| Recommend this page to a friend! |
| Author | ||
![]() |
|
Innovation award![]() Nominee: 1x | |||||||||||||||||||
| Classes of dan boorn | > | ANN Router | > | Download .zip .tar.gz | > | > | > | |||||
|
|||||||||||||||||||||||
| Groups | Applications | |||||||
| Groups | ||
| Networking | Networking protocols, clients and servers | View top rated classes | |
| Algorithms | Numerical and statistical algorithms | View top rated classes | |
| Artificial intelligence | Automation of tasks using human-like intelligence | View top rated classes |
| Innovation Award | ||
![]() April 2006 Number 3 Prize: One book of choice by O'Reilly |
Finding the fastest route between two nodes of a network is a difficult problem to solve using deterministic algorithms. An alternative means to solve this problem is to use fuzzy logic solutions. One kind of fuzzy logic solution consists on training neural networks to evaluate multiple routes, and then pick the route which yields the best results after a significant training period. This class simulates a network router using neural networks to find the fastest gateway port to deliver packets to the final network destination. Despite the routing solution that is evaluated on each moment may not be necessarily the best, continued training a neural networking may not only improve the solution over time, but may also automatically solve the problems caused by temporary or permanent connectivity interruptions in other routers in the way to the destination of each packet. Manuel Lemos |
| User ratings | ||
| There are not enough user ratings to display for this class. |
| Applications that use this class | ||
| Download all files: annrouter.tar.gz annrouter.zip |