This repository contains code for generating artificial data, running a task offloading process, and applying the NSGA-ii algorithm on an IoV system.
For a set of fixed tasks, the goal is to minimize their migation time and their computation time along a network of Road Side Units (RSU). These RSU are either edge servers (ES), or access points (AP).
To understand more about how the mathematical theory works, please refer to this document.
All the code is commented, what is important is to run the notebooks in the correct order.
Firstly, you must run the datagen.ipynb notebook to generate artificial data. Once that is done, you can run the offloading.ipynb notebook. That one runs the offloading process once, for a network. You can understand clearly how the process works.
The nsgaii.ipynb notebook contains the nsga2 algorithm function, which looks for a set of Pareto solutions, minimizing the max of both objective functions. It also tests multiple variables such as population size, amount of generations, etc..
The plots.ipynb notebook displays useful plots to understand the problem better, and also visualizes the pareto solutions for the best conditions found.
This folder contains all functions used in the notebooks. The NSGA-II algorithm is coded by hand, as well as the task offloading process.