Trident Scholar Abstracts 2003

Just as their precursors have for over one hundred years, the submariners of today’s Navy operate and perform diligently in an environment where they are not physically able to see their surroundings or the possible adversaries with whom they might be fighting. Far from considering themselves blind, these silent warriors have at their fingertips the capability to visualize everything around their ship via the advanced acoustic images provided by modern sonar systems.

Proper assessment of this information demands a sharply honed skill in recognizing subtle yet crucial details occurring at both large and small scales within the “waterfall” display screen viewed by the sonar operator. Thus, the principal goal of every submariner is to detect, classify, and track every contact as soon as possible with the greatest accuracy.

This Trident Scholar project seeks to apply the theory of wavelet based multi-resolutional optic flow to the specific problem of finding a more efficient way to analyze the time varying one-dimensional acoustic imagery supplied by sonar systems. It must be noted that previous work relating to optic flow has concentrated exclusively on the applications of two-dimensional visual image sequences. No research has yet been carried out to explore the intricacies of the theory in the one-dimensional context of sonar.

The study will mainly be focused upon uniting the sizeable and highly complicated notions of: wavelets as an instrument to distinguish changes occurring within the images at both coarse and fine levels of resolution; optic flow as a mechanism to describe the movement of distinct objects in an image sequence; and the reduction of the partial differential equations which govern optic flow into a system of solvable ordinary differential equations. To date, no effort has been made to harness these valuable ideas for acoustic investigation or even to consolidate them within the framework of a numerical interface which is compatible with current Navy users and systems such as MATLAB.

Benefits of such research applied in a context relevant to submariners might lead to faster detections and identifications of moving contacts as well as the ability to focus more clearly on changes in contact information amidst high levels of acoustic interference. Although the project may not create immediate tangible advances on our submarines, it will set the stage perfectly for the next generation of refinements for our real-time sonar systems.

This Trident Scholar project is an investigation of the nuclear structure of Tellurium-130. The purpose of nuclear structure research is to understand the features of the nuclear force that determine the balance between the various ways a nucleus can behavior. Theoretical model calculations are compared to experimental data to understand which models work better under which circumstances.

The tellurium nuclei have two valence protons with respect to a closed proton shell and a range of neutron numbers. Three different type of nuclear structure behavior are thought to be active in these nuclei: collective, two-particle, and particle-hole excitations known as intruders. Because there are seven stable even-even Te nuclei, one can study the evolution of these excitation modes over a wide range in neutron number, allowing for good quantitative interpretation of the nuclear structure.

For this project, experimental data have been taken at the University of Kentucky Nuclear Structure Laboratory using a technique called inelastic neutron scattering. By scattering neutrons off of Te-130, the nucleus can be excited, and the resulting de-excitation gamma rays recorded as the Te-130 excited states relax.  Through various analytical techniques, the energy level scheme will be constructed, and spectroscopic information such as lifetimes, level spins and parities, and decay branching ratios.  The behavior of the ¹³⁰Te nucleus will be examined from the viewpoints of the Interacting Boson Model-2, the General Collective Model, and the Particle-Vibrational Model.  Emphasis centers on understanding the interplay between particle and collective features and on the aspects of the nuclear forces and shell model orbitals that determine the relative importance of each model.  Often level energies are more important for evaluating whether a model has sufficient built-in complexity rather than distinguishing between physical descriptions.  Lifetime information is crucial to revealing the amplitudes of particle and collective components in the wavefunctions.  Results from these ¹³⁰Te studies will be combined with previous information on the other lighter tellurium nuclei ¹²⁰Te, ¹²²Te, ¹²⁴Te, ¹²⁶Te, and ¹²⁸Te to study the evolution of these structures across an isotopic chain.

Research and development of new platforms such as military aircraft is often costly and time-consuming. Therefore, it is important that platforms have as long of a service lifetime as possible. However, exposure to the environmental elements, particularly saltwater, is especially troublesome to the Naval aircraft because it shortens the lifetime of its aircraft.

The P-3C is an example of a versatile aircraft whose lifetime has been extended to the point that corrosion is becoming a concern. Structural components in the P-3C aircraft are currently composed of aluminum alloy AA7075-T6, which has high strength but limited corrosion resistance. In particular, AA7075-T6 is susceptible to a form of corrosion called environmentally assisted cracking (EAC), which can cause unexpected fracture of a loaded part. In 1999, Lockheed Martin Aeronautical Systems was granted funding to complete the Service Life Assessment Program for the P-3. One objective of this program was to identify a possible replacement for corrosion-prone AA7075-T6 components that would help reach the service goal life of 2015. There are two aluminum alloys that are currently under development as possible replacements for the aging AA7075-T6 parts: AA7249 and AA7150.

