FactoryMind

A Python-based production planning system for optimizing manufacturing operations. The system uses Mixed-Integer Linear Programming to minimize makespan while respecting machine capabilities, material availability, and operational constraints.

Features

• Production Optimization: MILP-based task assignment to minimize total production time
• Energy Consumption: Calculates energy usage based on machine power profiles (idle, loading, production)
• Realistic Time Modeling: Accounts for machine working hours, (un)loading time, batch capacities
• Data Generation: Built-in random factory data generator for testing and simulation
• Yield Rate Handling: Considers production waste and calculates actual quantities needed

Project Structure

project_work/
├── src/
│   ├── data/              # Static production data (materials, recipes, machines, orders)
│   ├── entities/          # Domain models (Machine, Recipe, Order, etc.)
│   ├── generator/         # Random data generator
│   ├── loader/            # Data loading and validation
│   ├── planner/           # Production planning and optimization
│   ├── schemas.py         # Pydantic validation schemas
│   └── main.py            # Entry point
└── README.md

Installation

Prerequisites

• Python 3.12+
• pip

Setup

1. Create and activate a virtual environment (recommended):

python3 -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

2. Install dependencies:

pip install -r src/requirements.txt

Usage

Run with static data

Use data from src/data/:

python3 -m src.main

Generate random data

Generate random factory data and run the planner:

python3 -m src.main -g

Or specify the number of entities to generate:

python3 -m src.main -g --materials 10 --recipes 5 --machines 3 --orders 2 --min-quantity 200

Command-line Options

• -g, --generate: Generate random data before running the planner
• --materials N: Number of materials to generate (default: random 5-15)
• --recipes N: Number of recipes to generate (default: random 3-10)
• --machines N: Number of machines to generate (default: random 2-5)
• --orders N: Number of orders to generate (default: random 1-3)
• --min-quantity X: Minimum quantity for each recipe in orders (default: 50 for pieces, 5.0 for kg/L)

Output

The planner produces a production schedule showing:

• Production assignments: Which machine produces which recipe
• Time estimates: Individual task times and overall makespan
• Energy consumption: KWh consumption per task and total
• Material validation: Stock availability checks

Example output:

PRODUCTION PLANNING
================================================================================

MACHINE: Packaging Unit A
--------------------------------------------------------------------------------
    RECIPE: Butter Biscuits
     - Quantity: 1500 pieces
     - Time: 1068.00 seconds (17.80 minutes)
     - Energy: 2739.6000 KWh

================================================================================
TOTAL ENERGY CONSUMPTION: 2739.6000 KWh
TOTAL WORK TIME (all machines): 1068.00 seconds (17.80 minutes)
MAKESPAN (calendar time): 1068.00 seconds (17.80 minutes)
================================================================================

Key Concepts

• Makespan: Total calendar time to complete all production (max of all machine times)
• Yield Rate: Percentage of good output vs. raw production (affects material requirements)
• Batch Capacity: Maximum quantity a machine can process in one batch
• Power Profile: Machine energy consumption at different states (idle, loading, producing)
• Max Working Hours: Daily operational limit for each machine

Dependencies

• pydantic: Data validation and settings management
• pulp: Linear programming and optimization
• Additional dependencies listed in src/requirements.txt