| Title |
Hibridizuotų genetinių algoritmų parametrų optimizavimo tyrimas / |
| Translation of Title |
Optimization of control parameters for hybridized genetic algorithms investigation. |
| Authors |
Mekionytė, Jurgita |
| Full Text |
|
| Pages |
46 |
| Keywords [eng] |
cutting stock problem ; genetic method ; Bottom Left ; Bottom Left Fill ; Bayesian approach |
| Abstract [eng] |
Cutting stock problem (CSP) is a relevant area of construction and stitching industries. Due to the combinatorial complexity of a problem, it is impossible to solve it exactly in every instance. The objective of the master thesis is to find an approximate solution of this problem. The problem was solved by employing the meta-heuristic genetic method with additional heuristics Bottom Left Fill (BLF) and Bottom Left (BL). To improve the efficiency of this method the Bayesian approach (BA) was applied for parameters optimization. This is a the novelty of the work. The methodology of solving cutting stock problems and optimization of control parameters for the proposed version of genetic algorithms was formulated and described. A comparative analysis of parameters optimzation was performed. The results of the comparative analysis and their application for genetic methds is the main part this master thesis. Thesis also includes a discussion about the development of software ot the optimal cutting stock problem. The creation of initial data is described. Recommendations for the further development of the program are considered as well. |
| Type |
Master thesis |
| Language |
Lithuanian |
| Publication date |
2011 |
