Adaptive Function to Ensure Robustness of MCMC-based Autonomous Decentralized Control Mechanism against Changing Environment

Abstract

In our previous work, we proposed an autonomous decentralized control mechanism (ADCM) to be able to control the probability distribution of a system performance variable on the Markov chain Monte Carlo method. As interesting features of our ADCM, we proved the probability distribution controlled by our ADCM is described by a few macro parameters, and discovered a law between the macro parameters and the external environment of the system in our ADCM. In this paper, on the basis of the law, we design an autonomous decentralized adaptive function to adapt to change in the external environment. This function ensures the robustness of our ADCM against changing environment. We apply our ADCM with the proposed adaptive function to a virtual machine placement problem in data center networks (DCNs). Simulation experiments confirm that the proposed adaptive function effectively deals with several DCN scenarios with changing environment.