Document Type

Thesis

Degree

Master of Science

Major

Computer Science

Date of Defense

11-16-2018

Graduate Advisor

Sharlee Climer, Department of Mathematics and Computer Science

Committee

Sharlee Climer, PhD

Badri Adhikari, PhD

Sanjiv K. Bhatia, PhD

Carlos Cruchaga, PhD

Abstract

With the advent of genetic sequencing, there was much hope of finding the inherited elements underlying complex diseases, such as late-onset Alzheimer’s disease (AD), but it has been a challenge to fully uncover the necessary information hidden in the data. A likely contributor to this failure is the fact that the pathogenesis of most complex diseases does not involve single markers working alone, but patterns of genetic markers interacting additively or epistatically. But as we move upwards beyond patterns of size two, it quickly becomes computationally infeasible to examine all combinations in the solution space. A common solution to solving this type of combinatorial optimization problem is to model it as a mixed-integer linear program (MIP) and solve it using the algorithm branch-and-cut, implemented by a commercial solver. However, with the trend of using increasing numbers of computing cores to increase computational power, there is a need for a different approach to solving MIPs that can utilize parallel environments. Here we show how a parallelized implementation of an alternative algorithm, cut-and-solve, can be used to solve this genetics problem faster than CPLEX, one of the leading commercial MIP solvers.

Share

COinS