Issues of data processing and storage allocations are the focus of a winning reflection by a computer science major in this year’s Huskies Showcase.
Subi Dangol, of Sauk Centre, won the award for Best Our Husky Compact Reflection in the Integrate Existing and Evolving Technologies dimension for her presentation “Improved Algorithms for a Common Assignment Problem.” Her faculty mentor is Ramnath Sarnath.
Huskies Showcase is being presented virtually on a volunteer basis for students, faculty and staff to interact with the projects on D2L Brightspace. Finalists and Husky Compact dimension winners were named this April. No awards will be presented for oral or poster presentation beyond finalist designations due to limitations of the virtual format.
The dimensions come from Our Husky Compact. The six dimensions embody the essential and cross-cutting attributes of a St. Cloud State University education. These attributes are developed over time and across and beyond the curriculum.
Presentation Abstract:
We are entering a world where massive amounts of data can be stored and processed by computers. When these large data sets are structured and analyzed correctly, it can provide valuable insights to help make better decisions. Unfortunately, handling large volume of data can be overwhelming and expensive.
To deal with the large data volume in a timely manner, we need to find new approaches and evaluate their effectiveness. In this research, we will be evaluating the effectiveness of new approaches to a real-world resource allocation problem called the Load Balancing with Overloading Penalties.
The main objective of this problem is to find the best assignment of users to their desired service providers (which could be a school, a hospital, or a mobile phone tower). This problem also addresses the issue of decrease in quality when a service provider is over-utilized by using a penalty variable. Due to the complexity and additional constraints of this problem, finding the best solution can very difficult and even impossible depending on the size of the data.
We have come up with a new approach to this problem that is more accurate than the existing solutions while still being time efficient. Our solution takes advantage of some known results from Graph Theory and existing solutions by breaking the problem down into different graph search algorithms for better data arrangement and faster processing. This new approach also has the possibility to open a new way to view this problem that could lead to more discoveries.
Reflection:
I believe this research has helped me achieve the goals of the dimension “Integrate Existing and Evolving Technologies.” In this research, we chose a problem that frequently arises in the real world so that we can contribute a new and better approach to this problem.
In addition, it can help all the fields associated with this problem. Through this, we were able to “critically evaluate [this project] and its impact on the society,” which is the first outcome of the dimension.
There were solutions that existed for this problem, but despite that we decided to investigate their nature and effectiveness. This “demonstrates our intellectual inquiry and curiosity characterized by innovation and divergent thinking.”
This project “engaged multiple perspectives” as Dr. Sarnath, Dr. Gunturi, and I collectively worked on evaluating the problem. We had meetings every week to discuss different approaches to this problem.
Many initial attempts were not successful, but through their expertise on algorithms and research, we were able come up with a better approach for this problem. This experience made me realize my passion, which is academics and research. I was able to gain a lot of research experience through this project. This will be extremely helpful as I pursue further education in this area of computer science.
