Scheduling Problem Implementation using Genetic Algorithm

Project for Nonlinear Programming and Evolutionary Algorithms course.

Revisited and updated 10.9.2024.

Overview

This repository contains an implementation of a genetic algorithm to solve the Scheduling Problem. The solution is written in Python and can be executed by running the following commands:

cd ./src
python3 ./main.py

Data

All input data required for the algorithm is stored in the data_timetable.txt file. An example of the data file format is provided below:

rooms: A, B, C, D, E
events(name, duration):
NPiEA - Predavanje, 180
Diskretna matematika - Predavanje, 120
Organizacija podataka - Predavanje, 120
NAiNS - Predavanje, 180
...

Configuration

You can customize the algorithm by adjusting constants in the ./common/constants.py (and ./main.py) file. The default values are:

POPULATION_SIZE = 600
MUTATION_RATE = 0.2
MUTATION_WIDTH = 6
GENERATIONS = 100
ELITISM_RATE = 0.01
# Set value to -1 to go through all generations.
BEST_STREAK_TO_EXIT_ON = 20
# main.py
PRINT_GENERATIONS = True

Feel free to experiment with these values to observe different behaviors of the genetic algorithm.

Requirements

Make sure you have Python 3.10.11 installed on your system.