EpiCity: An AI-Enabled Epidemic-Aware Smart City Health Intelligence Framework for Sustainable Urban Planning

Urban centres across the world remain structurally unprepared for epidemic events, lacking the real-time intelligence infrastructure needed to anticipate outbreaks, model population-level spread, and evaluate competing health interventions before crises escalate. This chapter presents EpiCity, an open-source, AI enabled epidemic-aware smart city health intelligence framework designed to embed probabilistic outbreak forecasting directly into the urban planning and administrative decision cycle. EpiCity integrates four technical contributions within a unified deployable dashboard: a hybrid ensemble probabilistic forecasting engine driven by more than 500 Monte Carlo simulations; an agent based urban population digital twin that models SEIRD-compartment epidemic dynamics across 200 heterogeneous agents over a 60-day simulation horizon; an intervention policy scenario comparison module for quantitative evaluation of non-pharmaceutical and pharmaceutical control strategies; and a Retrieval-Augmented Generation (RAG) explainable AI chatbot that translates model outputs into plain-language guidance for city officials without specialist epidemiological training. Real-world validation against Johns Hopkins CSSE COVID-19 surveillance data for the United States (June-July 2020) yielded a Pearson correlation of 0.88 between framework forecasts and observed case trajectories, with ensemble prediction intervals achieving 90% empirical coverage, confirming the calibration reliability of the uncertainty quantification pipeline. A representative simulation scenario demonstrated a 99% reduction in simulated peak case load relative to an unmitigated baseline, with an estimated 56 lives saved in the comparison period. Aligned with SDG 3 (Good Health and Well-being), SDG 11 (Sustainable Cities and Communities), and SDG 13 (Climate Action), EpiCity offers urban planners, public health officers, and city administrators a scientifically grounded yet practically accessible tool for evidence-based epidemic preparedness and resilient city governance. The code is available at https://github.com/Harmi-kotak22/Artificial_Life_Simulator.