USNA was provided with an AA7249 wide panel extrusion that is identical to the AA7075-T6 extrusion that forms part of the wing of the P-3C aircraft. The objective of this Trident project is to extend the ongoing evaluation of AA7249 extrusions to include additional corrosion testing, mechanical testing, and evaluations of the effects of processing on the wide panel extrusion, as well as additional AA7249 extrusions provided to USNA by industry. The complete study is executed in three phases: (1) evaluation of local properties such as hardness, strength, and electrical conductivity as a function of position within the extrusion, (2) evaluation of the microstructure (i.e., grain size, shape, and orientation) as a function of position within the AA7249 extrusions and (3) characterization and quantification of the EAC resistance as a function of the grain structure via a slow strain rate technique. This information will contribute to the ongoing evaluation of these alloys for replacement of AA7075-T6 in aerospace structures and contribute to a better basic understanding of the grain structure obtained in these parts and the effect of processing on EAC behavior.

This research focuses on the development of an optical calibration system in a thermo-luminescent dosimeter (TLD) reader. A reference light source, is used to calibrate the reader. Its output power must remain stable during system operation.

The Navy uses TLDs on all nuclear warships. Thermo-luminescent detectors measure the radiation dose a person receives. Inside of the TLD is a small crystal. Electrons in the crystal are excited to a higher energy state when they absorb radiation. This crystal is later taken out of the TLD. By heating the crystal it will emit light energy in proportion to the radiation dose. The TLD reader measures this energy to determine the radiation dose.

The TLD reader must be calibrated quite often to ensure it is getting accurate measurements. The problem is that currently the light source does not maintain a stable output over time. Bicron, the manufacturer of the readers, has implemented a radioactive isotope, rather than a LED, into one of its readers to try to resolve this problem, but this has not shown much improvement. This solution is not preferred because of the many restrictions placed on the movement of radioactive material. The problem could be solved if a LED or laser were able to have a stable output power.

Thus far testing shows that the LED used is too temperature sensitive to meet the stability demands for output power. With this in mind, this project will have four key parts. The first is to develop and verify the power stability of an LED. This power must be stable to within 1 % over an entire month. The second part is to integrate a data acquisition system that will record the optical power measurements. The project must then be packaged as a product that can be implemented in the Navy’s TLD readers. Finally, environmental testing and evaluation will be conducted to test the reliability and stability of proposed design.

The goal of this project is to resolve the physical properties of a sample of Near-Earth Objects from optical and infrared ground-based observations made at Kitt Peak and Mt. Lemmon, Arizona. The Trident Scholar for this project, MIDN 1/C Luke Dundon, spent two weeks in September 2002 working with professional astronomers to obtain these data on his sample of Near-Earth Objects. The telescopes used on Kitt Peak included the KPNO 84 inch telescope with the Simultaneous Quad Infrared Imaging Device (SQIID) infrared camera, the 90 inch optical telescope owned by the University of Arizona, as well as the 72 inch Spacewatch optical telescope. At Mt. Lemmon he used the 60 inch infrared photometer telescope for making further infrared measurements. He used these instruments for three-color photometry, as well as for lightcurves in the optical and infrared. (NOTE: The difference between a photometer-based telescope and the others was that the photometer only gives a numerical readout of the intensity of the radiation collected, while the other cameras produced an image of the part of the sky observed.)

These data will be used to constrain the physical properties of the sample. For each object, the light curve will be used to determine rotational properties. The taxonomic classification will be determined from three-color photometry and from calculating albedo where possible.

Near-Earth Objects are asteroids and comets on orbits that bring them close to Earth. They provide important information about the evolution of the solar system. They pose a threat to Earth. A collision with a large Near Earth Object would destroy the biosphere of the planet. In the future, as space exploration advances, they will be invaluable for mining of natural resources. Unfortunately, these objects have not been widely studied. This project will add to the body of knowledge of Near Earth Objects.

During amphibious operations, the movement of Marines from their host Navy ships the area of operations is extremely critical to mission accomplishment. The rapid and reliable dissemination of information is paramount. Both conventional radios and existing digital communications are insufficient to this task. Mobile ad hoc networking (MANET) is a potential solution to this problem. The Office of Naval Research (ONR) has sponsored projects to provide connectivity to Marines on the ground within a wide area relayed network. However, a host of technical challenges continue to impede the implementation of reliable and seamless networking.

