Past Projects

Over the years, the AML is completed more than 20 research projects. Below you can find the complete list, with brief descriptions.


2016–2017: Forecasting Natural Gas Demand
Sponsor: Constellation Energy

Description.  The AML team used advanced statistical methods to create a model for forecasting demand for natural gas in the residential/comercial and industrial markets. 
2015–2016: Particle Swarm Optimization Method for Automated Model Baselining
Sponsor: RTR Technologies, LLC

Description.  The AML team used a particle swarm optimization method to develop an algorithm to predict the ideal number of checkpoints needed to process passengers from international flights. 
2011–2013: Mathematical Modeling of Segregation Patterns in Urban Areas
Sponsor: US Department of Housing and Urban Development

Description.  The AML developed mathematical models of multi-group residential segregation in urban areas. The racial distribution of neighborhoods in urban areas has been front and center of the wider debate about racial disparity. The teams tackled this problem with probabilistic tools using Markov chain models to analyze and study the evolution of residential segregation at the census tract level.
2009–2011: Risk Analysis: Toxic Materials Transportation Security
Sponsor: Chemical Security Analysis Center (CSAC)

Description.  The AML worked with CSAC, a unit of the US Department of Homeland Security, to design a mathematical model to assess and manage the risk to the population involved in transporting toxic chemicals and to evaluate the available tools for solving the resulting mathematical problem. The student team adapted a minimum cost network flow model with randomized cost coefficients. This work was supported by the US DHS through the grant TCN 10-030.
Dennis Howell, Alexei Kolesnikov, Angel Kumchev, Patrick O’Neill and Matthew Tiger. Estimation of the commodity flow of chlorine from storage data. Journal of Transportation Security 5 (2012), 51–68.
2008–2009: A Study of the Deer Population in Baltimore County
Sponsor: Baltimore County Government

Description.  The AML worked with the Baltimore County Department of Environmental Protection and Resource Management and with the Towson University Environmental Science Program to estimate the deer population in Baltimore County and to model the population in the future.
2005–2008: The Mathematics of Geographic Profiling
Sponsor: National Institute of Justice

Description. The AML worked with the National Institute of Justice to determine the optimal police search area for a serial criminal. This is the question of how, given the location of the crimes committed by a single offender, to determine an optimal search area for that offender's home base. This work was supported by the National Institute of Justice through grants 2007-DE-BX-K005 and 2005-IJ-CX-K036. 
M. O'Leary. The Mathematics of Geographic Profiling. Journal of Investigative Psychology and Offender Profiling 6 (2009), 253–265.
2004–2005: Analysis of the Carroll Area Transit System
Sponsor: Carroll Area Transit Systems

Description. The AML worked with Carroll Area Transit Systems to provide statistical models that can be used to predict the optimal number of vehicles needed for the system. The team also analyzed the current operations of the system, and compared different bus scheduling algorithms. The team's results were published in the following article.
G. Han and J. Zimmerman. Predicting Demand for the Carroll Area Transit System. UMAP Journal 27.1 (2006).
2003–2005: Rural Baltimore County Domestic Well Analysis
Sponsor: Baltimore County Government

Description.  Two AML teams analysed data provided by the Baltimore County Department of Environmental Protection and Resource Management to look for patterns in the failure rates of rural domestic water wells. The teams' results were published in the following article.
K. Koepnick and X. Wang. Logistic Regression Analysis of Well Failures in Baltimore County. Journal of Data Science 7 (2009), 111-127.
2002–2003: An Analysis of the Optimal Staffing Level of the Baltimore City Fire Department  
Sponsor: Baltimore City Fire Department

Description.  This project determined an approximately optimal staffing level for the Baltimore City Fire Department. The team's results were published in the following article. 
A. Engel and C.L. May. Optimal Staffing at the Baltimore City Fire Department. UMAP Journal 26.1 (2005).
1998–1999: Validation and Enhancement of Applications of Models from Epidemiology to INFOSEC Assurance Metrics
Sponsor:  Science Applications International Corporation (SAIC)

