David J. Phillips
Mathematics Department
United States Naval Academy
Annapolis, MD 21401
410.293.6710/646.369.7819
dphillip@usna.edu, dphil73@gmail.com
http://www.usna.edu/Users/math/∼dphillip
EDUCATION
Ph.D. Columbia University. New York, NY. February 2007. Operations Research.
“Approximation algorithms and techniques in theory and practice”
Advisors: Garud Iyengar and Clifford Stein
M.Phil. Columbia University. New York, NY. May 2006. Operations Research.
M.S. Columbia University. New York, NY. February 2003. Operations Research.
B.A. with Honors Oberlin College. Oberlin, OH. May 1997. Mathematics.
ACADEMIC POSITIONS
Assistant Professor. Mathematics Department. United States Naval Academy. August 2011present.
Assistant Professor. Mathematics Department. College of William & Mary. August 2006-August
2011. on leave from August 2007-June 2008, Fall 2010.
Visiting Scholar. Computer Science Department. Brown University. Fall 2010.
Visiting Assistant Professor. Industrial Engineering and Operations Research Department,
Columbia University. August 2007-June 2008.
PROFESSIONAL EXPERIENCE
PA Consulting
February 1999 - June 2000. Analyst. Implemented and documented statistical and operations
research models using programming languages and applications including C/C++, Fortran,
Visual Basic, SAS, MatLab, and Excel. Some projects include:
• Lead programmer on Project Spreader, an energy plant valuation model incorporating Monte Carlo simulation, dynamic programming, linear programming, and finance
spread option modeling techniques.
• Lead programmer for the Generation Reliability model, which determines the
Loss-of-Load Expectation (LOLE) of a given set of power plants subject to planned
outage constraints. Program uses a greedy algorithm to schedule planned outages and
convolves forced outage probabilities to calculate LOLE.
The Lewin Group
August 1997 - February 1999. Research Assistant. Analyzed several data sets and implemented
microsimulation, regression analysis, and operations research models using programming languages and applications including Fortran, Visual Basic, SAS, and Excel. Some projects
include:
• Wrote and modified Fortran code implementing a microsimulation which forecasts health
care expenditures among Medicare recipients using the Medicare Current Beneficiary
Survey data-set. Wrote SAS code used to estimate coefficients used in the simulation.
• Implemented microsimulation in SAS analyzing impact of the 1997 Balanced Budget
Amendment upon Home-care agencies.
• Developed a software guideline in Visual Basic language to assist doctors in diagnosing
and treating community-acquired pneumonia.
GRANTS AND FELLOWSHIPS
NARC grant. Principal Investigator. Summer 2012, Summer 2013.
ONR grant. Principal Investigator. June 1, 2014–December 31, 2014. In review of my NARC grant
proposal, ONR choose to fund my grant and included additional support funds.
NSF DMS CSUMS. Theory, techniques, and research in computational mathematics. Senior faculty.
$668,775. Fall 2007-Summer 2011. Funding received while at William & Mary. The purpose of
this grant was to support undergraduate research into computational mathematics research. I plan
on using my experience with this grant to develop a similar program for undergraduate research at
USNA.
NASA/VSGC New Investigator. Principal Investigator. Summer 2009-Summer 2010. $20,000.
Funding received while at William & Mary.
REFEREED PUBLICATIONS
author order is alphabetical
13. Algorithmic Graph Theory and Sage, with D. Joyner, 305 pages.
• This text introduces graph theory and graph algorithms and describes implementations
of these algorithms in Sage, a Python library. The primary audience of the book is
mathematics, operations research, and computer science undergraduates.
• Cambridge University Press has approved the book subject to contract.
• Contribution: I have written three chapters for this book (approximately 70 pages) and
have written approximately 30 problems for 4 different chapters.
12. “Graph theoretic analysis of structural connectivity across the spectrum of Alzheimer’s disease”, with L. Jager, A. McGlaughlin, D. Ruth, and A. Soldan, Accepted to NeuroImage:Clinical.
45 pages.
• In this work, we use techniques from graph theory to help analyze magnetic resonance
images. In particular, we seek to differentiate images taken from healthy subjects from
images taken from subjects that have been diagnosed with dementia.
• Contribution: Professor Soldan and I are joint principal investigators of this work and
I am the first author of the paper. The majority of statistical expertise was provided
by CDR Ruth and some statistical support was provided by Professor Jager, although
the statistical code was written by ENS McGlaughlin and myself. This paper was based
on the Trident Project completed by ENS McGlaughlin who was primarily advised by
myself.
11. “On the Hadamard transform of monotone Boolean functions”, joint work with C. Celerier,
D. Joyner, and C. Melles. Tbilisi Mathematical Journal. Vol. 5(2) (2012), pp. 19-35.
• In this paper, we investigate and classify the Hadamard transforms of so-called bent
functions using algebraic and graph theory techniques. Bent functions are an important
class of functions in coding theory.
