Reconfigurable network topology design using algorithmic relaxation

We are sorry, this position has been filled.

Many network topology design problems in the context of High Performance Computing (HPC) and data centers are NP-hard problems. Yet these problems are important due to a network topology's potential impact in defining network performance. In this project, we intend on using various algorithmic techniques such a genetics algorithm, cross entropy method, etc to obtain approximate solutions to these complex problems. Much of the methods will be implemented in C++.

Lab: Lightwave Research Laboratory, Keren Bergman

Direct Supervisor: Keren Bergman

Position Dates: 5/20/2019 - 8/16/2019

Hours per Week: 40

Paid Position: Yes

Credit: Yes

Number of positions: 1

Qualifications: Basic optimization algorithms, programming

Eligibility: Junior (SEAS only)

Min Yee Teh, mt3126@columbia.edu