Reproducibility Materials for Closed-Loop Context-Aware Multi-Objective Network Orchestration for Adaptive Internet of Bio-Nano Things

Description

This dataset provides reproducibility materials supporting the controlled computational evaluation reported in the associated manuscript. It contains the validation configuration, controller-level summary, reported-equivalent run-level results, figure data, statistical comparison, scenario and parameter documentation, and figure-generation script. The evaluation comprises six policies (B1-B6), 20 nominal controlled runs, and 80 decision steps per run. The reported integrated policy B6 uses hysteresis and is compared with the B5 no-hysteresis policy. Under the specified benchmark scenario, B5 produced 800 configuration switches across the 20 runs and B6 produced 400, corresponding to a 50% reduction in configuration switching; fallback counts were unchanged.

Steps to reproduce

Download and extract all files in this dataset. Review README.md for the dataset structure, scope, reproducibility boundaries, and file descriptions. Use validation_parameters_and_scenarios.json to identify the computational evaluation parameters, scenario structure, controller definitions, constraints, and evaluation settings. Use validation_config.json to review the reported validation configuration, including 20 nominal computational runs and 80 decision steps per run. Use controller_summary.csv to inspect the reported aggregate outcomes for policies B1–B6, including mean decision cost, switching events, and fallback events. Use run_level_results.csv and statistical_comparison.csv to inspect the reported run-level and comparative results. Use paired_comparison.csv to examine the B5 versus B6 switching comparison. Use figure_data.csv with generate_figures.py to regenerate the supplied validation figures from the reported data. The principal reported comparison is between B5 (no hysteresis) and B6 (integrated hysteresis and fallback control): B5 produced 800 switching events and B6 produced 400 switching events across the 20 controlled runs, corresponding to a 50% reduction in configuration switching; fallback counts remained unchanged at 400. The dataset contains reproducibility materials and reported computational results. It does not constitute a complete executable reproduction of the controller because the publication-facing study does not retain all underlying implementation details, raw dynamic input traces, complete parameter vectors, or standalone controller-execution source code. Undocumented parameters are explicitly identified rather than inferred.

Institutions

• Aditya University

  Andhra Pradesh, Sūrampālem

Categories

Licence