Apppublic

abhi-bit-2/Genetic_Algorithm

sourceHugging Faceupdated 4mo agoView on Hugging Face

0likes

App README

Timetable Generator

This repository contains a Python-based timetable generation system for educational institutes. It leverages a Genetic Algorithm to produce optimized schedules without resource conflicts.

The system features both a desktop GUI built with tkinter and a modern web-based UI built with Gradio, backed by a FastAPI server. It is fully containerized and deployable to cloud platforms like Hugging Face Spaces.

🚀 Features

—Genetic Algorithm Optimization: Uses evolutionary techniques to resolve complex scheduling constraints.
—Web UI & API (Gradio + FastAPI): Interactive web interface for managing courses, tracking progress, and downloading schedules.
—Desktop GUI (Tkinter): Rich desktop application with frame-based navigation, real-time console redirection, and custom styling.
—Database Integration: SQLite backend for managing courses, teachers, branches, and class frequency requirements.
—Data Visualization & Export: View generated timetables in a grid format and export directly to Excel with professional formatting.

🧬 Algorithm & Core Data Structures

The application relies on highly optimized data structures to evaluate fitness and resolve scheduling conflicts rapidly.

Advanced Conflict Detection

—Sets for O(1) Lookups: Set data structures are extensively used to track unique elements and detect double-booking conflicts (teachers, rooms, branches) instantly. This $O(1)$ lookup time is critical for the fitness function, which must evaluate thousands of potential timetables during generation.

Constraint Resolution Engine

—Max Heap Implementation: A priority queue (implemented via Python's heapq) is used to optimize constraint resolution (fix_overassigned_classes and fix_class_assignments). By processing the most severe constraint violations first, the algorithm makes impactful changes early, leading to much faster convergence.
—Theoretical impact: Reduces constraint resolution time complexity from $O(m \times n)$ to $O(n \log n + m \log n)$ for $n$ classes and $m$ constraints.

Genetic Operators

—Tournament Selection: Efficient random sampling to select optimal parent schedules.
—Crossover & Mutation: Schedule slots are intelligently combined and mutated to explore the solution space effectively while maintaining valid ClassSlot objects.

🖥️ Desktop & Web Interfaces

Web Application (Cloud Ready)

The project includes a frontend.py Gradio application and an app.py FastAPI server. These allow you to generate and view schedules from any browser, making it ideal for cloud deployment.

Tkinter Desktop GUI (`modern_timetable_gui.py`)

The desktop interface provides a native experience featuring:

—Visual Design: A modern, card-based layout with consistent typography and color-coded feedback.
—Console Output Redirection: System logs and algorithm progress are streamed directly to a scrollable widget within the application using a custom StdoutRedirector.
—Workflow Optimization: Dedicated notebook tabs guide users logically through database setup, relationship configuration, class requirement specification, and timetable generation.

🐳 Docker Support & Deployment

A Dockerfile is provided for fully containerized execution. The setup uses a shell script (run.sh) to concurrently launch the FastAPI backend internally (port 8000) and expose the Gradio frontend externally (port 7860).

Building the Image

bash

docker build -t timetable-generator:latest .

Running the Container