Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

An intelligent agent-based simulation of human mobility in extreme urban morphologies

Abderaouf Bahi, Amel Ourici

TL;DR

A hybrid simulation framework that integrates agent-based modeling, reinforcement learning, reinforcement learning, supervised learning, and graph neural networks is developed, suggesting that efficient and sustainable mobility in extreme urban forms is achievable, provided adaptive AI systems, intelligent infrastructure, and real-time feedback mechanisms are implemented.

Abstract

This paper investigates the feasibility of human mobility in extreme urban morphologies, characterized by high-density vertical structures and linear city layouts. To assess whether agents can navigate efficiently within such unprecedented topologies, we develop a hybrid simulation framework that integrates agent-based modeling, reinforcement learning (RL), supervised learning, and graph neural networks (GNNs). The simulation captures multi-modal transportation behaviors across multiple vertical levels and varying density scenarios, using both synthetic data and real-world traces from high-density cities. Experiments show that the full AI-integrated architecture enables agents to achieve an average commute time of 7.8--8.4 minutes, a satisfaction rate exceeding 89\%, and a reachability index over 91\%, even during peak congestion periods. Ablation studies indicate that removing intelligent modules such as RL or GNN significantly degrades performance, with commute times increasing by up to 85\% and reachability falling below 70\%. Environmental modeling demonstrates low energy consumption and minimal CO$_2$ emissions when electric modes are prioritized. These results suggest that efficient and sustainable mobility in extreme urban forms is achievable, provided adaptive AI systems, intelligent infrastructure, and real-time feedback mechanisms are implemented.

An intelligent agent-based simulation of human mobility in extreme urban morphologies

TL;DR

A hybrid simulation framework that integrates agent-based modeling, reinforcement learning, reinforcement learning, supervised learning, and graph neural networks is developed, suggesting that efficient and sustainable mobility in extreme urban forms is achievable, provided adaptive AI systems, intelligent infrastructure, and real-time feedback mechanisms are implemented.

Abstract

This paper investigates the feasibility of human mobility in extreme urban morphologies, characterized by high-density vertical structures and linear city layouts. To assess whether agents can navigate efficiently within such unprecedented topologies, we develop a hybrid simulation framework that integrates agent-based modeling, reinforcement learning (RL), supervised learning, and graph neural networks (GNNs). The simulation captures multi-modal transportation behaviors across multiple vertical levels and varying density scenarios, using both synthetic data and real-world traces from high-density cities. Experiments show that the full AI-integrated architecture enables agents to achieve an average commute time of 7.8--8.4 minutes, a satisfaction rate exceeding 89\%, and a reachability index over 91\%, even during peak congestion periods. Ablation studies indicate that removing intelligent modules such as RL or GNN significantly degrades performance, with commute times increasing by up to 85\% and reachability falling below 70\%. Environmental modeling demonstrates low energy consumption and minimal CO emissions when electric modes are prioritized. These results suggest that efficient and sustainable mobility in extreme urban forms is achievable, provided adaptive AI systems, intelligent infrastructure, and real-time feedback mechanisms are implemented.

Paper Structure

This paper contains 30 sections, 21 equations, 5 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related works
  3. Methodology
  4. Hypothetical Model of The Line
  5. Simulation Scenario
  6. Implementation
  7. Reinforcement Learning (RL) for Autonomous Navigation
  8. Supervised Learning for Demand Prediction
  9. Graph Neural Networks (GNNs) for Route Optimization
  10. Agent Decision-Making Logic
  11. Calibration and Validation
  12. Behavior Modeling of Agents
  13. Experiments and Results
  14. Datasets
  15. Evaluation Metrics
  16. ...and 15 more sections

Figures (5)

  • Figure 1: NEOM’s The Line Chart
  • Figure 2: Agent-based simulation of human mobility in The Line
  • Figure 3: Heatmap of the ablation study on the synthetic dataset.
  • Figure 4: Heatmap of the ablation study on the real-world dataset.
  • Figure 5: Reachability Index across scenarios and model configurations.