HD-GEN: A High-Performance Software System for Human Mobility Data Generation Based on Patterns of Life

TL;DR

HD-GEN addresses the need for realistic, scalable synthetic human mobility data by unifying Pattern-of-Life (POL) trajectory generation with reproducible calibration, processing, and visualization in an end-to-end software platform. It introduces a parallel genetic-algorithm calibration against GeoLife-based metrics, a CPU-based generation engine with checkpointing, and a downstream pipeline that converts logs to GIS-ready datasets while enabling real-time visualization. The system demonstrates GeoLife+-calibrated datasets, massive trajectory data across multiple regions, and anomaly-injected mobility data, all produced with reproducible, cross-platform workflows. By combining realism from POL with scalable data generation and open-source tooling, HD-GEN provides a practical foundation for mobility analytics, ML, anomaly detection, and urban studies while preserving privacy and reproducibility. Similarity between real and simulated data is quantified using metrics $ADT$, $ADA$, $MXD$, and $MDD$ via the equation $Similarity(G,P) = 1 - \frac{1}{|M|} \sum_{k \in M} \frac{|k(P) - k(G)|}{k(G)}$, guiding calibration toward GeoLife-like behavior.

Abstract

Understanding individual-level human mobility is critical for a wide range of applications. As such, real-world trajectory datasets provide valuable insights into actual movement behaviors and patterns of life but are often constrained by data sparsity and participant bias. Synthetic data, by contrast, offers scalability and flexibility but frequently lacks realism. To address this gap, we introduce a comprehensive software pipeline for, generating, calibrating, processing, and visualizing large-scale individual-level human mobility datasets that combine the realism of empirical data with the control and extensibility of Patterns-of-Life simulations. Our system consists of four integrated components. (1) a data generation engine which constructs geographically grounded simulations using OpenStreetMap data to produce diverse mobility logs. (2) a genetic algorithm-based calibration module that fine-tunes simulation parameters to align with real-world mobility characteristics, such as daily trip counts and radius of gyration, enabling realistic behavioral modeling. (3) a data processing suite which transforms raw simulation logs into structured formats suitable for downstream applications, including model training and benchmarking, and (4) a visualization module that extracts key mobility patterns and insights from the processed datasets and presents them through intuitive visual analytics for improved interpretability.

Figures (11)

• Figure 1: Interaction between a person agent and the simulation environment for generating human mobility logs.
• Figure 2: End to end POL simulation pipeline showing parameter configuration, map generation, initialization, execution, checkpointing, and structured log generation
• Figure 3: Software architecture of HD-GEN.
• Figure 4: Deterministic checkpointing and cross platform execution of the simulation, showing how a paused run can be resumed and branched across Windows, macOS, and Linux machines to evaluate multiple scenarios from a shared state.
• Figure 5: Real time streaming architecture for POL simulation logs, showing how agent events are published during execution and consumed by external analytics frameworks through a read only streaming interface.