Artificial Intelligence
CITYMIND
Grid-based urban AI simulation solving planning and routing via ML algorithms.
OVERVIEW
A grid-based AI simulation solving distinct urban planning and routing challenges through advanced heuristic frameworks and machine learning algorithms. Five independent AI challenge modules were integrated into a real-time full-stack visualization.
KEY HIGHLIGHTS
- 01CSP layout planning with AC-3 arc consistency + MRV/LCV heuristics
- 02D* Lite + Genetic Algorithms for dynamic real-time routing; Parallel Tempering for ambulance placement
- 03All graph algorithms hand-implemented; Flask/SocketIO backend with PixiJS real-time visualization
TECHNICAL IMPLEMENTATION
Advanced AI Algorithms
Designed solutions for 5 challenges: CSP layout planning (AC-3 arc consistency, MRV/LCV heuristics), MST + Max-Flow road optimization, Parallel Tempering (Simulated Annealing variant) for minimax ambulance placement, D* Lite + Genetic Algorithms for dynamic routing under changing constraints, and K-Means clustering + Decision Trees for crime prediction.
Full-Stack Integration
Manually implemented all graph algorithms and data structures from scratch without external algorithm libraries. Integrated a Flask/SocketIO backend serving real-time simulation state with a PixiJS frontend providing animated grid visualization. Each AI step was streamed as discrete events to the client.
TECH STACK