🧠 Location Optimizer Backend (Flask API)

This is the Flask backend for the Location Optimization Tool, a geospatial decision support system. It provides RESTful APIs to solve:

• The P-Median problem for facility location optimization
• The Traveling Salesman Problem (TSP) using a Genetic Algorithm (GA)

The backend is designed to work with the React frontend deployed here and supports both synchronous and asynchronous operations.

---

⚙️ Tech Stack

• Flask for API framework
• PuLP for solving the P-Median Linear Program
• NumPy, Pandas for computation
• Gunicorn for production serving
• Flask-CORS for cross-origin communication
• dotenv for environment configuration

---

🚀 Deployment (Render)

• Backend URL: https://operations-research-project.onrender.com

---

📡 API Endpoints

🏭 POST /solve/p-median-coords

Solves the P-Median problem for selecting p facility locations given:

• A list of coordinates
• Fixed facility costs
• Value of p

📥 Request Body:

{
  "coordinates": [[lat1, lon1], [lat2, lon2], ...],
  "fixed_costs": [100, 200, 150, ...],
  "p_val": 3
}

📤 Response:

{
  "selected_facilities": [false, true, false, true],
  "total_cost": 374.52,
  "status": "Optimal"
}

🧠 Logic:

• Computes all pairwise distances using the Haversine formula
• Solves a Mixed-Integer Linear Program using PuLP

⚠️ Limitations:

• Assumes all points can be both clients and candidate facilities
• Returns [] and Infinity if LP is infeasible or not optimal

🧬 POST /solve/ga

Starts an asynchronous Genetic Algorithm to solve a TSP over the provided coordinates.

📥 Request Body:

{
  "coordinates": [[lat1, lon1], [lat2, lon2], ...],
  "generations": 1000,
  "pop_size": 200,
  "elite_size": 10,
  "mutation_rate": 0.01
}

📤 Response (immediate):

{
  "job_id": "abcd-1234"
}

⏳ GET /check-status/<job_id>

📤 Response:

{
  "status": "completed",
  "result": {
    "route_indices": [1, 4, 3, 2, 1],
    "route": [[...], [...], ...],
    "total_distance": 123.45,
    "computation_time": 2.34,
    "status": "success",
    "log": [
      {"generation": 1, "best_distance": 700.3},
      ...
    ]
  }
}

🧠 GA Logic:

• Uses real-world distance via the Haversine formula
• Implements selection, crossover (OX/CX), and mutation (swap/inversion)
• Logs progress over generations

⚠️ Limitations:

• No persistent storage: results vanish after app restarts
• Route indices assume original coordinate order
• Stochastic: may produce different solutions across runs

---

🧪 Local Setup

1. Clone & Install

git clone https://github.com/your-username/location-optimizer-backend.git
cd location-optimizer-backend
pip install -r requirements.txt

2. Configure Environment

Create a .env file:

FLASK_ENV=development
ALLOWED_ORIGINS=http://localhost:5173,http://localhost:3000

3. Run Locally

python -m src.api.routes

Or for production:

gunicorn src.api.routes:app

---

🧾 File Structure

├── src/
│   ├── api/
│   │   └── routes.py        # Main Flask routes & job manager
│   └── core/
│       ├── lp_solver.py     # P-Median optimizer using PuLP
│       └── ga_solver.py     # TSP optimizer using Genetic Algorithm
├── requirements.txt
└── .env

---

📌 Requirements

flask>=2.0.0
flask-cors>=3.0.10
pulp>=2.0.0
numpy>=1.19.0
pandas>=1.0.0
gunicorn==21.2.0
python-dotenv