Abstract:We introduce DeepABM, a framework for agent-based modeling that leverages geometric message passing of graph neural networks for simulating action and interactions over large agent populations. Using DeepABM allows scaling simulations to large agent populations in real-time and running them efficiently on GPU architectures. To demonstrate the effectiveness of DeepABM, we build DeepABM-COVID simulator to provide support for various non-pharmaceutical interventions (quarantine, exposure notification, vaccination, testing) for the COVID-19 pandemic, and can scale to populations of representative size in real-time on a GPU. Specifically, DeepABM-COVID can model 200 million interactions (over 100,000 agents across 180 time-steps) in 90 seconds, and is made available online to help researchers with modeling and analysis of various interventions. We explain various components of the framework and discuss results from one research study to evaluate the impact of delaying the second dose of the COVID-19 vaccine in collaboration with clinical and public health experts. While we simulate COVID-19 spread, the ideas introduced in the paper are generic and can be easily extend to other forms of agent-based simulations. Furthermore, while beyond scope of this document, DeepABM enables inverse agent-based simulations which can be used to learn physical parameters in the (micro) simulations using gradient-based optimization with large-scale real-world (macro) data. We are optimistic that the current work can have interesting implications for bringing ABM and AI communities closer.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to improve the scalability and efficiency of large - scale Agent - Based Models (ABM) through deep learning and graph neural network technologies. Specifically, the paper introduces a new framework named DeepABM, which uses geometric message passing in graph neural networks to simulate the behaviors and interactions of a large number of agents. DeepABM aims to achieve real - time large - scale agent population simulation and can run efficiently on GPU architectures. ### Specific Problems and Solutions: 1. **Limitations of Existing ABM Frameworks**: - **Efficiency Problem**: Traditional ABM frameworks such as Mesa and NetLogo are conceptually appealing but are inefficient when dealing with large - scale agent populations and are difficult to simulate in real - time. - **Scalability Problem**: Traditional ABM frameworks face challenges when expanding to large - scale agent populations, especially in cases where real - time simulation is required. 2. **Innovations of DeepABM**: - **Network - Centric Approach**: DeepABM adopts a network - centric approach, modeling around the interaction network between agents, and uses tensor operations and the message - passing mechanism of graph neural networks to simulate agent interactions and state transitions. - **Efficient Computation**: By taking advantage of the benefits of graph neural networks and the computing power of modern GPUs, DeepABM can efficiently handle the real - time simulation of large - scale agent populations. - **Parameter Calibration**: DeepABM supports the simultaneous calibration of a large number of input parameters using supervised gradient optimization methods, rather than relying solely on random search methods. 3. **Application Examples**: - **COVID - 19 Simulator**: The paper presents a specific application of DeepABM - DeepABM - COVID, which is used to simulate the spread dynamics of COVID - 19 and study the impact of various non - pharmaceutical interventions (such as isolation, exposure notification, testing, vaccination) on public health outcomes. - **Performance Improvement**: DeepABM - COVID can simulate 200 million interactions (involving 100,000 agents over 180 time steps) within 90 seconds, while the same simulation in the Mesa framework takes approximately 5 hours. ### Main Contributions of the Paper: - **Technical Breakthrough**: Proposed a new ABM framework based on graph neural networks, solving the efficiency and scalability problems of traditional ABM in large - scale agent population simulation. - **Practical Application**: Through the DeepABM - COVID simulator, demonstrated the practical application value of this framework in public health policy - making, especially in response to the COVID - 19 epidemic. - **Research Case**: Analyzed in detail the impact of delaying the second - dose vaccination strategy on the effectiveness of vaccine distribution, providing a scientific basis for public health decision - making. In conclusion, this paper solves the key problems of large - scale agent - based models in terms of efficiency and scalability by introducing the DeepABM framework, and shows its great potential in the public health field through specific COVID - 19 simulation applications.

DeepABM: Scalable, efficient and differentiable agent-based simulations via graph neural networks

CoV-ABM: A stochastic discrete-event agent-based framework to simulate spatiotemporal dynamics of COVID-19

On the limits of agency in agent-based models

Using neural networks to calibrate agent based models enables improved regional evidence for vaccine strategy and policy

A framework for modeling human behavior in large-scale agent-based epidemic simulations

Deep Learning-based Spatially Explicit Emulation of an Agent-Based Simulator for Pandemic in a City

High-Performance and Scalable Agent-Based Simulation with BioDynaMo

An AI-enabled Agent-Based Model and Its Application in Measles Outbreak Simulation for New Zealand

Agent based simulators for epidemic modelling: Simulating larger models using smaller ones

Large-scale agent-based simulations of online social networks

EconAgent: Large Language Model-Empowered Agents for Simulating Macroeconomic Activities

ABIDES: Towards High-Fidelity Market Simulation for AI Research

Detailed parallel social modeling for the analysis of COVID-19 spread

Automated and Distributed Statistical Analysis of Economic Agent-Based Models

Agent-Based Modelling Meets Generative AI in Social Network Simulations

Using machine learning as a surrogate model for agent-based simulations

Smart Agent-Based Modeling: On the Use of Large Language Models in Computer Simulations

Application of Machine Learning Techniques to an Agent-Based Model of Pantoea

OpenCity: A Scalable Platform to Simulate Urban Activities with Massive LLM Agents

Deep Learning in (and of) Agent-Based Models: A Prospectus