10. “First-order algorithms for packing-covering semidefinite programs”, joint work with G. Iyengar and C. Stein. Proceedings of the 2010 workshop on Modeling and Optimization: Theory
and Applications. 2012: 53–66.
• In this paper, we generalize some of our previous methods from [??] and [??] to solve a
larger class of semidefinite programs.
9. “Network topology measures”, joint work with R. Kincaid. Wiley Interdisciplinary Reviews:
Computational Statistics. 2011, 3(6):557–565.
• In this paper, we survey results from network topology measures, which are functions that
measure different properties of a graph such as connectivity. This work plays a principal
role in [??].
8. “Managing Magnetic Resonance Imaging Machines: Support tools for scheduling and planning”, joint work with A. Carpenter, L. Leemis, A. Papir, and G. Phillips. Health Care
Management Science. June 2011, 14(2): 158–173.
• In this work we developed a scheduling model and used techniques from approximation
algorithms to help determine patient schedules and resource plans for a radiology department.
• Contribution: Adam Carpenter and Alan Papir were my M.S. students advised by myself
on this project which was an optimization and simulation analysis of a resource planning
problem at a hospital. My contribution was in developing the optimization model and
writing all the optimization and simulation code. Adam, Alan, and Professor Leemis
developed and wrote the regression code. In addition, I wrote the manuscript.
7. “Approximating semidefinite packing problems”, joint work with G. Iyengar and C. Stein.
SIAM Journal on Optimization 21 (1), 2011.
• In this work, we generalize ideas from approximation algorithms for so-called vector packing problems to create simpler algorithms for a class of semidefinite programs. This work
was the journal version of [??] where the extra work involved a generalization of the methods described therein as well as an implementation. The implementation was extensively
tested and solves some of the largest semidefinite programs ever solved.
• Contribution: Professor Iyengar and Professor Stein and I were joint principal investigators on this work, which was based on my dissertation work. We all contributed equally
to the technical proofs. I conducted all of the computations and wrote most of the paper.
6. “ Feasible and accurate algorithms for covering semidefinite programs”, joint work with G.
Iyengar and C. Stein. Lecture Notes in Computer Science (3509): Proceedings of the 12th
Scandinavian Symposium and Workshops on Algorithms and Data Structures (SWAT 2010).
pp. 150–162.
5. “Visualizing and finding minimum delay flights.”, joint work with T. Dey and P. Steele. Journal
of Computational and Graphical Statistics. June 1, 2011, 20(2): 294–297.
• In this work, we created a graphical tool that uses Monte Carlo simuation and network
algorithms to generate predictions for the airline and travel route to take in order to
minimize delay time. The algorithm takes into account over twenty years of historical air
travel delay data.
• Contribution: Patrick Steele was an honors student jointly advised by Professor Dey
and myself. The project was a synthesis of optimization and statistical analysis. My
contribution was to develop the optimization model and writing the manuscript. Professor
Dey developed and wrote the statistical modeling code. Patrick wrote the optimization
and visualization code.
• Supplementary material constituted an additional eight pages.
4. “Models of malicious behavior in sponsored search”, joint work with G. Iyengar and C. Stein.
Proceedings of the 2007 Spring Simulation Multiconference (3). 2007: 143–151.
• In this work, we developed game theoretic models to investigate sponsored search auctions
such as the Google AdWords auction.
• Contribution: Professor Iyengar, Professor Stein, and I were joint principal investigators
on this work which, was based on my dissertation work.
3. “Approximation algorithms for semidefinite packing problems with applications to MAXCUT
and graph coloring”, joint work with G. Iyengar and C. Stein. Lecture Notes in Computer Science (3509): Proceedings of the 11th Conference on Integer Programming and Combinatorial
Optimization (IPCO XI). 2005: 152–166.
• In this work, we generalized algorithms from vector packing problems to solve the semidefinite programs used for MAXCUT and graph coloring.
2. “Scheduling an Industrial Production Facility”, joint work with E. Asgeirsson, J. Berry, C. A.
Phillips, C. Stein, and J. Wein. Lecture Notes in Computer Science (3064): Proceedings of the
10th Conference on Integer Programming and Combinatorial Optimization (IPCO X). 2004:
116–131.
• In this work we formulated an integer program and developed heuristics based on approximation algorithms to solve an industrial scheduling problem. The results of the model
and algorithm can be used to help decision makers determine resource plans.
• Contribution: My principal contribution was in both the coding and writing, which was
equally shared by Asgeirsson and I. I contributed to the model formulation equally with
my co-authors.
