Journal:Proceedings of the 28th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining 2022（CCF A类）
Place of Publication:Washington DC Convention Center August 14-18, 2022
Abstract:Reducing sensor requirements while keeping optimal control performances is crucial to many industrial control applications to achieve robust, low-cost, and computation-efficient controllers. However, existing feature selection solutions for the typical machine learning domain can hardly be applied in the domain of controls with changing dynamics. In this paper, a novel framework, namely the Dual-world embedded Attentive Feature Selection (D-AFS), can efficiently select the most relevant sensors for the system under dynamic control. Rather than the one world used in most Deep Reinforcement Learning (DRL), D-AFS has both the real-world and its virtual peer with twisted features. By analyzing the DRL's response in the two worlds, D-AFS can quantitatively identify respective features' importance towards control. A well-known active flow control problem, cylinder drag reduction, is used for evaluation. Results show that D-AFS successfully finds an optimized five-probe layout with 18.7\% drag reduction than the state-of-the-art solution with 151 probes and 49.2\% reduction than five-sensor layout by human experts. We also apply this solution to four OpenAI classical control cases. In all cases, D-AFS achieves the same or better sensor configurations than originally provided solutions. Results highlight, we argued, a new way to achieve efficient and optimal sensor designs for experimental or industrial systems.
Co-author:Zeng Wanxin
First Author:Wei Jiawen, Wang Fangyuan
Indexed by:Essay collection
Correspondence Author:Gui Ning
Translation or Not:no