CONTEXT-AWARE SELF-ORGANIZED RESOURCE ALLOCATION IN INTELLIGENT WATER INFORMATICS

초록

In this paper, we investigate the water resource allocation for smart water grids (SWG) with users’ context information. A water resource sharing scheme is proposed for efficient managing water resources with the aid of the intelligent water informatics. A novel spectral clustering algorithm is developed to classify end-users into different communities with respect to the endusers’ profiles. We characterize the dynamics of the SWG with the Markov decision process (MDP), and an online Q-learning aided water allocation algorithm is conceived by virtue of the MDP for adapting the dynamics of the SWG, thus improving the water utility efficiency. Numerical results demonstrate the advantages of the proposed scheme over the conventional static schemes.