Description.  This project investigated the use of epidemiological models to simulate and analyze the spread of computer viruses through a corporate network. The team's results were published in the following article.
J.L. Aron, R.A. Gove, M. O'Leary, S. Azadegan, and M.C. Schneider. The Benefits of a Notification Process in Addressing the Worsening Computer Virus Problem: Results of a Survey and a Simulation Model. Computers and Security 21 (2002), 142-163.
1996–1997: Customer Usage Profile for Fast-Packet Frame Relay  
Sponsor:  Bell Atlantic

Description.  The AML team developed customer usage profiles for the relay of fast packets through the communications network. Data communications engineering and queueing theory were used to develop models for projection of network impact, processor occupancy, and link utilization in the network serving Bell Atlantic's industrial and business clients.
1995–1996: A Computer-Adaptive Mathematics Placement Test
Sponsor:  Towson State University

Description.  The AML adapted a computer-based mathematics placement test from the nationally recognized series of tests available from the Mathematical Association of America. The AML test is computer-adaptive and uses information about the student's background to customize the level of the test and adjusts as the student takes the examination. Students are advised on placement in accordance with their skill level and their major.
1993–1994: Interface Between a Database System and Statistical Software
Sponsor:  Martin Marietta Aero and Naval Systems

Description. The AML team wrote a front and back end to a statistical package linked at the laboratory bench to a materials database used in the production of state-of-the-art marine and air craft.
1990–1992: Scheduling Production of Prepared Plated Media
Sponsor:  Becton Dickinson Microbiology Systems

Description. Plates treated with growth media used for cultivation of bacteria used in disease diagnosis are produced on several production lines at the Becton Dickinson Hunt Valley installation. The scheduling package designed by the AML optimizes the weekly schedule and takes account of inventory, quarantines, cleanup and setup times on three of the busiest lines.
1989–1990: An Enrollment Model for Resource Scheduling
Sponsor:  Towson State University

Description. An AML team developed software to help the registrar's office schedule courses based on the changing demands within majors and in the requirements for degrees.
1988–1989: Statistical Survey of Long-Term Care Patients
Sponsor:  BlueCross BlueShield of Maryland

Description. Towson actuarial science students and other mathematics majors investigated the length-of-stay of Marylanders in nursing homes for the purposes of helping the sponsor to evaluate the marketing and pricing strategy for long-term care insurance policies.
1986–1987: Estimation of Sales Tax Liability
Sponsor:  State of Maryland, Comptrollers Office

Description. Calculation of the liability of sales tax to be paid to the State of Maryland by retail companies doing business in the state had been obtained by teams of auditors combing through all of the business' records. The AML team wrote software to allow for just one or two auditors to select a stratified sample to estimate the tax liability with a high degree of accuracy.
1985–1986: Customer Service Queuing Model
Sponsor:  Citicorp of Maryland

Description. The Citicorp Choice credit card service desk had varying needs for operators and telephone lines. By setting up a real-time scheduling device, the AML team helped Citicorp meet its service goals on the help-line.
1984: Acoustical Pattern Recognition
Sponsor:  AAI Corporation

Description. Students in physics and mathematics joined together to study the acoustical signatures of several vehicles in an attempt to design a remote recognition device.
1983–1984: Collection of Cancelled Checks for Processing
Sponsor:  Union Trust Company of Maryland

Description. Union Trust had 85 branch offices in all parts of Maryland. The AML team developed an algorithm for the collection of cancelled checks from the branch offices for processing at the central Guilford Avenue center. The system was written to take into account the volume of check activity at branch offices, the monetary value of the checks, and the efficiency of routes.
1982–1983: Optimal Test Station Loading
Sponsor:  Westinghouse Electrical Corporation

Description. Westinghouse's Electronics Repair Center serviced parts for airplanes used by the Armed Forces. The AML team developed a scheduling program to help the service center minimize the time that a part is in the center for evaluation, repair and inspection.
1980–1981: Student Housing at Towson State University
Sponsor:  Towson State University

Description. The AML team investigated demographic and enrollment trends and projected the on-campus residential needs for the University. The projections were used in presenting the funding package to the state legislature.