Simulation datasets and Python code for off-corridor excursion bus operations

Published: 11 September 2025| Version 1 | DOI: 10.17632/h2xjtc3b7s.1
Contributor:
CLAUDIO SZWARCFITER

Description

This dataset contains the full inputs, simulation code, and outputs from a study of online accept/reject policies for off-corridor bus excursions in overlapping fixed-line systems. Contents include: Input scenarios: 22 parameterized cases (16 factorial core + 6 targeted add-ons) with corridor and side-stop networks, excursion arcs, demand processes, vehicle dispatches, and policy parameters. Simulation code: Python discrete-event simulator (sim_runner.py) implementing baseline, myopic-feasible, and slack-aware policies with feasibility safeguards (excursion budgets, return windows, headway controls). Outputs: Per-run logs (events, decisions, KPIs) and aggregated performance indicators. Documentation: Metadata files and manifests to ensure reproducibility, including seed management for independent replications. The dataset and code enable replication of all reported results and can be adapted to test alternative corridor geometries, demand patterns, and policy settings.

Files

Steps to reproduce

Steps to reproduce Download and unzip the dataset. Ensure Python ≥3.9 with numpy, pandas, pyyaml, and tqdm installed. Place sim_runner.py in the same directory as the scenarios_all22.zip file (or extract the scenarios/ folder). Run the simulator from the command line, for example: python sim_runner.py --zip scenarios_all22.zip --replicates 30 --out_root runs Outputs (event logs, decisions, KPIs, metadata) will be saved under runs/<scenario>/<policy>/<replicate>/. Aggregate results using the provided scripts or by re-computing statistics from kpis.csv.

Institutions

  • Holon Institute of Technology

Categories

Artificial Intelligence, Urban Studies, Operations Research, Computer Simulation, Agent-Based Modeling, Urban Mobility, Multi-Agent System, Bus Transportation

Licence