Research data for: An Adaptive Real-Time Decision-Support Framework for Dynamic School Transportation Reoptimization
Description
This repository provides the reproducible research package associated with the study “An Adaptive Real-Time Decision-Support Framework for Dynamic School Transportation Reoptimization.” The package contains the source code, benchmark instances, dynamic operational scenarios, experiment scripts, raw outputs, and processed results used to evaluate the Adaptive Real-Time Hybrid Genetic Search framework (ARTHGS) for the Dynamic School Transportation Reoptimization Problem (DSTRP). The DSTRP extends classical school bus routing to dynamic rolling-horizon operations with time-dependent travel times, hard-committed route segments, configurable transportation policies, and operational modification costs. The repository includes implementations of the main methods evaluated in the paper: No-Reopt-HGS, Greedy repair, ALNS, Reactive-ALNS-RH, ARTHGS, ARTHGS-GR, and ablated ARTHGS variants. It also includes the rolling-horizon CPLEX validation scripts, capacity-sensitivity experiments, and statistical analysis scripts used to generate the validation results reported in the revised manuscript. The benchmark data cover three school transportation policy families: single-school transportation, mixed-load transportation, and sequential multi-school transportation. The dynamic operational scenarios include static, mild, moderate, severe, and cascading conditions, with event streams representing traffic perturbations, vehicle delays, school-time perturbations, and localized route disruptions. The repository is organized to support reproducibility. It provides C++ source code for the optimization engine perturbations, vehicle delays, school-time perturbations, and localized route disruptions. The repository is organized to support reproducibility. It provides C++, Python scripts for experiment execution and result post-processing, benchmark instance files, scenario files, raw computational outputs, processed tables, and documentation explaining how to reproduce the main experiments and validation analyses.