Online quickest multiarmed bandit algorithm for distributive renewable energy resources

Abstract

Distributive renewable energy resource (DRER) system is one key research in smart grid technology. One important challenge is to select the best among many DRERs. In this paper, we develop an online quickest multiarmed bandit algorithm to determine the best choice of DRERs as few samples as possible, under the constraint of accuracy. We derive the close form for the confident interval and obtain an upper bound for the expected regret for the proposed scheme. From the simulation results, we can show that a user can effectively switch and select the best DRER with the minimum delay while balancing the exploitation and exploration.