PFMU (Fog-Based Unified Mobility Framework) is a decentralized intelligent transportation framework that jointly optimizes traffic signal control and smart parking allocation using fog computing and Multi-Agent Reinforcement Learning (MARL). The framework combines coordination-aware traffic control, congestion-aware parking allocation, and fog-level decision-making to reduce latency, improve throughput, enhance parking efficiency, and minimize network overhead. This repository provides a reproducible implementation of PFMU, benchmark baselines, scalability and robustness evaluations, ablation studies, statistical validation, and automated generation of publication-quality figures and tables for smart city and intelligent transportation system research.