Učni načrt predmeta

Predmet:
Stohastične optimizacijske metode
Course:
Stochastic Optimization Methods
Študijski program in stopnja /
Study programme and level
Študijska smer /
Study field
Letnik /
Academic year
Semester /
Semester
Informacijske in komunikacijske tehnologije, 2. stopnja Inteligentni sistemi in robotika 1 2
Information and Communication Technologies, 2nd cycle Intelligent Sytems and Robotics 1 2
Vrsta predmeta / Course type
Izbirni / Elective
Univerzitetna koda predmeta / University course code:
IKT2-614
Predavanja
Lectures
Seminar
Seminar
Vaje
Tutorial
Klinične vaje
work
Druge oblike
študija
Samost. delo
Individ. work
ECTS
30 30 30 210 10

*Navedena porazdelitev ur velja, če je vpisanih vsaj 15 študentov. Drugače se obseg izvedbe kontaktnih ur sorazmerno zmanjša in prenese v samostojno delo. / This distribution of hours is valid if at least 15 students are enrolled. Otherwise the contact hours are linearly reduced and transfered to individual work.

Nosilec predmeta / Course leader:
prof. dr. Bogdan Filipič
Sodelavci / Lecturers:
Jeziki / Languages:
Predavanja / Lectures:
slovenščina, angleščina / Slovenian, English
Vaje / Tutorial:
Pogoji za vključitev v delo oz. za opravljanje študijskih obveznosti:
Prerequisites:

Zaključen študijski program prve stopnje s področja naravoslovja, tehnike ali računalništva.

Student must complete first-cycle study programmes in natural sciences, technical disciplines or computer science.

Vsebina:
Content (Syllabus outline):

Uvod: optimizacija, optimizacijski problemi, dualnost minimizacije in maksimizacije. Vrste optimizacije: eksaktna in stohastična, analitična in empirična, zvezna in diskretna, statična in dinamična ter enokriterijska in večkriterijska. Optimizacija na osnovi numeričnih modelov. Primeri optimizacijskih problemov in vzroki za njihovo zahtevnost.

Stohastična optimizacija: stohastičnost podatkov in optimizacijskih postopkov, motivacija za stohastično optimizacijo, prednosti in slabosti stohastičnih optimizacijskih metod. Enostavni stohastični metodi: naključno preiskovanje in lokalna optimizacija.

Stohastični optimizacijski algoritmi: simulirano ohlajanje. Evolucijski algoritmi: genetski algoritmi, evolucijske strategije, evolucijsko programiranje, genetsko programiranje in diferencialna evolucija. Iskanje s tabuji, optimizacija z roji delcev, optimizacija s kolonijami mravelj. Lastnosti in primerjava algoritmov, primeri uporabe.

Vrednotenje rezultatov: statistična analiza rezultatov stohastičnih algoritmov, mere učinkovitosti in predstavljanje rezultatov.
Razlike med načrtovalskimi in rutinskimi
problemi ter testnimi in realnimi problemi.

Uporabni vidiki: nastavljanje vrednosti parametrov stohastičnih optimizacijskih algoritmov, hibridizacija algoritmov, večkriterijsko optimiranje in optimiranje s subjektivnim vrednotenjem rešitev. Značilna področja uporabe in študije praktičnih primerov iz načrtovanja in
modeliranja, analize empiričnih podatkov,
časovnega razporejanja opravil in upravljanja z
viri.

Introduction: Optimization, optimization problems, duality of minimization and maximization. Types of optimization: exact and stochastic, analytical and empirical, continuous and discrete, static and dynamic, single-objective and multi-objective. Optimization based on numerical models. Examples of optimization problems and sources of their difficulty.

Stochastic optimization: Stochasticity of data and optimization procedures, motivation for stochastic optimization, advantages and disadvantages of stochastic optimization methods. Simple stochastic methods: random search and local optimization.

Stochastic optimization algorithms: Simulated annealing. Evolutionary algorithms:
genetic algorithms, evolution strategies,
evolutionary programming, genetic programming and differential evolution. Tabu search, particle swarm optimization, ant colony optimization. Characteristics of the algorithms and their comparison, examples of application.

Evaluation of results: Statistical analysis of stochastic algorithm results, performance measures and presentation of results. Differences between design and routine problems, and between synthetic test problems and real-world problems.

Applied aspects: Setting parameter values in stochastic optimization algorithms, hybridization of algorithms, multi-objective optimization and
optimization with subjective evaluation of
solutions. Typical domains of application and
practical case studies from design and modeling,
empirical data analysis, scheduling and resource
management.

