This project implements Shapley-Bid Reputation Optimized Federated Learning (SBRO-FL), a method designed to address challenges in Federated Learning (FL), including data quality heterogeneity, compensation demands, and budget constraints. The SBRO-FL approach dynamically selects clients based on their historical performance and submitted bids while ensuring fairness and maintaining system stability.

This experiment is built on top of the FLEXible platform, an open-source framework for Federated Learning (FL). FLEXible provides a highly configurable and modular environment to simulate FL scenarios, making it an ideal choice for research and experimentation in FL.

This project is built using Python and the Conda package management system. To set up the environment on your local machine, ensure that you have Conda installed.

First, clone this repository to your local machine:

git clone https://github.com/ari-dasci/S-SBRO-FL.git

The environment for this project is defined in the environment.yml file. To create a new Conda environment from this file, run the following command:

conda env create -f environment.yml

• experiment.py: This is the entry point for running the experiments. It contains the experiment parameters and path settings needed to configure and run the experiments.
• result: This folder contains the results generated from the experiments. Each subfolder corresponds to specific experiment outcomes and configurations.

To start an experiment, navigate to the experiment.py file and adjust the parameters as needed for your desired setup. Results will be saved automatically in the result folder.

This code is released under the GNU Affero General Public License v3.0 (AGPL-3.0).
You may use, modify, and distribute it under the terms of the license.

If you use this code in your project, especially in a web-based service, you must make your modified source code available as well.