MANET routing protocols determine the best path that will allow data to travel throughout the network. The focus of past research has been developing these protocols for a ground environment. In a littoral environment, communication takes place while time-varying interference is present. The presence of this interference affects the quality and status of the links between nodes. Current routing protocols clearly define links as either functioning or non-functioning, but time-varying noise can cause links to be temporarily non-functioning. In such a case, the discovery of new routing path may not be necessary, and may actually congest the network unnecessarily. The fundamental factors in this aspect of routing protocol design include how a routing protocol deals with the discovery of neighboring nodes and classifies links in the network. The investigation of the effects of these factors in a routing protocol’s performance will be the basis for creating an improved routing protocol optimized for littoral operations.

Three existing MANET protocols: Ad Hoc on Demand Distance Vector (AODV), Dynamic Source Routing (DSR), Optimized Link State Routing (OLSR) are being investigated. AODV is being used as the baseline routing protocol to be modified because it already contains the option of configuring the type of neighbor discovery techniques. The effectiveness of the changes to these protocols are being evaluated using ns-2, an open-source network simulation tools. Simulation results are providing the basis for further development of existing protocols for both simulation and possible implementation using YP’s or Navy 44' as a test platform.

The final product in this project will be a protocol optimized for implementation in a littoral environment which will improve communications during troop transport: making adaptation to information and obstacles faster, increasing the coordination and effectiveness of amphibious landings and ultimately improving the success of the overall amphibious operation.

The goal of this Trident Scholar project is to use real time GPS data from a satellite to model atmospheric density throughout its orbit. The basic premise is that the difference between where a satellite is expected to be and where it actually is can be attributed to a number of factors. When all of these factors are accounted for, the remaining difference must be due to atmospheric drag. Atmospheric density at a point in space is directly related to the atmospheric drag experienced by a satellite at that point.

PC-Sat is a Naval Academy designed and built satellite, launched in September, 2001 from Kodiak, Alaska. The GPS receiver on board, built by DLR, Germany’s equivalent of NASA, was specifically designed for space use. It relayed 12 days of continuous position data in January, 2002 which will be used to calculate the atmospheric density. The data collected can be verified by comparing with data taken by NORAD (North American Aerospace Defense Command), which tracks everything orbiting the Earth.

In 2000-2001, John Young developed an algorithm to calculate atmospheric density using simulated data as a Trident project. He assumed, among other things, that the earth’s atmosphere rotated with the Earth. Mr. Young concluded that this assumption created a significant amount of error in the calculations. Therefore the current Trident Scholar project aims to improve the algorithm designed by Young and to use real data for analysis.

Currently the best model of atmospheric density is a function of altitude and neglects many factors. As the collection of data grows from various satellites, scientists can begin to develop a more accurate model of atmospheric density based on instantaneous information from a number of satellites. This project will contribute to the scientific community by adding to the database of ideas and information. It also has the potential to help focus the small satellite program here at the Naval Academy.

The goals of this Trident Scholar project are to develop a scale-model vehicle that is dynamically similar to a full size automobile, to design a driver assistance control system to help prevent accidents during an emergency turning maneuver, and to use the dynamically similar scale-model vehicle to test the control system.

The ability to compare the scale-model vehicle to actual automobiles requires that the two systems are dynamically similar. In other words, the scale-model and an actual automobile have identical reactions but on different scales. Dynamic similitude can be shown using the Buckingham-Pi Theorem by replacing the dimensional physical parameters with dimensionless products and ratios. These ratios are formed using basic units such as length, time, and mass. Dimensionless groups, known as Pi groups are formed from the ratios of physical parameters. Two systems are dynamically similar if the corresponding Pi groups are equal.

For this project, five Pi groups have been determined from eight parameters that govern a vehicle’s lateral position. In order to produce a dynamically similar scale vehicle, the five Pi groups are computed for the unmodified scale vehicle and compared to Pi groups for actual automobiles. Any differences between the Pi groups of the scale vehicle and Pi groups of actual automobiles can be resolved by modifying the scale vehicle.

Verification of the dimensional analysis is important because it may not be possible to match all the Pi group simultaneously. To verify the dynamic similitude between the scale vehicle and actual automobiles, the scale vehicle will perform a lane change maneuver on the test bed. There has been considerable research on this maneuver and the performance of the scale vehicle will be compared to published data obtained from automobiles performing the same maneuver.

