I came to UT Arlington to pursue my master’s in August 2007 in Industrial Engineering. I was a life-long student having never worked in the industry, so it was not a very big jump to go from B.Tech to a master’s and then on to a Ph.D. which I completed in December 2013. The operations research courses attracted me the most in my master’s, especially the ones taught by Dr. Corley, Dr. Rosenberger, and Dr. Chen. Even though I made some B’s, I had a lot of fun learning some amazing mathematics and statistics.

As a part of my master’s, I undertook research projects under Dr. Jay Rosenberger which gave me my first exposure to Ph.D. level research. The projects were done by previous students in the Center On Stochastic Modeling, Optimization, & Statistics (COSMOS) . My first impression of joining the COSMOS lab was when I peeked through the door and found that Spock from Star Trek was looking straight at me, disapprovingly. I started smiling when I realized it was only a life-sized replica. I was relieved to see that even though I was jumping into some serious research, there was still an element of humor to lighten the somber mood of the lab. Even though, quite frustratingly, when my own lack of knowledge and ignorance was brought into sharp focus, I looked at those times as a great learning opportunity. I worked on a nurse optimization project which was entering the testing stage at HEB hospital. After my master’s, I joined the Ph.D. program as a natural extension to my research work. Around that time, Dr. Anjomani came to Dr. Rosenberger for optimization of urban planning. The problem was to choose which would be the best land use to assign any given land piece. Each land piece would be given a certain rating regarding various land uses, e.g., high residential, low industrial, open areas, etc. It was a simple assignment optimization problem. We tried to use the methodology that was used for nurse optimization for urban planning optimization, but we hit a block which did not allow us to make any progress due to symmetry issues.

Given that we recognized we would not be able to make progress in the research in the direction we were headed, we decided to switch directions. I extended the simple land use optimization problems with more constraints by taking into account urban sprawl. After literature review, I recognized that urban sprawl had a lot of research towards solving it but there were not many comprehensive optimization models. So, the next step was to construct a mathematical model which would address all the various factors contributing towards urban sprawl. After the model was constructed, it was so large that CPLEX, an optimization software, would not even accept the model. It allowed us to experiment with Benders Decomposition method to reduce the model size and allow us to solve the model.

I also worked as a teaching assistant for the IMSE Department when I joined the master’s program. I sincerely enjoyed my time at UT Arlington since it gave me an opportunity to try a variety of things, e.g., salsa, yoga, aikido, etc.

I am currently working as a software engineer for E2open as their optimization specialist.

Written by Piyush Kumar, Ph.D. Industrial Engineering
IMSE Alumni 2013

Leave a Reply