A NEW APPROXIMATION BASED ON THE DIFFERENTIAL EVOLUTION ALGORITHM FOR THE GAUSSIAN Q-FUNCTION


Develi I.

INTERNATIONAL JOURNAL OF INNOVATIVE COMPUTING INFORMATION AND CONTROL, cilt.8, ss.7095-7102, 2012 (SCI-Expanded) identifier identifier

Özet

This paper presents a new approach that combines two well-known approximations in order to improve the approximation accuracy of the Gaussian Q-function. Differential evolution, which can be considered as a simple and efficient evolutionary algorithm for optimizing real-valued optimization problems, is applied to estimate the parameters of the new approximation. The results produced by the presented approximation are compared with some other Q-function approaches from the literature. It is shown that the new approximation outperforms previously reported approximations and can provide an accurate representation of the Gaussian Q-function, especially for small arguments.