Title :
Simulation Research Based on an Improved Genetic Algorithm
Author :
Jiang Jing ; Tan, Boxue ; Meng, Lidong ; Jiang, Jing
Author_Institution :
Sch. of Electr. & Electron. Eng., Shandong Univ. of Technol., Zibo, China
Abstract :
Premature convergence is the main obstacle to the application of genetic algorithm. This paper makes improvement on traditional genetic algorithm by linear scale transformation of fitness function, using self-adaptive crossover and mutation probability and adopting close relative breeding avoidance method. Simulation results show that the improved algorithm outperforms traditional genetic algorithm in terms of convergent speed and the ability to find a global optimum.
Keywords :
convergence; genetic algorithms; neural nets; probability; close relative breeding avoidance method; fitness function; genetic algorithm; linear scale transformation; premature convergence; self adaptive crossover probability; self adaptive mutation probability; Artificial neural networks; Convergence; Equations; Gallium; Genetic algorithms; Genetics; Optimization; close relative breeding avoidance; fitness function; genetic algorithm; self-adaptive;
Conference_Titel :
Intelligence Information Processing and Trusted Computing (IPTC), 2010 International Symposium on
Conference_Location :
Huanggang
Print_ISBN :
978-1-4244-8148-4
Electronic_ISBN :
978-0-7695-4196-9
DOI :
10.1109/IPTC.2010.76
