Solving TSP by Transiently Chaotic Neural Networks
Shyan-Shiou Chen, Chih-Wen Shih · InTech eBooks · 2008
It has been more than two decades since artificial neural networks were employed to solve TSP. Among the efforts in improving the performance of this computational scheme, substantial achievements have been made in incorporating chaos into the system and developing mathematical analysis for finding the parameters in the chaotic regime and convergent regime. There are several advantages in employing the piecewise linear activation function. We have observed that the TCNN with piecewise linear activation function has better performance than with the logistic activation function in the