1. “Closed on-line bin packing”, joint work with E. Asgeirsson, U. Ayesta, E. Coffman, J. Etra,
P. Momčilović, V. Vokhshoori, Z. Wang, and J. Wolfe. Acta Cybernetica 15 (2002), no. 3.
361–367.
• We found and proved the best possible algorithm for a variant of bin packing problems.
• Contribution: Professor Coffman posed an open problem to our class which we solved
jointly.
PAPERS SUBMITTED/IN PREPARATION
1. (with R. Kincaid, S. Kunkler*, and M.D. Lamar) “Algorithm and Complexity for a Network
Assortativity Measure,” Submitted to Networks, 15 pages.
• In this work, we develop the first known polynomial time algorithm to find subgraphs
of a given nodal degree with minimum or maximum generalized Randić index. We show
that when a connected subgraph is required than the problem is NP-Hard. Finally, we
provide computations and applications of our method.
• Contribution: Professor Kincaid, Professor Lamar, and I jointly advised Sarah Kunkler on
her honors research. My principal contribution was in the development of the algorithm
and the complexity proof. Also, I contributed to roughly 40% of the computations.
2. (with C. Celerier*, D. Joyner, C. Melles, and S. Walsh*) “Edge-weighted Cayley graphs and
p-ary bent functions” currently submitted to Integers, 64 pages.
• We investigate the connections of graphs to more general version of bent functions than
considered in [11] using algebraic techniques.
• Contribution: ENS Walsh was an honors student of Professor Joyner who continued
research that ENS Celerier worked on in [??] below. Professor Joyner, Professor Melles,
and I are joint principal investigators of this work.
3. (with R. M. Lewis and R. Zhang*) “Solving semidefinite programs over symmetric, diagonally
dominant matrices,” in preparation. 20 pages.
• In this paper, we develop an algorithm to solve a class of semidefinite programs used in
network design problems.
• Contribution: I have done the majority of the technical analysis for this work and roughly
40% of the programming. I have also done the majority of the writing.
COURSES TAUGHT
Introduction to Operations Research, Fall 2006, Fall 2007, Spring 2012, Spring 2013, Spring
2015
Linear Programming, Fall 2006, Fall 2008, Fall 2009
Nonlinear Programming, Spring 2007
Network Optimization, Spring 2007, Spring 2009, Spring 2011, Fall 2014
Scheduling, Spring 2008
Integer Programming, Spring 2008, Fall 2012
Calculus I, Fall 2008, Spring 2011, Fall 2014
Calculus III, Fall 2011
Foundations of Mathematics, Spring 2009, Fall 2009, Spring 2010
Convex Optimization, Spring 2010, Spring 2014
Simulation modeling, Fall 2013, Spring 2014
RESEARCH STUDENTS
ENS Alec McGlaughlin, USNA (B.S. ‘14). Thesis: Analyzing and Assessing Brain Structure with Graph Connectivity Measures. Currently pursuing an M.S. in Bioengineering at Stanford
University.
ENS Taylor Boosmann, USNA (B.S. ‘13). Thesis: Staff scheduling academic radiologists.
Currently working for the Navy as an analyst.
ENS Robert Crutcher, USNA (B.S. ‘13). Thesis: Modeling the fight against a pathogen
population as a two-player game. Currently pursuing an M.D. at Harvard Medical School.
Sarah Kunkler, College of William & Mary (B.S. ‘12). Thesis: Algorithms and Complexity
for a Network Assortativity Measure. Currently working for Microsoft.
Patrick Steele, College of William & Mary (B.S. ‘11). Thesis: Minimizing Flight Delay.
Currently pursuing a Ph.D. in Operations Research at Cornell University.
Austin Powell, B. S., M.S., (William & Mary, ’10). The Maximum Cut Problem: Investigating
Computational Approaches. 2010.
Alan Papir, B. S. (Columbia University, ’09), M. S., (William & Mary, ’10). Currently
working for Booz Allen Hamilton.
Adam Carpenter, B. S., (William & Mary, ’08), M. S., (William & Mary, ’09). Currently
working for the Thomas Jefferson National Accelerator Facility .
Erik Vargo, B. S. (Lebanon Valley College, ’07), M.S., (William & Mary, ’09). Currently
at the University of Virginia for a Ph.D. in Systems and Information Engineering.
Alex Mills, B.S., (William & Mary ’07), Ph.D. (UNC-Chapel Hill, ‘12). Dr. Mills is
now an Assistant Professor at the Kelley School of Business at Indiana University in Bloomington,
Indiana.
Matthew Hanson, B.A., (William & Mary ’07). Currently working at the National Bureau
of Economic Research.
PROFESSIONAL ACTIVITIES
Planning Committee Member of Brown University’s Institute for Computational and Experimental
Research in Mathematics (ICERM) workshop, Semidefinite Programming and Graph Algorithms,
February 2014.
Session chair at the Informs Computing Society Conference, January 2011.
