Anandhu Dileep

I studied physics and mathematics during my undergrad years and graduated with a Master of Mathematics in 2020. During my master's, I worked on discrete optimization during my thesis work, and I applied available optimization methods for two case studies related to a vehicle routing problem in agriculture. Thereafter I developed a novel optimization algorithm for the problem, based on the well-known petal heuristics. After my master's, I optimized the algorithm further and it is currently in the production stage as a research article.

Meanwhile, as I started studying machine learning, I began to see some works making use of machine learning in optimization for data forecast for defining problem instances with uncertainty, and also some other works making use of optimization for interpretable or explainable machine learning algorithms. Then I found out the even more direct application of machine learning, or more specifically reinforcement learning, for the design of hyperheuristics for solving wide classes of optimization problems. This is the research topic that I am currently starting to work on. It is very exciting to me as I will get to know more about reinforcement learning which is an integral part of modern robotics and apply very similar strategies in the field of optimization in the hope of making wonders as was done in robotics. 

Besides this topic, I am also fascinated by the mathematical programming approaches for discrete optimization, which usually is not scalable but still, the theory behind the attempts to make it scalable seems very beautiful in many cases. I also like to read about formal logic and theoretical computer science and also about some of the important historical developments in Mathematics to understand the process of how important ideas slowly developed over the years. My other hobbies include watching good movies, playing football, and having long walks.