DocumentCode :
2981225
Title :
Population initialization in the context of a biased, problem-specific mutation
Author :
Arabas, Jaroslaw ; Kozdrowski, Stanislaw
Author_Institution :
Inst. of Electron. Fundamentals, Warsaw Univ. of Technol., Poland
fYear :
1998
fDate :
4-9 May 1998
Firstpage :
769
Lastpage :
774
Abstract :
Several aspects of incorporating domain-specific knowledge into evolutionary computations are discussed. Some interesting properties are described which have been observed during deeper investigations of an algorithm for telecommunication network design. The algorithm was presented by J. Arabas (1995) and compared to other design techniques by S. Kozdrowski et al. (1997) showing comparable or better performance. In this paper, most attention is paid to the proper initialization of the population and to its interference with a knowledge-based mutation operator. Experimental results for several design problems are included to exemplify the observations
Keywords :
CAD; genetic algorithms; mathematical operators; network synthesis; telecommunication computing; telecommunication networks; biased problem-specific mutation; domain-specific knowledge; evolutionary computation; interference; knowledge-based mutation operator; performance; population initialization; telecommunication network design algorithm; Algorithm design and analysis; Data structures; Encoding; Equations; Evolutionary computation; Genetic mutations; Interference; Robustness; Telecommunication computing; Telecommunication traffic;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Evolutionary Computation Proceedings, 1998. IEEE World Congress on Computational Intelligence., The 1998 IEEE International Conference on
Conference_Location :
Anchorage, AK
Print_ISBN :
0-7803-4869-9
Type :
conf
DOI :
10.1109/ICEC.1998.700149
Filename :
700149
Link To Document :
بازگشت