OPPONENT´S ASSESSMENT ON DIPLOMA THESIS
Student´s name and surname: Pandiyaraj Gnanasekar
Name of the diploma thesis: Development of Evolution Algorithm for Shop Scheduling Problem Supervisor of the thesis: Ing. František Koblasa, Ph.D.
Opponent: Ing. Pavel Raška, Ph.D.
1. Diploma thesis evaluation
Evaluation excellent excellent
minus very good very good
minus good failed
Meeting the goal and fulfilling
task of the thesis X
Quality of conducted survey X
Methodology of solutions X
Expert level of the thesis X
Merit of the thesis and its
potential applicability of results X
Formal and graphic level of the
thesis X
Student´s personal approach X
Mark x in the corresponding box.
Supervisor´s final evaluation is based on his/her overall subjective evaluation.
Grading is stated literally in the article no. 5, neither by a number, nor by a letter.
2. Comments and remarks on diploma thesis:
The aim of the student´s thesis was to study the shop scheduling problem using different variants of the simple genetic algorithm. The presented master thesis is very well structured into logically related chapters, which explain the principle and the reasons for the use of a genetic algorithm for the scheduling problem to the reader in a simple way. I really appreciate the explanation of the principles of optimization algorithms by verbal description and by simple pictures.
The literature review is missing a more detailed description of types of scheduling algorithms used for the shop scheduling problem, a basic division of algorithms and a list of possible methods
designed for an optimization of the shop scheduling problem. I also lack literature references in some chapters e.g. chapter 3.7.
The principle of two-point crossing, illustrated in Figure 9 in chapter 3.5.2, is probably incorrect. The inner part of the child chromosome does not correspond to any part of the parents' chromosome.
3. Questions about diploma thesis:
What is the difference between a bit and a gene?
How many times have the optimization experiments been repeated for each setting of the optimization algorithm to avoid the effect of random behavior of the optimization?
Could a different type of evaluation be used for the evaluation of the optimization experiments instead of the arithmetic mean?
Are there any cases where you would recommend using the Modified Simple Genetic Algorithm - Operation Based Representation (MSGAOB) instead of the Modified Simple Genetic Algorithm - Job Order Representation (MSGAJO).
For what other optimization problems would you recommend using MSGA?
What other optimization methods would you recommend for the shop scheduling problem?
4. Opponent´s statement whether the diploma thesis meets the academic title requirements and whether it is recommended for defense:
The master thesis can be criticized only for some small mistakes on the formal side (occasional grammar mistakes, typos, incorrect reference to a non-existent chapter in the work). All the specified goals were met, and the thesis is very interesting with very high practical value. The thesis meets the academic title requirements and I recommend this thesis for defense.
5. Opponent´s grading:
I appreciate the use of a modified genetic algorithm for problems associated with the shop scheduling problem in the presented master thesis. These algorithms are promising for the future development of planning methods not only in the field of industrial engineering. The thesis is in my opinion worth the grade ‘very good’.
Date 17.6.2020 in Pilsen
………....
Opponent’s signature
TECHNICKÁ UNIVERZITA V LIBERCI | Fakulta strojní | Studentská 1402/2 | 461 17 Liberec 1 tel.: +420 485 35x xxx | jméno.příjmení@tul.cz | www.fs.tul.cz | IČ: 467 47 885 | DIČ: CZ 467 47 885