Co-organizer for the Summer CSUMS Research Experience for Undergraduates, 2008-2010.
Co-organizer for the GMU-WM Workshop on Undergraduate Research in Computational Mathematics, William & Mary, March 2009.
Referee for Mathematical Programming, Mathematics of Operations Research, ACM Transactions on
Algorithm, ACM-SIAM Symposium on Discrete Algorithms, MPS Conference on Integer Programming and Combinatorial Optimization, INFORMS Journal on Computing, Journal of Algorithms,
Naval Research Logistics, Linear Algebra and its Applications, and Health Care Management Science.
Member of INFORMS and the Association for Computing Machinery.
SELECTED PRESENTATIONS
“Scheduling with stochastic processing times in a medical clinic”. INFORMS Computing Society
Conference. January 2015. Richmond, VA.
“Solving SDPs over Symmetric, Diagonally Dominant Matrices”.
November 2014. San Francisco, CA.
INFORMS Annual meeting.
“Appointment scheduling and resource planning under uncertainty”. Lawrence Livermore National
Laboratory Seminar. November 2014. Livermore, CA.
“Scheduling and resource planning: models for a glaucoma clinic”. Walter Reed Military Medical
Center Opthalmology/Optometry Departmental Retreat. October 2014. Gaithersburg, MD.
“Structural MRI connectivity analyses using graph theory approaches to differentiate subjects across
the spectrum of AD”. Dementia Consortium Colloquia, Neurology Department, The Johns Hopkins
University Medical School. April 2014. Baltimore, MD.
“MRI Magnet Planning”. International Symposium on Mathematical Programming. August 2012.
Berlin, Germany.
“First-order methods for finding weighted graphs with high algebraic connectivity”. Discrete Math
Seminar, Columbia University. December 2010. New York, NY.
“Packing and covering semidefinite programs: algorithms and examples”, Computer Science Theory
Lunch, Brown University. November 2010. Providence, RI.
“Applying first-order methods to packing and covering semidefinite programs”. Modeling and Optimization: Theory and Applications 2010. Bethlehem, PA. August 2010.
“Algorithms and applications for packing and covering semidefinite programs ”. Technische Universität-Berlin Colloqium. Berlin, Germany. June 2010.
“Feasible and accurate algorithms for covering semidefinite programs ”. The 12th Scandanavian
Symposium and Workshops on Algorithm Theory. Bergen, Norway. June 2010.
“Solving the Sparse PCA SDP Relaxation ”. INFORMS Annual Meeting. San Diego, CA. October
2009.
“Data Expo 2009: Minimizing the probability of experiencing a flight delay ”. Joint Statistics
Meeting. Poster presentation. Awarded Honorable Distinction. Washington, DC. August 2009
“Managing magnetic resonance imaging machines: optimal purchases and schedules”. Ninth Workshop on Models and Algorithms for Planning and Scheduling Problems. Abbey Rolduc, The Netherlands. July 2009.
“Rounding time-indexed project scheduling integer programs to determine resource planning”. IN-
FORMS Computing Society Annual Meeting. Charleston, SC. January 2009.
“Scheduling Patients in a Radiology Department”. INFORMS Annual Meeting. Washington, DC.
October 2008.
“Graph optimization in medicine and mining”. CSUMS, The College of William & Mary. Williamsburg, VA. March 2008.
“Semidefinite packing problems”. Discrete Math Seminar, Columbia University. November 2007.
New York, NY.
“Models of malicious behavior in sponsored search”. Spring Sim 2007, Norfolk, VA. March 2007.
“Distribution-free strategies for budget-constrained bidders in sequential auctions”. Workshop on
Alternative Solution Concepts for Mechanism Design, EC 2006. University of Michigan, Ann Arbor.
June, 2006.
“Google Keywords auctions”. Cowles Foundation Conference on Optimization. Yale University.
New Haven, NH. February, 2006.
“Experimental Behavior of Algorithms for Semidefinite Relaxations”. INFORMS Annual meeting.
New Orleans, LA. November 2005.
“Semidefinite packing, a lower bound”. INFORMS Annual meeting. New Orleans, LA. November
2005.
“Approximating semidefinite packing problems”. IPCO XI. TU-Berlin, Berlin, Germany. June 2005.
“Approximating semidefinite packing problems”. IBM, Yorktown Heights, NY. March 2005.
“Scheduling an industrial facility”. INFORMS Annual meeting. Atlanta, GA. November 2003.
COMPUTER SKILLS
Languages/Scripting/Technical Software
Python, C/C++, R/Splus, Fortran, Gurobi, CPLEX, AMPL, SAS, LATEX, bash scripting,
Visual Basic, Pascal, Matlab, Mathematica, Excel, @Risk.
Operating Systems
Linux (various), Unix (various), Mac OS X, Windows