PhD Dissertation Proposal by: Shaon Bhatta Shuvo
Date:Thursday, October 5, 2023
Time: 11:00 am – 12:30 pm.
Location: Essex Hall Room 122
Abstract: In the face of challenges posed by infectious diseases such as COVID-19, there is a pressing need to revisit and refine our epidemiological modelling techniques for enhanced future preparedness. Proper intervention policies, rooted in accurate modelling, are pivotal; they not only limit the immediate impact of such diseases but also curtail their long-term societal and economic ramifications. This dissertation proposal sets its sights on a pioneering approach, aiming to merge the intricate detail of Agent-Based Modeling (ABM) with the predictive structure of mathematical compartmental models like Susceptible, Exposed, Infectious, Recovered, and Dead (SEIRD). ABM is renowned for its capability to simulate complex relationships in decision-making processes, yet it grapples with challenges related to scalability and intricate parameter estimation. Compartmental models such as SEIRD, and its extended versions, while structured, often have limited decision-making adaptability, and can struggle with changing parameters over time, potentially oversimplifying disease dynamics in varied populations. To navigate these issues, this research proposes a hybrid model, leveraging the strengths of both methodologies. A vital component of this proposal is the incorporation of N-step Deep Q Reinforcement Learning (N-Step DQRL). With its capability to learn and optimize from extensive datasets, N-Step DQRL is anticipated to play a transformative role in devising and fine-tuning intervention policies, ensuring timely and effective responses. Through this integrative approach, the objective is to provide decision-makers with a robust modelling tool, priming for a proactive and data-informed stance against looming health threats.
Internal Reader: Dr. Dan Wu
Internal Reader: Dr. Saeed Samet
External Reader: Dr. Abdulkadir A. Hussein
Advisor: Dr. Ziad Kobti, Dr. Narayan Kar