Once the scale vehicle is shown to be dynamically similar and the results are verified, a control system will be designed to alter a driver’s steering input. This control system will prevent the driver from steering the automobile too aggressively and causing the vehicle to roll over or spinout. The control system will be implemented in the control system of the scale vehicle and an emergency turning situation will be simulated using the test bed apparatus.

A Bluetooth based “power node” is being developed for monitoring and controlling power systems onboard US Navy Vessels. For this application a “power node” is defined as an electronic system, which collects information from many sensors and makes appropriate control decisions based on the occurrence of well-defined events. Bluetooth is a low cost, low power wireless standard, which is incorporated on each “power node.” This wireless standard allows networking of several “power nodes.” An important advantage of this system is that it can be configured for many shipboard applications.

The Bluetooth standard uses a spread-spectrum modulation scheme that allows reliable communication between “power nodes” within several sub-networks (piconets) in the same physical location. We are striving to design a robust wireless network that maintains reliable connectivity among nodes even when the communication channels are altered by the opening and closing of watertight doors. Versatility is achieved by the use of the Motorola MC68HC908JB8 microcontroller in the “power node.”  This microcontroller is in-circuit-programmable, allowing rapid software changes. The goal of this project is to have a working prototype network consisting of several “power nodes” to be tested on the ex-USS America.

The September 11^th bombing of the World Trade Centers in New York illustrated the many problems associated with rescuing the survivors of a collapsed building. The fact that those trapped beneath the rubble only have a couple days to live combined with the massive amount of debris that need to be searched makes the urban search-and-rescue mission extremely daunting. In addition, the environment within a pile of debris is very unstable and often too dangerous for entry by humans. The solution to such a complicated problem lies in robots capable of quickly exploring a collapsed building and pinpointing the location of any survivors.

The overall goal of this Trident Scholar project is to build an urban search-and-rescue robot. The primary goal is to develop a physical structure that will be unique and versatile enough to traverse different terrain challenges. A design's effectiveness will be judged on its ability to overcome the pre-selected terrain types. The secondary goal is to simplify operator input by establishing a 'fly-by-wire' style controller, wherein user inputs will be mapped to motor motion commands that will be developed using a physics simulator. Initially, a simulation model of the robot needs to be created with responses that closely correlate to the responses of the actual prototype. Using the simulation, various methods, including genetic algorithms, will be used to develop locomotive methods, or 'gaits,' for various types of motion, i.e., lines and curves. The movements will then be translated to the physical prototype and their effectiveness (the ability to move as predicted) analyzed. The end result will be a robot with a versatile mobility controlled using simple operator inputs.

This Trident Scholar project involves an investigation into applying capillary pumped loops to cool shipboard electronics. A capillary pumped loop (CPL) consists of an evaporator plate and a condenser connected by piping into a loop. The system is driven by the pumping head produced through capillary action by a wick in the evaporator plate and thus operates without the assistance of any pumps. Inside the loop, the working fluid, water, will evaporate inside the evaporator plate instead of being heated sensibly by the heat produced from the simulated electronics package. Current electronic cooling systems employ forced air convection, which can only produce a maximum heat flux removal of 3.39 W/cm². Many electronic systems, processors, RF devices and packages being considered by the Navy and the civilian world are going to soon exceed this limit within the next two years.

This investigation seeks to use a vertical flat plate evaporator to cool simulated electronics packages and the hot spots created by these packages which are mounted to the evaporator plate’s surface. The goals of this investigation include designing a vertical flat plate evaporator to achieve at least 10 W/ cm² of heat removal, and testing its performance in a series of experiments designed to help prove the feasibly for CPL technology to be applied to Naval vessels at sea. Some of the parameters to be studied include minimizing the effect of subcooling on the liquid return line by insulating the cold plate, piping, and evaporator plate and varying the flow rate through the cold plate from 5 to 0 gallons per minute, testing the performance of the vertical flat plate evaporator under a variety of pitch and tilt orientations up to forty-five degrees to show that the pumping head produced in the wick can overcome the pitch and roll of a ship and sea, and finally to ascertain the maximum capillary pumping head produced by the evaporator plate by closing a needle valve to increase flow losses. Subcooling occurs in the condenser and as the vapor is condensed into a liquid; for optimum CPL performance, the temperature of the liquid should remain near the saturation point and not drop to far below this point so that most of the heat removed is utilized by latent heat transfer rather than by sensible heat transfer. The maximum pumping head produced by the evaporator plate will give engineers information on how far apart the condenser and the evaporator plate can be for the loop to still function.

