Collaborative reinforcement learning of autonomic behaviour

This paper introduces Collaborative Reinforcement Learning (CRL), a coordination model for solving system-wide optimisation problems in distributed systems where there is no support for global state. In CRL the autonomic properties of a distributed system emerge from the coordination of individual agents solving discrete optimisation problems using Reinforcement Learning. In the context of an ad hoc routing protocol, we show how system-wide optimisation in CRL can be used to establish and maintain autonomic properties for decentralised distributed systems.