AI-Powered CPU Scheduling Simulator

Overview

This project is an advanced CPU Process Scheduling Simulator built with Flask (Python). It allows users to visualize and compare classical CPU scheduling algorithms through a web-based interface.

What sets this simulator apart is its Machine Learning pipeline. It has been specifically trained on over 122,000 empirical operating system snapshots to predict true CPU Burst Times using RandomForestRegressor, GradientBoostingRegressor, and Exponential Weighted Moving Averages (EWMA) synchronously. By stripping away stochastic OS noise, the models achieve up to 0.999 mathematically verified Test R² scores using strictly the core structural features of a process.

Features

• Algorithms Supported:
  • FCFS (First Come First Serve)
  • SJF (Shortest Job First - Non-preemptive)
  • SRTF (Shortest Remaining Time First - Preemptive SJF)
  • Priority Scheduling (Preemptive & Non-preemptive)
  • CFS (Completely Fair Scheduler - Weight based)
  • Round Robin (Configurable Time Quantum)
• Synchronous ML Prediction Workflow: Input a process's hardware telemetry and watch 3 separate AI/Heuristic models simulate and predict its Burst Time instantly side-by-side via a dynamic comparative UI.
• Dimension Reduction: Models are rigorously hyper-optimized to predict burst times accurately using only 4 deterministic features: Voluntary Context Switches, Memory %, I/O Read Bytes, and Num Threads.
• Visualization: Generates dynamic HTML/CSS Gantt charts for process execution without heavy canvas libraries.

Tech Stack

• Backend: Python, Flask, Pandas, Scikit-Learn (Random Forest, Gradient Boosting)
• Frontend: HTML5, CSS3, JavaScript (Vanilla DOM rendering)
• Logic: Custom Python implementations of ML metrics parsing and core scheduling algorithms.

Installation Setup

1. Clone the repository:

   git clone <repository-url>
   cd <project-folder>

2. Create a Virtual Environment:

   python3 -m venv .venv
   source .venv/bin/activate  # Mac/Linux
   # .venv\Scripts\activate   # Windows

3. Install Dependencies:

   pip install -r requirements.txt

How to Run

1. Start the Flask Application:

   python3 app.py

2. Open in Browser: Navigate to http://127.0.0.1:5000

API Endpoints (Machine Learning)

• Predict Burst Time (Multi-Model Synchronous):

  • POST /predict-burst
    • Evaluates CPU burst time computationally across Random Forest, Gradient Boosting, and EMMA.
    • Request body:

      {
        "name": "chrome.exe", 
        "memory_percent": 0.32, 
        "num_threads": 20, 
        "io_read_bytes": 1390784, 
        "num_ctx_switches_voluntary": 27499 
      }

    • Response:

      { 
        "rf": 1.352, "gb": 0.942, "emma": 1.250, "unit": "seconds" 
      }

• Retrain System Database:

  • POST /train-all
    • Reads process_data.csv, shuffles a strict 70-30 Train/Test split to prevent data leakage, refits the ML pipelines, and returns rigorous MAE/R² evaluation metrics.

• Model Info:

  • GET /model-info
    • Returns operational metadata about the predictors and empirical telemetry features.

Project Structure

/
├── app.py                  # Main Flask application entry point
├── requirements.txt        # Project dependencies
├── README.md               # Project documentation
├── src/
│   ├── algorithms/         # Classical scheduling algorithm functions
│   ├── ml/                 # Machine learning environment
│   │   ├── process_data.csv        # 122k row empirical telemetry dataset
│   │   ├── rf_predictor.py         # Random Forest pipeline & logic
│   │   ├── gb_predictor.py         # Gradient Boosting pipeline & logic
│   │   ├── rf_model.pkl            # Persistent RF parameters
│   │   └── gb_model.pkl            # Persistent GB parameters
│   ├── models/             # OOP Data structures (e.g. Process)
│   └── utils/              # Evaluation & helper utilities
├── static/
│   └── style.css           # UI layout and Grid/Flex styling
└── templates/
    └── index.html          # Main comparative Dashboard & Simulator bounds

---

Created as part of the AI-Powered CPU Scheduling architectural overhaul.