Project Goals 

Preparing optimal examination timetables is extremely important for the universities, students, and teachers. Nevertheless, this is a complex task to be done manually as many constraints, regarding students, teachers, and facilities, need to be considered. Therefore, it is quite interesting, from the scientific perspective, as well as from practical point of view at the universities, to look at how such timetables could be generated in an automated manner.

The objective of this project is to precisely deal with this challenge, by developing new optimisation methods based on metaheuristics and artificial intelligence techniques for the problem of examination timetabling at the Faculty of Electrical and Computer Engineering, University of Prishtina, as well as for the existing variants of the examination problems in the literature. We aim to develop hybrid optimisation methods and automated algorithm selection for this envisioned problem. In addition, we shall develop a protype implementation for the examination timetabling at the Faculty of Electrical and Computer Engineering, University of Prishtina.

The project will contribute to the area of examination timetabling and will enable to boost the cooperation between University of Prishtina, University of Sarajevo and TU Wien. In addition, this project will represent a huge opportunity to young Kosovar and Bosnian researchers to improve their research background in the fields of artificial intelligence, optimisation, and automated timetabling.

Project duration

July 2021 - May 2023

Team 

TU Wien: Technische Universität Wien

• Nysret Musliu (project coordinator)

Universiteti i Prishtinës (University of Prishtina)

• Kadri Sylejmani (team leader)

• Labeat Arbneshi (university assistant, Research topic: Adding constraints of exam administrators to the examination timetabling problem)

Master students 

• Fjolla Beqiri (Master thesis: Ant Colony Optimization for examination timetabling problem)

• Synim Selimi (Master student: Iterated Local Search for the examination timetabling problem with constructive-based initial solution)

• Mihrije Kadriu (Master thesis: Developing a system for automated examination timetabling based on local search algorithms; Bachelor thesis: Developing a web tool for validation of the solutions for the examination timetabling problem)

• Ardit Baloku (Master student: A memetic algorithm approach for the relaxed version of examination timetabling problem) 

Bachelor students

• Arita Alidemaj (Bachelor thesis: Building a prototypal web based system for examination timetabling by utilizing state of the art algorithms)

• Elsa Nishori (Bachelor thesis: Analysis of solutions for examination timetabling problem using grouping algorithms)

• Drinor Ahmeti (Bachelor thesis: Developing a web tool for pre-processing real-life test instances for the examination timetabling problem)

• Durajet Mustafa (Bachelor thesis: Developing a software tool for examination timetabling at the Faculty of Electrical and Computer Engineering - University of Prishtina)

• Betim Thaqi (Bachelor thesis: Developing an Integer Linear Programming model for the examination timetabling problem at the Faculty of Electrical and Computer Engineering - University of Prishtina)

• Enis Berisha (Bachelor thesis: Building a web app for visualizing the quality of the solutions for the examination timetabling problem)

• Valtrin Shala (Bachelor thesis: Developing a web tool for aggregation of test data for the examination timetabling)

• Shpat Gashi (Bachelor thesis: Automated examination timetabling via Simulated Annealing)

• Laurent Arifaj (Bachelor thesis: A Hybrid Local Search-Based Approach for Solving the Relaxed Version of Exam Timetalbing Problem)

• Rina Lutolli (Bachelor thesis: Developing a web application for tuning the parameters of the algorithms for automated examination timetabling)

• Esma Krasniqi (Developing a web tool for pre-processing real-life test data of the examination timetabling problem at the Faculty of Electrical and Computer Engineering)

• Kushtrim Haziri (Bachelor thesis: Developing a web application for prepossessing the dataset of the examination timetabling problem)

• Kreshnik Malaj (Bachelor thesis: Developing a web tool for visualization of the solutions for the examination timetabling problem)

• Rrezarta Mustafa (Bachelor thesis: Developing test scripts for unit testing of the algorithm for the examination timetabling problem)

University of Sarajevo

Maida Islamagić (teaching assistant)

Nejra Selimović (teaching assistant)

Collaborators

University of Udine

Andrea Schaerf (Full Professor)

Publications 

Synim Selimi, Labeat Arbneshi, Kadri Sylejmani and Nysret Musliu, Iterated Local Search for the examination timetabling problem with constructive-based initial solution, Proceedings of the 13th International Conference on the Practice and Theory of Automated Timetabling - PATAT 2022: Volume III (2022), ISBN: 978-0-9929984-5-5. [View paper]

Developed software

• A solver for the examination timetabling problem using Iterated Local Search metaheuristic

• A prototypal system for examination timetabling

• An Integer Linear Programming Model for the examination timetabling problem at the Faculty of Electrical and Computer Engineering, University of Prishtina 

• A memetic algorithm solver for the relaxed version of the examination timetabling problem

• A Simulated Annealing Approach for the examination timetabling problem

• A hybrid meta-heuristic solver combining Variable Neighbourhood Search and Simulated Annealing for the relaxed version of the examination timetabling problem

• A solver incorporating the number theory, probability, and greedy techniques for the Examination Timetabling Problem

• A tool on the web that aggregates input data and converts it into a CSV format

• A web-based examination timetable validator 

• A web tool for examination timetabling by using the state of the art solver based on MiniZinc

• Examination timetable creator for the case of Faculty of Electrical and Computer Engineering

• A web tool for visualization of the solution constraint

Test instances

• Test set with the details for exam administrators (generated based on the original test set created by Battistutt et al. 2020)

• Test set with some relaxed hard and soft constraints (generated based on the original test set created by Battistutt et al. 2020)

Solution results

...