Reliability-Based Grain Harvest Transport Fleet Sizing: Reproducibility Package

This reproducibility package supports the study “Beyond average capacity matching: Field-calibrated reliability-based fleet sizing for grain harvest transport under uncertainty”.

The archive contains the complete Python implementation of the mechanistic harvest-transport model, Monte Carlo reliability analysis, parametric-bootstrap assessment of parameter uncertainty, sensitivity and robustness analyses, derived numerical outputs, and publication figures. The workflow reproduces the reported reliability-based fleet-size thresholds for direct field-to-receiver grain haulage under stochastic operating conditions.

The model uses field-derived calibration summaries for harvester operating speed, grain-truck road speed, and unloading duration, with n = 200 observations for each stochastic input. Historical raw GPS and receiver-timing records are not redistributed because they are not contained in the redistributable source archive, and no reconstructed observations are presented as raw data.

Version 1.0.0 was implemented in Python 3.13.5. The archived run uses a fixed primary random seed of 20260915, 500,000 operational Monte Carlo realisations, and 500 parametric-bootstrap refits with 10,000 inner realisations per refit. The package includes pinned environment specifications, integrity hashes, derived CSV outputs, and all quantitative figures generated by the analysis script.