Temeljna literatura in viri / Readings:

Izbrana poglavja iz naslednjih knjig: / Selected chapters from the following books:
- A. E. Eiben, and J. E. Smith. Introduction to Evolutionary Computing, 2nd edition. Springer, 2015. ISBN
978-3-662-44873-1.
- A. Kaveh. Advances in Metaheuristic Algorithms for Optimal Design of Structures. Springer, 2014. ISBN
978-3-319-05548-0.
- F. Neumann, and C. Witt. Bioinspired Computation in Combinatorial Optimization. Springer, 2010. ISBN
978-3-642-16543-6.
- G. Rozenberg, Th. Bäck, and J. N. Kok (Eds.). Handbook of Natural Computing. Springer, 2012. ISBN
978-3-540-92909-3.
- E.-G. Talbi. Metaheuristics: From Design to Implementation. Wiley, 2009. ISBN 978-0-470-27858-1.

Cilji in kompetence:
Objectives and competences:

Cilji predmeta so (a) posredovati temeljna znanja o
stohastičnih optimizacijskih metodah, (b)
predstaviti vrste stohastičnih optimizacijskih
algoritmov ter njihove prednosti in slabosti, (c)
predstaviti metodologijo vrednotenja rezultatov
stohastičnih optimizacijskih algoritmov in njihovega prilagajanja za reševanje specifičnih vrst problemov, (d) pokazati njihovo praktično
uporabnost.

Študenti, ki bodo uspešno končali ta predmet,
bodo obvladali osnove stohastične optimizacije in bodo usposobljeni za uporabo stohastičnih algoritmov v reševanju zahtevnih optimizacijskih problemov in spremljanje nadaljnjega razvoja na tem področju.

The course objectives are to (a) to give essential
knowledge on stochastic optimization methods, (b)
present the types of stochastic optimization algorithms, and their advantages and drawbacks, (c)
present the methodology of evaluating the results
of stochastic optimization algorithms and their
adaptation for solving specific types of problems, (d) show their practical potential.

The students who will successfully complete this
course will master the basics of stochastic optimization and will be capable of applying stochastic algorithms in solving demanding optimization problems and following further development in this field.

Predvideni študijski rezultati:
Intendeded learning outcomes:

Študenti bodo z uspešno opravljenimi obveznostmi tega predmeta pridobili:
- poznavanje metodologije stohastičnega optimiranja, njene uporabnosti, prednosti in slabosti,
- sposobnost identifikacije optimizacijskih problemov in prepoznavanja primernosti uporabe stohastičnih metod pri njihovem reševanju,
- sposobnost izbire ustreznega stohastičnega optimizacijskega algoritma in njegovega prilagajanja danemu problemu,
- obvladovanje uglaševanja parametrov optimizacijskih algoritmov in vrednotenja njihovih rezultatov,
- usposobljenost za samostojno reševanje zahtevnih realnih optimizacijskih problemov s stohastičnimi algoritmi.

Students successfully completing this course will acquire:
- Knowledge of stochastic optimization methodology, its applicability, strengths and weaknesses,
- Ability to identify optimization problems and recognize suitability of applying stochastic methods in solving optimization problems,
- Ability to select an appropriate stochastic optimization algorithm and adjust it to a given problem,
- Mastering the setting of optimization algorithm parameters and evaluation of their results,
- Ability to solve demanding real-world optimization problems using stochastic algorithms.

Metode poučevanja in učenja:
Learning and teaching methods:

Predavanja, seminar, konzultacije, samostojno delo

Lectures, seminar, consultations, individual work

Načini ocenjevanja:
Delež v % / Weight in %
Assesment:
Seminarska naloga
50 %
Seminar work
Pisni ali ustni izpit
50 %
Written or oral exam
Reference nosilca / Lecturer's references:
1. VODOPIJA, Aljoša, TUŠAR, Tea, FILIPIČ, Bogdan. Characterization of constrained continuous multiobjective optimization problems: A performance space perspective. IEEE Transactions on Evolutionary Computation, 2024 (early access), doi: 10.1109/TEVC.2024.3366659.
2. VODOPIJA, Aljoša, TUŠAR, Tea, FILIPIČ, Bogdan. Characterization of constrained continuous multiobjective optimization problems: A feature space perspective. Information Sciences, 2022, vol. 607, pp. 244-262, doi: 10.1016/j.ins.2022.05.106.
3. VODOPIJA, Aljoša, STORK, Jörg, BARTZ-BEIELSTEIN, Thomas, FILIPIČ, Bogdan. Elevator group control as a constrained multiobjective optimization problem. Applied Soft Computing, 2022, vol. 115, pp. 108277-1-108277-14, doi: 10.1016/j.asoc.2021.108277.
4. KOBLAR, Valentin, FILIPIČ, Bogdan. Evolutionary design of a system for online surface roughness measurements. Mathematics, 2021, vol. 9, no. 16, pp. 1904-1-1904-18, doi: 10.3390/math9161904.
5. ZUPANČIČ, Jernej, FILIPIČ, Bogdan, GAMS, Matjaž. Genetic-programming-based multi-objective optimization of strategies for home energy-management systems. Energy, 2020, vol. 203, pp. 117769-1-117769-15, doi: 10.1016/j.energy.2020.117769.