A Hybrid Relay Admissibility and Security Weighting Architecture for Dynamic Adversarial Resilience in Wireless Sensor Networks

Abstract

This repository contains the complete simulation engine, canonical results, statistical analysis pipeline, and verification/audit trail supporting the manuscript submitted to Scientific Reports. The archive includes: a custom discrete-event WSN simulator implementing a BIS-weighted relay admissibility filter combined with a reinforcement-learning-based adaptive security weighting mechanism, evaluated against seven dynamic adversarial behavior types (static, Nash-equilibrium-optimal, WSNDS-calibrated, On-Off, retreating, step-change, and Q-learning-adaptive attackers); the canonical simulation dataset (1,560+ runs, deterministically seeded for full reproducibility); the statistical analysis scripts generating every table and figure in the manuscript, including Bonferroni-corrected multi-comparison testing; and a complete provenance and verification record documenting the methodological corrections made during peer review, including a corrected Nash equilibrium derivation and several results re-verified as statistically null. All simulation parameters, seeds, and dependency versions are pinned for exact reproducibility. See README.md for reproduction instructions and a guide to the repository's structure. Keywords: wireless sensor networks, trust-aware routing, adaptive adversarial behavior, reinforcement learning, Bayesian inconsistency scoring, Nash equilibrium, selective forwarding, cumulative intercepted packets