Adaptive Approaches to Distributed Resource Allocation

Abstract

The problem of allocating scarce, reusable resources over time to interconnected tasks in uncertain and changing environments, in order to optimize a performance measure, arises in many real-world domains. The paper examines several recent distributed optimization approaches to this problem and compares their properties, such as the guarantees of finding (near-)optimal solutions, their robustness against disturbances or against imprecise, uncertain models, with a special emphasis on adaptive capabilities. The paper also presents a reinforcement learning based distributed resource control system and argues that this method represent one of the most promising approaches to handling resource allocation problems in the presence of uncertainties.