Previous [1] [2] [3] [4] [5] [6] [7] [8] [9] [10]

Journal of Inforamtion Science and Engineering, Vol.17 No.1, pp.35-45 (January 2001)

A Self-Organizing Genetic Algorithms
With a Eugenic Strategy

Kao-Shing Hwang, Jeng-Yih Chiou and Yuan-Pao Hsu
Department of Electrical Engineering
National Chung-Cheng University
Chiayi, Taiwan 711, R.O.C.

A Genetic Algorithm (GA) is general optimization technique suitable for solving nonlinear, multi-constraints, and combinatorial optimization problems. However, it may be slow in converging of failing to reach the global optimum. A novel strategy called Self-Organizing Eugenics Strategy (SOES) is proposed to overcome these shortcomings. In the proposed algorithm, a simplified adaptive resonance theory neural network (ART) is embedded to generate schemata, and the simulated annealing algorithm (SA) is applied to guiding the search toward an optimal solution. To illustrate its improved performance, the method is used to solve an optimization problem. Based on the results of experiments, the method demonstrates a better performance than the traditional GA and other genetic operations.

Keywords: genetic algorithm, self-organizing, eugenics strategy, adaptive resonance theory, simulated annealing algorithm

Full Text () Retrieve PDF document (200101_03.pdf : 3,563,874 bytes)

Received December 24, 1999; revised March 28, 2000; accepted June 27, 2000.
Communicated by Gen-Huey Chen.