Student Activities
Advisor: Muhammad Usman, Ph.D
Stander Symposium 2013
Student -
Nicholas Haynes
We
demonstrate numerically the eventual time-periodicity of the solutions of the
Korteweg-de Vries equation with periodic forcing at the boundary using the sinc-collocation
method. This method approximates the space dimension of the solution with a
cardinal expansion of sinc functions, thus allowing
the avoidance of a costly finite difference grid for a third-order boundary
value problem. The first-order time derivative is approximated with a weighted
finite difference method. The sinc-collocation method was found to be more
robust and more efficient than other numerical schemes when applied to this
problem
2. Simulation of Nonlinear Waves Using Sinc Collocation-Interpolation
Student -
Eric A Gerwin, Jessica E Steve
In this project we explore the Sinc collocation method to solve an initial and boundary value
problem of nonlinear wave equation. The Sinc collocation method is based
upon interpolation technique, by discretizing the
function and its spatial derivatives using linear
combination of translated Sinc functions. Our project will focus on multiple
boundary conditions such as the well known Dirichlet and Neumann
conditions. Our project will also focus on two
established nonlinear partial differential equations: the
Sine-Gordon equation and the Kortweg-de Vries equation.
3. Exploring the Sinc-Collocation Method for Solving the Integro-Differential Equation
Student -
Han Li
In this project we study the Sinc
approximation method to solve a family of integral differential equations. First
we will apply the Sinc-collocation
method to solve the second order Fredholm integro-differential
equation. Numerical results and examples demonstrate the reliability and
efficiency of
this method. Secondly, various types of integro-differential equations are
solved by Sinc-collocation technique and the numerical results
are compared, to explore the stability of this method.
Stander Symposium 2012
Student -
Nicholas Haynes
A recent paper in the Journal of General Physiology disproved the hypothesis
that the ciliary axoneme and the basal bodies of cilia impose selective barriers
to the movement of proteins into and out of the the cilium using a combination
of numerical modeling and observation with confocal and multiphoton microscopy.
We compare the accuracy and computational efficiency of the numerical method
used in the paper, known as the method of lines, to another method, known as
sinc collocation, and discuss the possible use of other methods for improving
the algorithm.
2. Stability Analysis of a Model for In Vitro Inhibition of Cancer Cell Mutation
Student -
Chris Yakopcic
Human homeostasis is the body's ability to
physiologically regulate its inner environment to ensure its stability in
response to changes in the outside environment. An inability to maintain
homeostasis may lead to death or disease, which is caused by a condition known
as homeostatic imbalance. Normal cells follow the homeostasis when they
proliferate and cancer cells do not. This work describes a model consisting of
three reaction-diffusion equations representing in vitro interaction between two
drugs. One inhibits proliferation of cancerous cells, and the other destroys
these cells. A stability analysis of the model is performed with and without
diffusion applied to the model. MATLAB is used to perform the stability analysis
of the model.
3. Numerical Investigation into a Computational Approximation of Bifurcation Curves
Student -
Joshua R Craven
In this project, I use computational tools to study the bifurcations in
nonlinear oscillators. Matlab is first used to determine the slow flow phase
portrait of each region and the characteristics of each critical point. Next,
the parameters are discretized and for each set of values we find the locations
of the real critical points and the eigenvalues of the Jacobian matrix. With
this knowledge, we can approximate the bifurcation diagram. These results are
compared with results from preexisting software.
Student -
Alyssa C Lesko
IL-2 treatments have recently been identified to significantly reduce
metastatic melanoma tumors and in some cases eliminate them. The problem with
these treatments is that a set plan of administration varies from patient to
patient and methods for determining treatment steps are still in the process of
being developed. Previous research by Asad Usman and colleagues has used a
numerical technique using MATLAB to decide treatment protocols. This research
used the MATLABâ??s built in ode15 function to addresses treatment procedures
including the starting and stopping of each treatment and the period in between
each treatment. Building on this data and existing model, my project will
explore several other numerical techniques such as ode23 and ode45 solvers,
Eulerâ??s method, and the predictor corrector method to study IL-2 treatments in
metastatic melanoma patients. A comparison will be made using error plots and
tables, and a stability analysis using pplane7 will be investigated.
