Spring 2024 Online Course

The material for the METROPOLIS 2 Online Course, Spring 2024, is distributed to the course's participants only. No part of this material may be reproduced or transmitted without the prior permission of Professor André de Palma. ©

Summary

This course aims at introducing the participants to the METROPOLIS2 simulator. The participants will get a general understanding of what METROPOLIS2 is able to do and how it is doing it. They will learn about the input and output data of METROPOLIS2. At the end of the course, the participants will be able to run a simulation from scratch and to analyze its output.

Organization of the course

The course is divided into 7 two-hour sessions. The sessions will take place online, via TEAMS, on Wednesdays from 5 to 7 p.m. Paris time (to cover all the timezones in Europe and America).

At the end of each session, the participants will be assigned a task. These tasks are optional. It is emphasized, however, that the best way to learn is through practice, so completing these tasks is strongly encouraged.

Pre-requisites

The participants should be comfortable working with computers as they will need to manipulate JSON, CSV and Parquet files (these formats will be introduced during Session 1). To manipulate data for large-scale simulations, knowledge of a scripting language (such as Python, R, or Julia) is strongly recommended.

The participants should also be familiar with usual concepts of transportation economics (e.g., the bottleneck model with alpha-beta-gamma preferences, discrete-choice models).

Session 1: Introduction

Date: Wednesday, March 13th

Topic: Introduction to the course, overview of METROPOLIS2 (theory, features, terminology); downloading the simulator and reading the documentation; METROPOLIS2 flow process and the input / output formats.

Task: Download and install the simulator.

Session slides (Part 1: André de Palma)

Session slides (Part 2: Lucas Javaudin)

Session 2: Supply side

Date: Wednesday, March 27th

Topic: Congestion modeling; definition of road-network data (edges and vehicle types); computing travel times; importing from OpenStreetMap.

Task: Create a road network (by hand or with a script); run some routing queries on the road network.

Session slides

Session 3: Demand side

Date: Wednesday, April 3rd

Topic: Demand modeling; definition of demand data (agents, travel alternative, road and virtual trip, utility functions); computing choices.

Task: Generate a population of agents (by hand or with a script); run a basic mode-choice model.

Session slides

Session 4: Equilibrium and convergence

Date: Wednesday, April 10th

Topic: Learning model; technical parameters of the simulation; understanding convergence.

Task: Define parameters; run a simulation.

Session slides

Session 5: Simulation output

Date: Wednesday, April 17th

Topic: Understanding the summary file; reading and understanding the dis-aggregated output data (agents, legs, routes, network conditions).

Task: Create a graph from the results of the simulation run.

Session slides

Session 6: Input side of Paris' simulation

Date: Wednesday, May 22nd

Topic: Generation of METROPOLIS2 input to simulate Paris' urban area, with a focus on the use of synthetic population and the computation of public-transit travel times.

Session slides

Session 7: Output side of Paris' simulation

Date: Wednesday, June 5th

Topic: Interpretation of METROPOLIS2 output from the Paris' urban area simulations, with a focus on the computation of CO2 emissions, fuel consumption and air pollution and the policy evaluation of the low emission zone.

Session slides (Part 1: Romuald Le Frioux)

Session slides (Part 2: Lucas Javaudin)