NSGA-II Nursing Service Location Optimization in London

Project Overview

This project demonstrates two approaches for multi-objective optimization using the NSGA-II genetic algorithm to determine optimal locations for nursing service centers.
Two case studies are included:

• London — simulated data
• Shanghai — real-world data sourced from a published dataset

The Shanghai dataset originates from the paper:

Efficient methods for a bi-objective nursing home location and allocation problem: A case study
Applied Soft Computing, 2018
DOI: 10.1016/j.asoc.2018.01.014
Dataset: http://pan.baidu.com/s/1kUFD3wV

---

Purpose

This repository is used to support and reproduce our paper presented at a conference.
It does not aim to fully reproduce the referenced Shanghai study.
Instead, it is inspired by its approach and uses the dataset for demonstration purposes — showing how the NSGA-II algorithm can be implemented and adapted to solve similar nonlinear, constrained, multi-objective optimization problems in domains such as healthcare service network design, especially using Python.

---

Key Features

• Implementation of NSGA-II from scratch in MATLAB and python.
• Two regional case studies (London, Shanghai)
• Example scripts for visualization and Pareto front plotting
• Modular design for easy adaptation to other datasets

---

Note

This code is intended for research and educational use only.
The Shanghai dataset remains the intellectual property of the original authors and is used here solely for demonstration and non-commercial academic purposes.

Project Structure

1. Streamlit Demo App

• File: london-overlay-nsga-map-app.py
• Purpose: Interactive, visual demonstration of NSGA-II for nursing center placement in London.
• Features:
  • Animated genetic algorithm evolution over generations.
  • Interactive map with candidate locations and patient heatmap.
  • Slider and step button for exploring generations.
  • Designed for teaching, demo, and intuitive understanding.

Demo

---

2. NSGA-II Reproduction & Analysis

• File: nsga_reproduced.py
• Purpose: Detailed, reproducible implementation of NSGA-II for scientific analysis and paper reproduction.
• Features:
  • Loads real candidate and demand data from Excel files in data/.
  • Compares NSGA-II optimization with a baseline method.
  • Plots spatial solutions and Pareto fronts.
  • Outputs results for publication and benchmarking.

---

Requirements

Streamlit Demo

See requirements-streamlit.txt for minimal dependencies.

NSGA-II Reproduction

See requirements-nsga.txt for scientific dependencies.

Merged Requirements

If you want to install everything at once, use requirements.txt.

---

Notes

• The Streamlit app uses simulated data (random lat/lon within London's bounding box).
• The NSGA-II reproduction uses real candidate and demand data from Excel files in the data/ folder.
• Both parts use the pymoo library for multi-objective optimization.
• No external APIs (e.g., Google Maps) are used, making this easy to run offline or in classroom settings.

How to run

Streamlit Demo

• Python 3.7+
• pip

Installation

First, create a virtual environment (optional but recommended):

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

Then install dependencies from requirements-streamlit.txt or requirements.txt.

pip install -r requirements.txt

Running the App

streamlit run london-overlay-nsga-map-app.py

This will open a new browser window with the interactive map interface.

NSGA-II Reproduction

1. Install dependencies from requirements-nsga.txt or requirements.txt.
2. Ensure your Excel data files are in the data/ folder.
3. Run the script:

 python nsga_reproduced.py

Project development process / Methodology

Optional Extensions

• Include real healthcare facility locations as constraints.
• Replace the objective function with real-world service efficiency metrics.
• Extend the analysis to other cities or datasets.

License

This software is licensed for NON-COMMERCIAL use only. For commercial use, please contact prof.dr.rer.nat.lu@gmail.com.

See LICENSE for license details.

Project QR Code

Scan to visit the GitHub repository: