Stochastic Approach to Uncertainty Control in Multiphysics Systems

Abstract

The problem of uncertainty analysis in multi-component systems is considered. As an example a problem of decision making under uncertainty in task of modeling of carbon balance and analysis of greenhouse gases (GHG) emissions using satellite tools was considered. Approaches to decision making under uncertainty are described: interval, fuzzy, and stochastic assessments. Different approaches and algorithms to calculate carbon and GHG emissions are described. For every algorithm (deterministic inventory, ecological modeling, and satellite control of emissions) errors and uncertainties are analyzed and estimated. Algorithms for uncertainty analysis are presented. Algorithm for analysis of components of uncertainty of vegetation productivity assessment using satellite data is proposed. Uncertainty component analysis allows understand important properties of the system studied and its feedback to anthropogenic load and climate impact. It was demonstrated that the comprehensive analysis of uncertainties not only reduces errors but also obtains new knowledge about the systems studied.