem.

6:00 Poster Session

Location: Bost Auditorium North

Posters may be set up between 4:00p and 4:30p

GESTALT PRINCIPLES APPLIED TO SOFTWARE-ENGINEERING DIAGRAMS: AN INITIAL STUDY

Krystle Lemon*, Edward B. Allen, Jeffrey Carver, and Gary Bradshaw, Mississippi State University. Mississippi State, MS 39762

Discovering root-causes of comprehension errors in software design is important to prevent their presence in software systems. This research synthesizes software engineering and Gestalt principles of similarity, proximity, continuity for the purpose of discovering whether certain visual attributes of diagrams (dashed arrows, severe complexity, etc.) can affect the accuracy and efficiency of understanding correct relationships amongst the entities in the diagram. Twenty-seven subjects viewed diagrams of different types and answered questions about them. The experiment tested whether two dependent variables, accuracy and response time, were significantly affected by independent variables, diagram type (simple1, simple2, complex), Gestalt principles (good vs. bad), and forward/backward (question order). The results of this study indicated that the Gestalt principles did affect the comprehension in the complex diagrams. Further post hoc analysis was conducted on the data that revealed reaction time speed up occurred in all diagrams because of the type of questions presented to the subjects. Distance metrics such as line length, number of lines crossing, lines going into a box, lines coming out of a box, lines per diagram, boxes per diagram, and number of bends per line were measured and used as additional variables in the analysis. Interactions between number of lines per diagram, number of bends per line, and number of lines going into a box help to predict the average reaction time of the subjects.

ONTOLOGY ENGINEERING: METADATA MANAGEMIENT IN A RAPID PROTOTYPING CAPABILITY FOR NASA EARTH SCIENCES

Robert A. Elliott, Sr., Mississippi State University, Mississippi State, MS 39762

NASA, Mississippi State University, and other research organizations, are collaborating to develop a Rapid Prototyping Capability to advance Earth sciences. The Rapid Prototyping Capability will manipulate a wide variety of metadata and our research in Ontology Engineering will support more uniform handling of metadata. Ontology services and ontology engineering practices would facilitate experiments using NASA science results obtained from satellite and sensor data in the context of the Rapid Prototyping Capability. This poster defines ontology and introduces ontology engineering principles that describe the creation of ontologies, the tools used in ontology creation and the methodologies used in practice. Ontologies provide formal definitions of data and support sharing of knowledge bases. Ontology engineering will be applied to data sharing issues thereby enabling the merging of ontologies (knowledge bases) and creating new knowledge and data models. Ontology engineering practices are a basis of sound ontology creation and are rapidly growing as more user communities standardize this discipline. Our future work will apply ontology engineering to advance Earth science through the NASA Rapid Prototyping Capability.

AN ANALYSIS OF TEST COVERAGE FOR THE RULE BASED SYSTEM, LOCI

Stephen Medders*, Mississippi State University, Mississippi State, MS 39762

Rule based systems use a set of logical rules to decide which computations to do given the set of data. Loci is a rule based system that compiles a set of rules, automatically designs a parallel architecture for running the computations, and then returns results of the computations based on queries to Loci. CHEM code uses Loci to build models of partial differential equations (PDE) for use in computational fluid dynamics. These models are tested using a test suite called Quick Test. Quick Test runs a set of test data on the models used by CHEM code to verify that the PDE's in the models are in fact being computed by Loci. Currently, there is no method used to show the adequacy of the tests in terms of coverage. Traditional coverage tools measure lines of source code covered, logical paths covered, and logical conditions covered. However, these do not cover rule usage. Rule coverage of the test set should include data flow between rules and fault propagation from one rule to another. Fault injection, inserting a fault by altering a value or the code itself, is one method to check this. Future work will include performing this fault injection on the Quick Test suite, as well as identifying data flow paths and rule dependency graphs in order to evaluate rule coverage of the Quick Test suite.

FRIDAY MORNING

Hunter Henry Executive Room 12

9:00 USING ALICE TO DECREASE ATTRITION IN CS1

Constance Bland*, Timothy Holston, and Udar Kumar Kudikyala, Mississippi Valley State University, Itta Bena, MS 38941

A disturbing trend has resulted over the last decade as it relates to the number of students pursuing undergraduate computer science degrees. Specifically, the entering student population at Mississippi Valley State University (MVSU) has begun to mirror the national trend of showing a decline in the number of students declaring computer science as a major. The decline in the number of incoming students declaring computer science as a major coupled with the large attrition rate are the two factors most responsible for this alarming trend. In this paper, we discuss our strategy to affect the growing attrition rate of computer science majors. During the CS0 course, students are provided with an introduction to programming using Alice, a 3-D animation programming environment developed at Carnegie Mellon University. This introduction to Alice serves the dual purpose of molding students' work habits into ones which facilitate their success in the major as well as providing a mechanism for demonstrating abstract programming concepts in a concrete manner. Preliminary data indicates that students exposed to programming using Alice seamlessly develop traits and characteristics which are necessary to succeed in the discipline of computer science and subsequently are retained in the major at a higher rate than their counterparts.

