riccardocadei / Metropolis-Hastings-for-optimization

Deploying a 5G Network in a country

Metropolis-Hastings algorithm for optimization

Problem

In telecommunications, 5G networks are the next generation of broadband cellular networks. Telecommunication companies are actively testing and starting to roll them out in different parts of the world. The task of this project is to deliver a roadmap to test this new network while optimizing the cost of the maintenance of the new installations. We propose a Metropolis-Hastings based algorithm to find an approximate solution to this problem.

More information about the problem can be found in the folder references.

Team

The project is accomplished by team Aquarium with members:

• Riccardo Cadei: @riccardocadei
• Anita Dürr: @AnitaDurr
• Loïc Busson: @loicbusson

Environment

The project has been developed and test with python3.6. The required library for running the models are numpy,pandas andscipy. The library for visualization is matplotlib.

Project structure

Istances generator

DatasetGenerator.py: generators of instances of the problem

Algorithm

helpers.py: useful functions to define the problem

markov_algos.py: Metropolis-Hasting based algorithm implementation

Notebook

main.ipynb: hyper-parameter tuning and answers to the problem

Report

references/report.pdf: a 4-pages report of the complete solution

references/pb_statement.pdf: the problem statement of this project