Using Fuzzy Adaptive Genetic Algorithm for Function Optimization

Author(s):  
Yo-Ping Huang ◽  
Yueh-Tsun Chang ◽  
Frode-Eika Sandnes
2011 ◽  
Vol 403-408 ◽  
pp. 2598-2601
Author(s):  
Lan Yao ◽  
Yu Lian Jiang ◽  
Jian Xiao

The critical operators for genetic algorithms to avoid premature and improve globe convergence is the adaptive selection of crossover probability and mutation probability. This paper proposed an improved fuzzy adaptive genetic algorithm in which the variance of population and individual fitness value are used to measure the overall population diversity and individual difference, meanwhile, both of them are applied to design fuzzy reference system for adaptively estimation of crossover probability and mutation probability. Simulation results of function optimization show that the new algorithm can converge faster and is more effective at avoiding premature convergence in comparison with standard genetic algorithm.


2012 ◽  
Vol 532-533 ◽  
pp. 1636-1639
Author(s):  
Hong Lian Shen ◽  
Feng Lin Cheng ◽  
Huan Ru Ren

A numeric method of solving nonlinear equation group is proposed. The problem of solving nonlinear equation group is equivalently changed to the problem of function optimization, and then a solution is obtained by adaptive genetic algorithm, considering it as the initial solution of Levenberg-Marquardt algorithm, a more accurate solution can be obtained, as a result, time efficiency is improved.


2018 ◽  
Vol 29 (1) ◽  
pp. 409-422 ◽  
Author(s):  
Marco Vannucci ◽  
Valentina Colla ◽  
Stefano Dettori ◽  
Vincenzo Iannino

Abstract In the industrial and manufacturing fields, many problems require tuning of the parameters of complex models by means of exploitation of empirical data. In some cases, the use of analytical methods for the determination of such parameters is not applicable; thus, heuristic methods are employed. One of the main disadvantages of these approaches is the risk of converging to “suboptimal” solutions. In this article, the use of a novel type of genetic algorithm is proposed to overcome this drawback. This approach exploits a fuzzy inference system that controls the search strategies of genetic algorithm on the basis of the real-time status of the optimization process. In this article, this method is tested on classical optimization problems and on three industrial applications that put into evidence the improvement of the capability of avoiding the local minima and the acceleration of the search process.


Sign in / Sign up

Export Citation Format

Share Document