The vertical flat plate evaporator is designed to allow for heat removal from both sides and it is designed to fit in a standard Navy commercial off the shelf (COTS) electronics cabinet. This investigation represents the first attempt at designing a CPL evaporator plate for the shipboard use of cooling electronic components. A successful vertical flat plate design will then be considered for installation into COTS electronics cabinets because it is much more efficient than air cooled systems currently installed in Navy ships and submarines.

The recent development of lightweight, ultra-strong, high endurance composites has opened a world of possibilities in the field of aircraft design. Yet, with all their advantages, the full nature of composite structures is not truly known. For instance, it has been shown that transverse shear deformation tends to have a much greater impact on the characteristics of composite wings, than on wings composed of classical materials. However, current studies taking this fact into consideration are limited to low-speed, incompressible flow environments. Modern high performance aircraft rarely operate in these environments. In order to accurately determine the effects of transverse shear deformation on composite wings in high-speed, compressible flow, a new model must be created.

The purpose of this research will be to develop that model for a generic wing. This model can then be used to solve for the aeroelastic properties of a refined wing to include flutter, divergence speed, and flutter mode shapes. These properties are all crucial in determining the characteristics and safety of a given wing. In the classical model, it is assumed that the transverse shear flexibility of the wing is zero due to extremely high values of the modulus of rigidity. This allowed for transverse shear effects to be ignored. However, it has been shown that in the case of composites, which have a much smaller modulus of rigidity, transverse shear effects cannot be ignored, often having an impact of up to fifty percent on vital the wing characteristics.

The derivation of this model begins with the three dimensional displacement equations for a generic wing. Once these are determined, the boundary conditions necessary to solve the displacement equations can be developed using Hamilton’s variational principle, which is a statement of conservation of energy. Inserting in the coefficients of rigidity and the aerodynamic forces gives the governing equations. Finally, using the method of Laplace transforms, the governing equations can be solved, resulting in a matrix, which incorporates all the geometric, aerodynamic, and material properties of the wing.

It is assumed that the effect of transverse shear deformation on composite wings in high-speed, compressible flow will be significant. This project will not only verify that hypothesis, but also give an accurate statement of just how significant those effects will be. This knowledge will allow engineers to alter their design with a greater awareness of what is causing the problem rather than just having to guess and then check their work.

This Trident Scholar project involves the investigation of modular robots that use only magnetic forces to self-reconfigure and maintain a network of independent robots. Modular robots are usually identical electro-mechanical devices. Alone, each module has limited functionality. As a group or network, however, they can form almost anything, i.e. using one type of “Lego block” to build a larger structure. In this project, each module is independent. This means that everything that it requires to function is onboard, including power and information processing. In addition, this project will use modules that latch and self-reconfigure autonomously using permanent magnets and momentary electro-magnets.

The goals of this Trident Scholar project are to determine the effectiveness of a network of that utilizes prototype magnetically connected and actuated modules; to develop a self-reconfiguration algorithm; and to extrapolate the fundamental principles discovered on to large scale prototypes to networks of numerous small scale modules.

This project contains two major phases. First, a generic module design is to be developed so that several modules are able to self-reconfigure in a network. This design will not be optimized due to time constraints, but will provide observable effectiveness results, such as reaction of outside forces and reconfiguration speed. The second phase will involve a generic program for each module to enable the network to autonomously self-reconfigure.

It is expected that the research will find the large scale prototype modules to be less effective in latching and reconfiguring than mechanically linked and actuated modules. However, if this effectiveness is related to a micro scale, as is one of the goals of the research, it is expected that magnets will produce better network attributes. The project is expected to successfully perform several network reconfigurations that can be achieved using a self-reconfiguration algorithm using minimal time and space knowledge. Finally, as stated, the research expects to yield a successful modular self-reconfigurable network whose principles are easily and readily extrapolated to networks of thousands of modules on a micro scale.

Modular robots have the potential to be cheap, highly fault tolerant, and extremely adaptable, thus opening up the possibility of almost unbounded applications. This Trident Research is part of the current research effort to study modular robotics. It aims to generate novel information about modular network design and reconfiguration, adding to the body of knowledge of usable modular ideas. It is hoped that such research can later be applied to such diverse areas as medicine, space exploration, and search and rescue.