9:30 E-SERVICING: THE USES AND IMPORTANCE OF THE NEW IEEE STANDARD

Sumanth Yenduri*, A. Louise Perkins, Smvanthi Munagala, and Famaz Zand, University of Southern Mississippi, Long Beach, MS 39560

Everybody has to participate in protecting environment and that encompasses producers and consumers to recycle electronic tools and equipment. Organizations and governments all around the world are now pressurizing everybody into e-protection through various rules, protocols and directives. Recently, a new IEEE standard 1680 has been released with the aim to attract everybody. The standard will make recycling huge amounts of outdated equipment kept in waste storage facilities much easier. The standard was announced in May at the IEEE International Symposium on Electronics and Environment. The symposium was attended by academicians, industry personnel, manufacturers, consumers, environmentalists, and engineers. The "Standard for Environmental Assessment of Personal Computer Products" is the first IEEE standard and the first US standard on computer recycling. The Green Electronics Council (GEC) has listings for various products that comply with the standard proposed although the compliance is not mandatory. The standard covers only for computers and monitors as of now. As new computer technologies make many "not-so-old" computers obsolete although they are still functional accounts to the bulk of electronic waste produced (236.1 million tons only in US in 2003). Currently, less than 1/5th of the e-waste is being recycled. The standard promotes extension of the life of the computers and also makes them easier to recycle. Modern day computers are difficult to recycle and may have hazardous material which cannot be recycled. There is no clear indication of what needs to be recycled and what not. In this paper, we outline the standard in more understandable terms by describing the guidelines, uses and the significance to adopt it. The need to understand the practices mentioned in the standard and to apply them is needed in today's world than every before.

10:00 A BLUEPRINT FOR AN RFID SMART FLOOR FOR EFFICIENT SHIPYARD ASSEMBLY YARD INVENTORY TRACKING

Andrew Strelzoff* and Tulio Sulbaren, University of Southern Mississippi, Hattiesburg, MS 39406

One of the major challenges for the industries serving the NAVY and the Air force is the tracking of current and future inventory of parts required for ship building. The tracking of parts is very important because it has a direct impact on the construction schedule and cost of the ships.

One of the possible technologies that could be used to track ship building parts is radio frequency identification (RFID) systems. Unfortunately, a blueprint to implement RFID technology with the shipyard conditions has not been developed. Shipyards poses special problems for RFID technology such as: 1-Scale of the operation stretched across thousands of feet, 2-Large number of parts in the tens of thousands, 3-Diversity of parts ranging from tiny wiring leads to entire engines weighing thousands of pounds, 4-Large metal objects such as hull assemblies or metal tubing which may reflect radio signals causing false or confused signals, 5-Sensor vulnerability to interference from shipyard manufacturing and test operations and 6-Sensitivity of shipyard instruments to interference from the tracking system. This paper presents the design parameters for an outdoor RFID "smart floor" to track thousands of parts through a shipyard. Based on the design parameters, this paper summarizes the currently available RFID technology and provides a blueprint for a durable, accurate, scalable, effective inventory tracking system based on RFID technology.

10:30 WIRELESS SENSOR-BASED LOCATION FOR MANUFACTURING

Shirong Du* and John N. Daigle, University of Mississippi, University, MS 38677

It is a difficult task to locate and track a large number of RFID-tagged metal containers by wireless sensor-based approaches inside a manufacturing plant due to the hostile and dynamically changing radio propagation environment. We present a top-container detection and location strategy to locate all the containers, including those from which the location sensors cannot receive any signal or any direct signal. Our strategy is based on the ad hoc information gathering and appropriate location methodologies. We also investigate the performance of various methodologies in locating containers under the radio Rayleigh fading model on manufacturing plant premise. The methodologies can be divided into two categories: non-distance-based methodologies, which include position centroid and RSS pattern matching, and distance-based methodologies, which include triangulation and optimization. We examine the problems in triangulation location and use a least mean-square error optimization technique to locate containers. We find that when distance measurements from a small number of sensors are available and the measurement error is large, the optimization methodology can achieve lower location error than triangulation. When many distance measurements are available, triangulation has lower location error than optimization, but optimization performs much faster than triangulation.

11:00 Business Meeting

FRIDAY AFTERNOON

Bost Auditorium South

1:15 MISSISSIPPI CENTER FOR SUPERCOMPUTING RESEARCH (MCSR) USER ADVISORY GROUP MEETING, POSTER SESSION, AND SPECIAL SUBSESSION ON SUPERCOMPUTING

David Roach, University of Mississippi, University, MS 38677

The Mississippi Center for Supercomputing Research was established in 1987 by the Mississippi Legislature and the Institutions of Higher Learning (IHL) in order to provide high performance supercomputing (HPC) support for research and instruction at all state universities. The Mississippi Supercomputer User Advisory Committee (MSUAG) was established by the IHL Research Consortium to provide user input and advice to MCSR management and technical staff on policies and procedures for the Center's operations. It includes member representatives from all IHL institutions. The Advisory Group will meet at this MAS conference. Mr. David G. Roach, Director of the MCSR, will conduct the meeting. The agenda includes an update on MCSR HPC facilities and services, introduction of new MCSR staff members, and site reports and ongoing research updates by MSUAG representatives. A poster session, showcasing research projects that utilize MCSR facilities and services, will follow the Advisory Group Meeting. A Special HPC Subsession of the Mathematics, Computer Science, and Statistics Division, sponsored by the MCSR, will also be held to serve as a forum on supercomputing in which faculty and graduate student researchers will have the opportunity to describe their research projects that involve HPC, Internet2, Grid Computing, Visualization, Network Security, Computer Systems Administration, and the use of MCSR resources. IHL faculty and graduate students, with an interest in HPC and/or MCSR facilities and services, are also invited to attend and participate.