Comparison of the Q-routing and shortest path routing algorithm
Firat Tekiner, Zabih F. Ghassemlooy, T. Srikanth · Northumbria Research Link (Northumbria University) · 2004
In this paper, we compare the self-adaptive QRouting and dual reinforcement Q-Routing algorithms with the conventional shortest path routing algorithm. The Q-Routing algorithm embeds a learning policy at every node to adapt itself to the changing network conditions, which leads to a synchronised routing information, in order to achieve a shortest delivery time. Unlike Q-Routing, the shortest path routing algorithm routes the packets based on the link with the least delay irrespective of the traffic pattern on the link. Here simulations are carried out under problematic conditions by taking into account the node and link failures. Results show that the adaptive (learning) approach performed better than the traditional non-adaptive approach under problematic conditions.