User‐Driven Routing Algorithm Application for CDN Flow
Abdelhamid Mellouk, Saïd Hoceini, Hai Anh Tran · 2013
This chapter presents a new quality of experience (QoE)-based routing algorithm, called QoE QLearning-based Adaptive Routing, which is based on a bio-inspired mechanism and uses the Q-learning approach, an algorithm of reinforcement learning. The chapter briefly presents the knowledge base of the Reinforcement Learning and Q-routing. Next, it focuses on the routing algorithm proposal. The chapter explains routing algorithm, called QQAR protocol. It surveys the reinforcement learning theory and then describes the QQAR algorithm by mapping the RL model onto the routing model. A section deals with the mathematical model of RL in detail. The chapter discusses the attempt of applying the Q-learning algorithm to adaptive routing protocol, called Q-routing. The goal of Q-routing is to balance the exploitation and exploration phase. The chapter surveys some related works of Q-routing. Finally, the chapter shows the experimental results of two proposals: the routing algorithm QQAR and the new server selection method.