Genetic algorithms (GAs) provide robust yet efficient procedures to find near-optimal solutions in complex and large-scale problem spaces. Although GAs are considered among the most successful techniques for parameter optimization, current GA paradigms do not perform well in non-stationary environments. This paper presents and investigates novel operators for real-coded GAs for improving their performance in rapidly-changing, dynamic parameter optimization problems. Experimental results are given that indicate that the real-coded GA with proposed methods outperforms a standard binary-coded and a Gray-coded GA on several test problems.
展开▼