Optimizing Urban Design for Pandemics Using Reinforcement Learning and Multi-objective Optimization

Abstract

The present study demonstrates a novel approach to leveraging reinforcement learning and multi-objective optimization for enhancing urban preparedness against pandemics. The role of urban design in preventing the spread of infectious diseases is significant, as evidenced by the COVID-19 pandemic, highlighting the need for preparedness for potential future pandemics. The method proposed in this study employs a hybrid approach of reinforcement learning and multi-objective optimization to identify optimal solutions for urban design that effectively reconcile diverse objectives, including but not limited to public health, economic viability, and environmental sustainability. The findings obtained from a simulated outbreak demonstrate that the proposed approach exhibits superior performance in comparison to the currently available methods. This suggests that it could be used to help plan cities for future pandemics. The utilization of reinforcement learning has the potential to enhance urban planning by employing a reward-based mechanism to instruct an agent on the prevention of a pandemic outbreak. The consideration of multiple objectives simultaneously can lead to further enhancement in the optimization process, which is commonly referred to as multi-objective optimization. The proposed methodology has the potential to mitigate the transmission of pandemics while taking into account the economic ramifications and the standard of living. The findings of this investigation illustrate the feasibility of utilizing reinforcement learning and multi-objective optimization techniques for the purpose of optimizing urban design interventions aimed at mitigating pandemics. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.