Agile Methodologies Course: AGH University of Science and Technology, Krakow, Poland
Urban street lighting is a critical pillar of public safety and infrastructure. However, static systems often result in excessive energy consumption and high operational costs. This repository presents a Hybrid Artificial Intelligence framework designed to transform conventional lighting into a context-aware, energy-efficient, and safety-critical system.
The project integrates data-driven predictive modeling with interpretable fuzzy reasoning and multi-objective evolutionary optimization. By simulating a professional workspace environment using Jira-Scrum and GitHub, this repository demonstrates a robust pipeline for smart city infrastructure development.
├── Notebook/
│ └── SmartStreetLighting.ipynb # Core implementation: regression, fuzzy logic, and optimization.
├── ResearchPaper/
│ ├── assets/ # Figures, plots, and experimental data visualizations.
│ ├── Main_Report.tex # Comprehensive technical report (Full Version).
│ ├── Short_Report.tex # Executive summary/workshop submission (Shortened).
| └── SmartStreetLighting.pdf # Compiled report of the shortened version
├── AI-StudentProjects-A.pdf # Supplemental technical guidelines.
├── AI-StudentProjects-B.pdf # Supplemental technical guidelines.
├── GECCO-StudentWorkshop2025.pdf # Award-winning workshop submission (GECCO 2025) used as a guidline.
├── LICENSE # MIT License documentation.
└── README.md # Project documentation.
The framework employs a three-layer architecture to ensure both performance and transparency:
- Predictive Analysis: A data-driven regression model forecasts short-term lighting demand by processing heterogeneous sensor inputs, including ambient light levels, occupancy rates, and traffic density.
- Interpretable Reasoning: A fuzzy-rule-based expert layer embeds domain-specific safety logic. This ensures reliable illumination under uncertain or extreme environmental conditions where purely data-driven models might fail.
- Multi-Objective Optimization: Fuzzy membership functions and control thresholds are tuned using a multi-objective evolutionary algorithm. This produces a Pareto-optimal set of solutions that balance energy conservation against predicted safety risks.
- Energy Efficiency: Achieves substantial reduction in municipal energy consumption compared to static or reactive systems.
- Operational Safety: Maintains high safety standards through rule-based logic that prevents under-illumination in high-traffic or high-risk periods.
- Explainability (XAI): Unlike "black-box" models, the fuzzy reasoning layer provides human-readable decision paths, facilitating municipal trust and regulatory compliance.
- Python 3.x
- Jupyter Notebook / JupyterLab
- LaTeX Environment (for compiling research papers)
- Clone the repository: git clone https://github.com/owoMarciN/AgileProject.git
- Navigate to the
Notebook/directory to explore the implementation. - Open
SmartStreetLighting.ipynbto view the data preprocessing, model training, and optimization results.
Distributed under the MIT License. See LICENSE for more information.
© 2026 AgileProject Team - Smart City Infrastructure Division.