Optimizing Urban Mobility Restrictions: a Multi-Agent System (MAS) for SARS-CoV-2

TL;DR

This paper presents a multi-agent system model for simulating SARS-CoV-2 spread in Toronto, incorporating detailed mobility and demographic data to evaluate the impact of restrictions and policies.

Contribution

It introduces a novel MAS-based modeling framework that integrates rich spatial, temporal, and epidemiological data for urban COVID-19 simulation.

Findings

Mobility restrictions influence contagion dynamics.

The model reproduces complex emerging behaviors.

Simulations can inform policy decisions.

Abstract

Infectious epidemics can be simulated by employing dynamical processes as interactions on network structures. Here, we introduce techniques from the Multi-Agent System (MAS) domain in order to account for individual level characterization of societal dynamics for the SARS-CoV-2 pandemic. We hypothesize that a MAS model which considers rich spatial demographics, hourly mobility data and daily contagion information from the metropolitan area of Toronto can explain significant emerging behavior. To investigate this hypothesis we designed, with our modeling framework of choice, GAMA, an accurate environment which can be tuned to reproduce mobility and healthcare data, in our case coming from TomTom's API and Toronto's Open Data. We observed that some interesting contagion phenomena are directly influenced by mobility restrictions and curfew policies. We conclude that while our model is able to reproduce non-trivial emerging properties, large-scale simulation are needed to further investigate the role of different parameters. Finally, providing such an end-to-end model can be critical for policy-makers to compare their outcomes with past strategies in order to devise better plans for future measures.