An intelligent, full-stack scheduling application that uses a hybrid AI solver to tackle the NP-hard problem of university course timetabling. This project finds not just a valid schedule, but an optimal one, balancing a dense network of real-world constraints.
- Interactive Web UI: A full-featured frontend built using Gemini, with vanilla HTML/CSS/JS for creating and managing courses, professors, rooms, and time slots. This acts as a platform to feature the powerful backend.
- Complex Constraint Modeling: Models the problem as a Constraint Satisfaction Problem (CSP) with a rich set of both Hard Constraints (e.g., no double-booking) and Soft Constraints (e.g., professor preferences).
- Hybrid AI Solver: Implements a novel Three-Stage Solver Architecture that combines the speed of greedy algorithms with the robustness of heuristic optimization.
- Explainable AI: Provides a clear, step-by-step output of the solving process, demonstrating the journey from a broken or suboptimal state to a valid, high-quality solution.
- Full-Stack Integration: The entire AI engine is served via a FastAPI backend, providing a robust REST API for the frontend client.
| Area | Technology |
|---|---|
| Backend | Python, FastAPI, Pydantic, Uvicorn |
| AI Core | Hill-Climbing, Simulated Annealing (Custom Implementation) |
The intelligence of this project lies in its resilient, multi-stage approach. This is not a brute-force solver; it's a sophisticated pipeline designed to mimic intelligent problem-solving.
[Random Schedule] -> [STAGE 1] --(Stuck?)--> [STAGE 2] --(Valid)--> [STAGE 3] -> [OPTIMIZED SCHEDULE]
- Algorithm: Greedy Hill-Climbing Local Search.
- Goal: Aggressively find any valid schedule as fast as possible by minimizing hard constraint violations. If it finds a zero-violation state, it succeeds and passes the result directly to Stage 3.
- Activation: This stage only runs if the fast Hill-Climbing search gets stuck in a local minimum.
- Algorithm: Simulated Annealing.
- Goal: To intelligently escape the trap. It can make "worse" moves to navigate out of the local minimum and find a path to a valid, zero-violation state.
- Activation: This final stage only operates on a schedule that has been proven 100% valid.
- Algorithm: Simulated Annealing.
- Goal: To maximize the "Happiness Score." It explores the vast space of valid schedules, intelligently trading off soft constraints to find a demonstrably superior, high-quality final result.
- Clone the Repository
git clone https://github.com/Emeralden/University-Schedule-Planner
- Install Dependencies
pip install -r requirements.txt
- Launch the Server
uvicorn backend.main:app --reload
- Open the Application
- Navigate to
http://127.0.0.1:8000/in your web browser.
- Navigate to