Quick Overview
| Detail | Description |
| Course Duration | 3 days (December 5–7, 2023) |
| Delivery Format | In-person: Workshops, lectures, and guided labs |
| Level | Intermediate (basic modeling or programming background suggested) |
| Start Date | December 5, 2023 |
| Effort | 8 hours/day (total 24 hours) |
| Language | English |
| Certificate | Yes (minimum 80% attendance + project submission required) |
| Participants | 25–30 researchers, professionals, and graduate students |
Course Description
This intensive 3-day course provides practical training in agent-based modeling (ABM) using the GAMA platform, a powerful open-source environment for simulating complex socio-environmental systems.
Participants will:
- Build models such as Schelling’s segregation, traffic flow, and land use dynamics.
- Integrate GIS datasets to simulate real-world spatial scenarios.
- Use GAML (GAMA Modeling Language) to define agent behavior and environmental dynamics.
- Learn advanced techniques like model calibration, batch processing with headless mode, and hybrid modeling with mathematical equations.
- Collaborate on a group project addressing a real-world problem (e.g., flood evacuation, urban heat islands, or informal settlement growth).
- Network with global ABM experts and peers through interactive sessions and social events.
Learning Objectives
Master Core ABM Concepts:
- Understand key ABM principles (emergence, adaptation, feedback loops).
- Translate real-world phenomena into agent-environment-interaction models.
GAMA Platform Proficiency:
- Script agent logic in GAML using attributes, reflexes, actions, and state machines.
- Integrate and preprocess GIS layers (e.g., shapefiles, raster maps, OpenStreetMap).
- Use GAMA’s 3D viewer, charts, and dashboard tools for real-time analysis.
Advanced Modeling Techniques:
- Run multi-scenario simulations and automate them using batch mode and headless execution.
- Apply sensitivity analysis, model calibration, and meta-modeling techniques.
- Combine equation-based modeling (e.g., differential equations for population dynamics) with ABM.
- Deploy simulations using server mode for collaborative workflows or large-scale runs.
Key Topics:
Day 1:
Introduction to Agent-Based Modeling and GAMA.
- GAMA interface walkthrough and basic GAML syntax.
- Lab: Schelling’s segregation model (customization, visualization, extensions).
- GIS integration: Importing shapefiles, working with real-world coordinates, layering.
Day 2:
- Traffic modeling case study: Micro- and mesoscopic agent behaviors.
- Workshop: Building a multi-agent traffic model using Hanoi traffic network data.
- Headless mode: Running multiple experiments, exporting outputs, automation scripts.
- Parameter exploration: Sensitivity analysis and model calibration with CSV inputs.
Day 3:
- Hybrid modeling: Integrating system dynamics (e.g., predator-prey, SIR models) using ODEs.
- Final group project: Simulate a real-world scenario (urban resilience, public health, etc.).
- Presentations and peer feedback.
- Closing and certification.
Tools & Technologies
GAMA Platform v1.9.1
- GIS extension for raster/vector spatial data.
- 3D visualization toolkit.
- Headless mode for batch runs and scalability.
- Web-based dashboard and plotting tools.
GIS Software:
- QGIS for spatial data preprocessing (e.g., merging layers, clipping, coordinate transformations).
Scripting & Analysis:
- GAML for modeling.
- Optional Python or R for output post-processing and statistical evaluation.
3D Rendering:
- Integration with Blender or ParaView for high-quality simulation output visualization.
Target Audience
- Researchers in urban systems, social dynamics, and ecological modeling.
- Urban and transportation planners, environmental consultants.
- Academics and graduate students in computer science, engineering, or public policy.
- Data scientists and system designers interested in simulation-based analysis.
Prerequisites:
- Some familiarity with programming (e.g., Python, Java, or NetLogo) is beneficial.
- Prior exposure to GIS concepts is recommended but not mandatory.
Instructors
- Dr. Patrick Taillandier – Lead GAMA developer; expert in ABM for smart mobility systems.
- Dr. Alexis Drogoul – GAMA creator; focuses on disaster risk and large-scale societal modeling.
- Arthur Brugiere – Specializes in GIS-3D integration and visual storytelling for ABMs.
- Tri Nguyen Huu – Developers behind GAMA’s headless/server modules; emphasize scalable simulation pipelines.
- Nguyen Ngoc Doanh – Modeling and simulation: Overview.
Learning Format
- Live Lectures: Interactive, code-along sessions with real-time debugging.
- Hands-On Labs: Use Hanoi traffic and Malaysian land use data to build models from scratch.
- Group Projects: Participants form teams to design, run, and present a complete GAMA simulation.
- Peer Reviews: Exchange feedback during project presentations.
Assessment & Certification
Evaluation Criteria:
Final Project (Model + Report): 100%
- Model must demonstrate spatial integration, parameterization, and visualization.
- Report includes objective, agent structure, parameter tuning, and insights.
Certificate:
- Issued digitally by VinUniversity and the GAMA development team.
- Includes personalized feedback on final project quality and simulation design.
Contact & Further Info
- Email: ummisco.sea@gmail.com
- Course Coordinator: Do Bui Khanh Linh (linh.dbk@vinuni.edu.vn)
Tags
#AgentBasedModeling #GAMAPlatform #SimulationScience #UrbanMobility #GIS #3DVisualization #SystemThinking #EnvironmentalModeling #HeadlessSimulation
