Skip to Main Content
This paper describes an approach which uses a machine learning model in a packet scheduling problem to improve the end-to-end delay. A good Packet scheduling discipline allows to achieve QoS differentiation and to optimize the queuing delay. In a dynamically changing environment this discipline should be also adaptive to the new traffic conditions. We model this problem as a multi-agent system which consists of a whole of autonomous learning agents that interact with the environment. We define the learning problem as a decentralized process using a general mathematical framework, namely Markovian Decision Processes which are an effective tool in the modelling of the decision-making in uncertain dynamic environments. Coordination between agents occurs through communication governed by an ant colony model.