The goal of this Trident Scholar project is to compare tank testing and computational fluid dynamics in their uses to naval architecture. The traditional process of ship performance prediction is hydrodynamic tank testing where a scale model of the ship is towed to measure the forces scalable to full size. Each trial includes different conditions such as speed, rudder angle, yaw angle, or heel angle. The results can then be scaled up for the actual ship. However, proposed design changes are expensive and time-consuming to analyze as they require new models.

Since the advent of the computer, the use of Computational Fluid Dynamics (CFD) is available for ship performance prediction. CFD takes a digitized ship hull, places it onto water panels of a virtual sea, and calculates the water pressure on each part of the hull. This provides lift and drag predictions. In CFD, there is no problem with modeling similarity. Design changes are relatively easy and inexpensive to evaluate, but large computing resource requirements force numerical simplifications in the code. Nonetheless, CFD is approaching the accuracy of the tank.

The project will compare ship performance prediction methods using a simple PC-based CFD code developed by engineers at NSWC-Carderock versus a complex mainframe code at Patuxent River NAS. These results will be compared to tank test data developed in the USNA Hydrodynamics Laboratory. The new Navy 44 Sail Training Craft (STC) will serve as the baseline design due to its design complexity and intended future service with the Naval Academy. Differences in each set of results will be explored in areas such as lifting surfaces, leeway, free surface effects, and turbulence. To test the use of CFD as a design tool, a new rudder will be developed. The significance of this project is that it may be proven that the inexpensive CFD is a valuable tool for naval architecture in the design of ships. Furthermore, using a velocity prediction program, an accurate prediction of the sailing characteristics of the boat will be calculated, allowing for design improvements which will potentially benefit the Command Seamanship and Navigation Training Squadron and Varsity Offshore Sailing Team.

This Trident Scholar project involves an investigation of the influence unmanned-aircraft characteristics have on landing precision. The goal of this analysis is the determination of a minimum set of performance and maneuverability criteria that ensure satisfactory carrier landing precision and consistency. The first step in this project is to design and write a comprehensive computer model to simulate unmanned carrier landings. The model will include ship dynamics, atmospheric turbulence, navigational information errors, dynamic airplane control structure, and aircraft dynamics. The airplane senses and correct for three disturbances: wind gusts, unsteady ship motion, and navigational information errors. The second step involves varying aircraft attributes and recording landing performance statistics for each configuration. Some examples of aircraft attributes are elevator effectiveness, roll control power, throttle lag, and center of gravity location. The influence of each attribute will be extracted from the landing statistics. This will be done for various weather conditions because the ship motion increases with wave size and the gust intensity increases with average wind velocity. The third step is the development of preliminary criteria to be refined and verified with test cases. The fourth and final step is to generate a report presenting the above process and the final aircraft design limits. The required aircraft performance is expected to be relaxed from the standards for manned aircraft. The significance of the anticipated performance requirements is reduced airplane cost and increased design freedom.

This Trident Project investigates academic scheduling problems at the U.S. Naval Academy. The focus is on devising methods to construct good final exam schedules. A method to improve existing course schedules and facilitate course changes is also proposed.

Several possible techniques to treat the final exam problem that will be examined include: the Descent Method, Simulated Annealing, partial search methods, genetic algorithms, and graph theoretic approaches. The Descent Method is a discrete analogue of hill-climbing techniques from multivariable calculus. This technique searches a set of schedules by exhaustively analyzing schedules in local neighborhoods along a search path. Simulated Annealing is an alternative method inspired by the annealing process in chemistry. By using probabilistic techniques, this method allows a wider search of the space of schedules. Partial search techniques seek to build a schedule day-by-day but backtrack if the partially completed schedule is already worse than previously completed schedules. This methodology, known as the branch and bound method, is a standard paradigm in Artificial Intelligence. Genetic algorithms involve mating “parent” schedules to form favorable “offspring” schedules and then subjecting these new schedules to local mutation. Finally, there are also graph theoretic approaches to the exam-scheduling problem. One promising approach involves the computation of the chromatic number of a particular graph created from the course information. These methods give different heuristics that can be used to tackle the exam-scheduling problem.

The exam-scheduling problem is an example of an NP-hard problem – a class of problems for which no fast algorithms are known. After further theoretical and practical investigation, one of the methods discussed above will be implemented in a computer program. Our goal is to improve on the method of exam scheduling currently in use at the Naval Academy.

The last project facilitates section changes at the Academy. This is currently done on an ad-hoc basis, but could be streamlined by using a centralized barter system. This barter technique would accept input listing desired section changes and would identify multi-student section changes that lead to better class schedules. A prototype computer program to find such sections changes will be devised.