5. Mathematical Study of the Foot and Mouth Outbreak Model
Student - Jungmi Johnson
The foot and mouth outbreak in the UK in 2001 was a disastrous event for
the country and the economic. The disease did not only cost UK government so
much money to stop the disease, but it also affected the tourism industry.
Mathematical epidemic models can provide clear strategy for minimizing the
effect of such a disease, determining the expected manner of its progression in
the event of a future outbreak based upon the latest available data on the
epidemic. This project is to explore how to minimize the cost, how to contain
the disease in minimal time, and how realistic these models will be considering
the limitation of the model. Numerical and qualitative tools such as MATLAB's
built in ode solver will be used.
6. Applying Mathematical Epidemic Modeling to Discover Commercially Beneficial Outbreak Control Methods
Student -
Michael Ciesa
This project is a mathematical analysis and computational study of the 2001 foot
and mouth epidemic in the UK. This model includes an application of the SIR
model, developed by W. O. Kermack and A. G. McKendrick, with two additional
factors: vaccination and incubation period infectives. The incubation period
infectives represent the population of individuals infected with the disease
that do not show symptoms, but still have the possibility of infecting other
individuals.
7. Qualitative Study of an SIR epidemic model with an asymptotically homogeneous transmission function
Student -
Karoline E. Hoffman
I will be exploring and analyzing an SIR epidemic model. This particular
model has an asymptotically homogeneous transmission function which means the
transmission rate is proportional to the fraction of the number of infective
individuals to the total population. I will also look at a qualitative analysis
of the model and then discuss the implications of the results of the model.
Stander Symposium 2011
A Numerical Study of In Vitro Inhibition of Mutation of Cancer Cells Using Two Different
Student -
Giacomo Flora
The growth of in-vitro cancer cells has been studied
using two numerical methods: the Predictor-Corrector and the Operator Splitting
method. The mathematical model developed by Dey (2000) is used, which consists
of three reaction-diffusion equations representing in vitro interaction between
two drugs, one which inhibits the proliferation of the cancer cells and the
other which destroy these cells. The solutions resulting from the application of
the two methods are in excellent agreement. In addition stability analyses of
model and diffusion free case have been performed.
Stander Symposium 2010
1. A Graphical User Interface for Solving the Falkner-Skan Equation (Download all MATLAB codes here)
Student -
Giacomo Flora
A Graphical User Interface (GUI) has been developed to solve the Falkner-Skan
equation. This famous nonlinear third order Falkner-Skan equation on infinite
interval describes several fluid dynamic problems under varying the value of two
constant coefficients. The developed GUI enables the user to input these two
coefficients, which will characterize the behavior of the corresponding
solution, and the parameters necessary for the iterative methods used to solve
the equation. At this regard, a shooting method developed by Zhang J. and Chen
B. has been adopted. The required parameters for the numerical solution are
represented by the initial shooting angle, the initial free boundary and two
tolerance criteria.
Student -
Elham Negahdary
In this work we revisit some relatively new techniques based on radial basis
functions (RBFs) to interpolate, boundary-value and initial-boundary-value
problems with high degrees of localization in space and/or time. First, we
generate an initial discretization using equally spaced points and find the RBF
approximation of the function. Next, we compute the interpolation error at
points halfway between the nodes. Points at which the error exceeds a threshold
become centers, and centers that lie between two points with error less than a
smaller threshold are removed. The two end points are always left intact. We
also adapt the shape parameters of RBFs based on the node spacing to prevent the
growth of the conditioning of the interpolation matrix. The shape parameter of
each center is chosen based on the spacing with nearest neighbors, and the RBF
approximation is recomputed using the new center set. Recent work in the
literature on radial basis functions method has shown some promising results in
terms of accuracy and efficiency to solve higher order nonlinear partial
differential equations. Since radial basis functions methods are completely
meshfree, requiring only interpolation nodes and a set of points called centers
defining the radial basis functions. Adaptive radial basis function approach is
based on refining and coarsening nodes based on shape parameter, interpolation
error and condition number of the interpolation matrix.
