A hybrid framework for pedestrian simulation that allows pedestrians to dynamically switch models based on density-based zones in the environment.
Workflow:
Prerequisites:
- GAMA 1.8.2 (https://gama-platform.org/wiki/OlderVersions)
- Python:
- Install packages in calibration_python: pip install -r requirements.txt
How to run calibration:
- Prepare simulation in GAMA headless mode
- Copy all folders in calibration_python/headless into headless folder in GAMA installation folder.
- Set the path for gama_headless_folder variable in GA_calibration_parallel.py
- Set the number of processes for parallel execution in GA_calibration_parallel.py.
- In calibration_python folder, run: python GA_calibration_parallel.py --population_size 128 --num_generations 10000 --mutation_rate 0.01
How to run local sensitivity analysis:
- In calibration_python folder, run: python local_sensitivity_analysis.py
Cite this work: Dang H-T, Gaudou B, Verstaevel N. HyPedSim: A Multi-Level Crowd-Simulation Framework—Methodology, Calibration, and Validation. Sensors. 2024; 24(5):1639. https://doi.org/10.3390/s24051639