Research Internship for (Under)Graduate Students
One of the research topics was Developing metaheuristic algorithms
for optimization problems
Short description: The main research topics are directed towards the
development of mathematical models and (meta)heuristic optimization methods
for various world-known optimization problems (optimization on graphs,
scheduling, transportation, location, etc). Beside the application of
different general purpose exact solution methods (CPLEX, Gurobi, LINGO,
etc.), problem specific exact and heuristic algorithms will be developed.
Although working with various metaheuristic methods, we particularly promote
the ones developed by Serbian researchers: Variable Neighborhood Search
(VNS) and Bee Colony Optimization (BCO). In addition, our current research
project investigates parallelization, theoretical and empirical evaluation
of metaheuristics. Our interest is also directed towards the integration of
Artificial Intelligence (AI) and optimization methods to deal with real-life
optimization problems that occur in science and industry.
Supervisors: Tatjana Davidović,
tanjad@mi.sanu.ac.rs; Dragan Urošević
draganu@mi.sanu.ac.rs; and Tatjana Jakšić-Krüger
tatjana@mi.sanu.ac.rs;
Prerequisites: Good programming skills, C(C++), Java, Python.
Suggested material:
Presentation of the obtained results at
Seminar on Computer Science and Applied Mathematics is available
here.
The students and their supervisors would like to thank the professors
included in this internship: accademician Dragoš Cvetković, and Nataša
Milosavljević. Your valuable help made our work easier and more successful.
from the Operations research and
management science discipline.
Dušan Ramljak, Penn State University, Great Valey, Philadelphia, USA;
Raka Jovanović, Qatar Environment and Energy Research Institute,
Hamad Bin Khalifa University, Doha, Qatar;
Milica Anđelić, Department of Mathematics, Kuwait University;
Slobodan Jelić, Faculty of Civil Engineering, University of Belgrade,
Serbia.
Subject: Variable neighborhood Search for
Spectral Reconstruction of Graphs (in Serbian).
Subject: Bee Colony Optimization for
Training of Artificial Neural networks (in Serbian).
Subject: Automatic Tunig of Parameters for
Metaheuristic Methods (Manual for irace in Serbian).
Subject: Ant Colony Optimization for Clustering
Incomplete Data (in Serbian).
Subject: Bee Colony Optimization for
Feature Selection Problem.
Subject: Metaheuristic Approach to Spectral
Reconstruction of Graphs (co-authored with P. Ćirković, P. Đorđević, and
T. Davidović, published in the Proceedings of MOTOR 2022 conference).
Subject: Comparative Analysis of
Heuristic Approaches to P||Cmax, Proc. 11th International Conference
on Operations Research and Enterprise Systems, ICORES 2022 (virtual),
Feb. 3-5, 2022, pp. 259-266.
DOI: 10.5220/0011008500003117 (co-authored with T. Davidović, T.
Jakšić-Krüger, and D. Ramljak).
Subject: Difficulty
Esimation of Combinatorial Optimization Problems using Machine Learning
(co-authored with D. Ramljak, T. Jakšić-Krüger, T. Davidović,
D. Ostojić, and A. Haridas, published in Symmetry 2023, 15(1),
140:1-32,
DOI: 10.3390/sym15010140).
Subject: Lower Bounds for P||Cmax, (in preparation,
co-authored with D. Ostojić, T. Jakšić-Krüger, T. Davidović, and D. Ramljak).
Subject: Neighborhood Management in
Variable Intensity Neighborhood Search, (presented at SMSCG 2023, co-authored with
T. Davidović).
Marko Milenković, Faculty of Science and Mathematics,
University of Niš
Subject: General Variable neighborhood Search for
Spectral Reconstruction of Graphs (co-authored with P. Đorđević, M.
Anđelić, and T. Davidović, presented at EURO 2024 conference, paper in
preparation).
Subject: Maximum Diversity Problem with Capacity and Budget Constraints
Publications: Bee Colony Optimization for Maximum Diversity
Problem with Capacity and Budget Constraints, 4th International Conference
on Smart Grid and Renewable Energy (SGRE 2024), Doha, Qatar, Jan. 08-10,
2024, pp. 1-6.
DOI: 10.1109/SGRE59715.2024.10428871 and
General Variable Neighborhood Search for Maximum
Diversity Problem with Capacity and Budget Constraints, (under
review, co-authored with D. Urošević, T. Davidović, and R. Jovanović).
Subject: A Framework for Proof-of-Useful-Work Consensus Protocol (co-authored
with D. Ramljak, B. Sharma, M. Todorović, T. Davidović, published in Proc.
2nd Blockchain and Cryptocurrency Conference (B2C' 2023), Corfu,
Greece, Oct. 18-20, 2023, pp. 68-71.).
Subject: Heuristics for P||Cmax Scheduling problem (co-authored
with D. Ostojić, D. Ramljak, T. Jakšić-Kruger, and T. Davidović, paper in
preparation).
Subject: Mashine Learning and Metaheuristics (supervised by T. Jakšić-Kruger
and T. Davidović).
Subject: Difficulty Estimation for P||Cmax problem instances (so-authored
with S. Jelić, T. Jakšić-Kruger, and T. Davidović, presented at Artificial
Intelligence conference, Belgrade, Dec. 26-27, 2023, paper in preparation).
Subject: Minimum Steiner Tree Problem (supervised by S. Jelić, R. Jovanović,
D. Urošević, and T. Davidović).
Subject: Fixed Set Search Applied to the Max-Cut Problem,
(co-authored with R. Jovanović, D. Urošević, and T. Davidović, published in
Proc. IEEE 8th International Energy Conference (ENERGYCON 2024), Doha,
Qatar, March 04-07, 2024, pp. 1-6.
DOI: 10.1109/ENERGYCON58629.2024.10488777).
Radoš Jojić, Faculty of Mathematics, University of Belgrade, and
Natalija Ranđelović, Faculty of Science and Mathematics,
University of Niš
Subject: Sustainability and fairness of Blockchain Consensus
Protocol
(co-authored with D. Ramljak, M. Todorović, and T. Davidović).
Presentations:
Balancing Efficiency and Fairness in Blockchain: A
Proof-Of-Useful-Work-Based Blockchain Consensus Protocol, Artificial
Intelligence Conference, Dec. 26-27, 2024, Belgrade, Serbia
Aleksa Đorđević, Oxford University
Subject: Price-Collecting Group Steiner Tree Problem
(co-authored with S. Jelić, R. Jovanović, D. Urošević, and T. Davidović),
paper in preparation).
Subject: Colored Travelling Salesmen Problem, (
co-authored with D. Urošević, R. Jovanović, and T. Davidović, paper in
preparation).