A Reinforcement Learning Approach to the Internet QoS Routing Problems

抄録

We present a new algorithm based on reinforcement learning for packet scheduling in routers with QoS requirements. In our approach, reinforcement learning is used to learn a scheduling policy in response to feedback from the network about the delay experienced by each traffic priority class. We construct a new traffic regulator with the stochastic learning automaton, which does not require prior knowledge of the statistics of each traffic flow and can adapt to changing traffic requirements and loads.