Heuristics genetic algorithm using 80/20 rule

Author

Li, B. ; Jiang, W.S.

Author_Institution

Res. Inst. of Autom. Control, East China Univ. of Sci. & Technol., Shanghai, China

fYear

1996

fDate

2-6 Dec 1996

Firstpage

436

Lastpage

438

Abstract

Genetic algorithm (GA) has been widely used in optimizing difficult problems. Many achievements have been published so far. In practice, premature convergence and evolving too slowly are the common problems we often meet when a simple GA (SGA) is used. Over the years, many modifications have been suggested to alleviate the difficulties. This paper introduces an improved genetic algorithm (IGA) using 80/20 rule. SGA and the improved GA are both used to solve scheduling problems. The global optimum can not be obtained by SGA at all because of the serious premature convergence problem. When we use the improved GA, the results become much better. The global minimum or approximate global minimum can be obtained in a short time. The premature convergence problem has also been solved

Keywords

convergence of numerical methods; genetic algorithms; scheduling; search problems; 80/20 rule; approximate global minimum; convergence; heuristics genetic algorithm; optimization; scheduling problems; Automatic control; Biological cells; Convergence; Databases; Earth; Genetic algorithms; Genetic mutations; Robustness; Thumb;

fLanguage

English

Publisher

ieee

Conference_Titel

Industrial Technology, 1996. (ICIT '96), Proceedings of The IEEE International Conference on

Conference_Location

Shanghai

Print_ISBN

0-7803-3104-4

Type

conf

DOI

10.1109/ICIT.1996.601625

Filename

601625