Summary
International Symposium on Nonlinear Theory and Its Applications
2015
Session Number:B2L-A
Session:
Number:B2L-A-1
A Method for Enhancing Robustness of MCMC-Based Autonomous Decentralized Control
Masaya Yokota, Yusuke Sakumoto, Masaki Aida,
pp.411-414
Publication Date:2015/12/1
Online ISSN:2188-5079
Summary:
ADC (Autonomous Decentralized Control) is being actively discussed to realize a scalable control scheme for large-scale and wide-area systems. The previous work has proposed the MCMC (Markov Chain Monte Carlo)-based ADC for indirectly controlling the entire system, but has not fully discussed for its adaptability against environmental fluctuations. In this paper, we propose a method to enhance the robustness of the MCMC-based ADC against severe environmental fluctuations. In addition, we clarify the effectiveness of the proposed method for severe environmental fluctuation through applying to the virtual machine placement problem in data centers.