This repository contains the all code produced for the course 'Genetic and Evolutionary Computing' (H02D1a) at KU Leuven. The goal of this course is to familiarize the students with genetic algorithms and for students to be able to produce their own algorithm. In this repository the asymmetric travelling salesman (ATSP) problem has beel 'solved' using a evolutionary algorithm. Afterwards the strenghts and weaknesses have been assessed in a report. The scoring for this course is based fully on the project. Twelve of the twenty points are earned by the algorithm itself based on the performance on test problems relative to the other students (peers). The other eight points are earned during the oral examination in which the professor asks some questions about your implementation and some basic questions about evolutionary algorithms.
The assignments limits the algorithm in two ways: all code should be written in a single file (r0701014.py) and you are not allowed to use more than two cores (this is not a course parallel computing). A very basic algorithm is implemented during the group phase in the exercise sessions. This algorithm is expanded upon during the individual phase. Some test problems (the csv files) are provided in order to assess the performance of the algorithm. Note that the examination uses different unseen test problems in order to avoid overfitting. In order to fully understand all the functions and design decisions, it is recommend to read the report and the papers in the bibliography.
The report started from a template with example questions that could be answered in each section. These questions did not need to be answered but they could give a guideline. The hard limit was 10 pages.
The algorithm, the report and the oral examination together scored 19/20.