Alumni Directory

Vijay Kalivarapu
PhD, Summer 2008

Home Dept: Mechanical Engineering

PhD Dissertation title: Improving solution characteristics of particle swarm optimization through the use of digital pheromones, parallelization, and graphical processing units (GPUs)

PhD Advisor: Eliot Winer

Area of PhD research: Optimization has its foundations dating back to the days of Newton, Lagrange, Cauchy, and Leibnitz when differential calculus methods were developed to minimize and maximize analytical functions. Substantial progress in optimization became more prominent in the mid to late twentieth century when digital computers showed promise in offloading analytical problem solving into numerical methods through computer code for faster evaluations of designs. Deterministic optimization methods such as steepest descent, conjugate gradient and Newton's methods are known for their robustness in iteratively reducing the objective function value for minimization problems. However, they are primarily suitable for solving single objective function problems that are unimodal and continuous. With increased sophistication in engineering problems, multimodal and multi-objective problems have become more prevalent drastically reducing the effectiveness of deterministic methods. This led to the development of heuristic methods, particularly evolutionary methods such as Genetic Algorithms, Ant Colony Optimization, and Particle Swarm Optimization. These methods have multiple design points exploring the design space over iterations as opposed to a single design point as in the case of deterministic methods. Evolutionary methods come with the capability to solve multimodal discontinuous design spaces with increased reliability and efficiency, but at considerable computational expense.

Employment upon graduating:
academicPost Doctoral Student
Iowa State University
Ames, IA

<---- Back to search results