3. Kinetic Modeling of A Spherical Catalytic Particle
Student -
Fadhel Zammouri
It is critical for chemical engineers to understand the kinetics behavior
associated with catalytic particles. This is crucial in the design and
fabrication of catalytic reactors. In this work, a mathematical model for the
interplay between the rates of molecular transport (diffusion) and the intrinsic
activity (chemical kinetics) is studied. The concept of effectiveness factor in
catalytic first order chemical reactions, exothermic and endothermic, is
addressed in detail. Also the behavior of chemical reaction rates and the
temperature gradient in the catalytic particle with respect to stability is
discussed. The Numerical solution of the model has been computed for various
values of the parameters. For all of our computations and numerical simulations
we have used MATLAB.
4. Measles Epidemic: Studying the Spread Using Numerical Techniques
Student -
Jaye S. Flavin
Studying past epidemics is a necessary step in understanding and preventing the
spread of future contagions. The measles epidemic in New York in the mid-1960s
is an ideal case study for mathematical epidemiology because of the detailed
records kept on those infected and the unique properties of measles as a
disease. Using a Computer Algebra System (CAS), we will revisit the qualitative
properties of the measles epidemic model and compare solutions using different
numerical techniques.
Undergraduate Research in Mathematical Biology
Advisors: Muhammad Usman, Ph.D,
Amit Singh, Ph.D (Department of Biology)
Stander Symposium 2010
5. A Computational Study of the Fitzhugh-Nagumo Action Potential System
Student(s) -
Joseph R Salomone, Anna M. Stcyr, Angela Q. Umstead
The most celebrated example of mathematical modeling is the Hodgkin-Huxley model
of nerve physiology. Their experiments were carried out on a giant axon of a
squid, which was large enough for the implantation of electrodes. The
Hodgkin-Huxley mathematical model for nerve cell action potential is a system of
four coupled ordinary differential equations. The Fitzhugh-Nagumo two-variable
action potential system behaves qualitatively like the Hodgkin-Huxley
space-clamped system. Being simpler by two variables, action potentials and
other properties of the Hodgkin-Huxley model may be visualized as phase-plane
plots. We use MATLAB to study the numerical solutions as well as the qualitative
properties of the model.
6. Mathematical Modeling of H1N1 Flu
Student(s) -
William E. Balbach Nathan B. Frantz Brett R. Mershman William T. Weger
Mathematical models have been used to understand the dynamics of infectious
diseases and to predict the future epidemic or pandemics. In 2009, a new strain
of the influenza A (H1N1) virus spread rapidly throughout the world. This “swine
flu” as it is commonly known, increased to what is considered an epidemic in a
matter of months. In order to understand the spread of this virus, and similar
patterns in future outbreaks, we study a simplified SIR mathematical model to
answer some epidemiological questions. We solve the model numerically and also
study the qualitative properties of the model. It is important
to mention that a solution of a mathematical model is not necessarily a solution
to the real problem, but a solution to a simplified idealization of the real
world problem.
7. Mathematical Modelling of Infectious Diseases
Student(s) -
Kevin M. George, Branden J. King
Study of infectious diseases has become more important with increased global
connectivity and personal contact. The discovery of the microscope in the 17th
century caused a revolution in biology by revealing otherwise invisible.
Mathematics broadly interpreted is a more general non optical microscope.
Mathematical model helps to understand dynamics of how they spread, how many
people are infected, resist the infection, or recover. In this work we study
infectious diseases models qualitatively. These mathematical models are solved
numerically using MATLAB.
Student Research 2009
Miriam Poteet (Currently a Ph.D. Student at AFIT)
Math Clinic: Stability of steady state solutions of the forced Kuramoto-Sivashinky(KS) equation
Stander Symposium 2009 Poster Presentation
David Aaby (Currently a graduate student in Biostatistics at University of Michigan)
Undergraduate Research in Mathematical Biology: A comparative study of numerical methods for the Hodgkin-Huxley model of nerve cell action potential, (Citation)
Advisors: Muhammad Usman, (Mathematics), Amit Singh, (Biology)
Stander Symposium 2009 Poster Presentation