A Reinforcement Learning Framework for Dynamic Resource Allocation: First Results.

This paper addresses the problem of dynamic resource allocation among multiple entities sharing a common set of resources. A solution approach is presented based on combining the reinforcement learning methodology with function approximation architectures. An implementation of this approach in Solaris 10 demonstrated a robust near-optimal performance on a simple problem of transferring CPUs among resource partitions so as to match the stochastically changing workload in each partition, both for large and small CPU migration costs