micropurc

micropurc estimates the perturbed utility route choice (microPURC) model from trip-level data. It is data-agnostic: a network, an origin-destination demand, and a linear-in-parameters utility declared with the model-spec DSL are all that is needed to simulate data or estimate parameters. It accompanies Fosgerau, Nielsen, Rasmussen, and Yao, Estimating the perturbed utility route choice model with trip-level data.

The package is organized into subpackages:

• micropurc.core – the estimator and its numerical machinery;

• micropurc.forward – forward solvers behind the ForwardSolver protocol (PIQP by default);

• micropurc.simulation – a data-generating process and route sampler;

• micropurc.spec – a schema and Python DSL for specifying utilities.

Contents

• Getting started
  • Installation
  • Requirements
  • A first estimation
  • Troubleshooting
• Examples
  • An end-to-end pipeline
  • Specifying utilities
  • Simulating data
  • Archiving and analyzing runs
• API reference
  • Network
  • Demands
  • Configuration
  • Model specification
  • Forward solvers
  • Core estimation
  • Simulation
  • Reporting

Acknowledgments

The forward route-choice quadratic programs are solved with PIQP (Schwan et al., 2023), whose C++ sources are vendored and compiled into micropurc’s native extension under its BSD-2-Clause license.

Indices

• Index

• Module Index