Efficient Implementation of the Bayesian Inversion by MCMC with Acceleration of Posterior Sampling Using Surrogate Models

2021 ◽  
pp. 846-853
Author(s):  
Simona Domesová ◽  
Michal Béreš ◽  
Radim Blaheta
Keyword(s):  
Surrogate Models ◽  
Efficient Implementation ◽  
Bayesian Inversion ◽  
Posterior Sampling

scholarly journals Cardiovascular Modeling With Adapted Parametric Inference

2018 ◽  
Vol 62 ◽  
pp. 91-107 ◽  
Author(s):  
Didier Lucor ◽  
Olivier P. Le Maître
Keyword(s):  
Approximation Error ◽  
Surrogate Models ◽  
Patient Specific ◽  
System Response ◽  
Cardiovascular Parameter ◽  
Monte Carlo Algorithms ◽  
Statistical Framework ◽  
Posterior Sampling ◽  
Powerful Approach ◽  
Parameter Values

Computational modeling of the cardiovascular system, promoted by the advance of fluid-structure interaction numerical methods, has made great progress towards the development of patient-specific numerical aids to diagnosis, risk prediction, intervention and clinical treatment. Nevertheless, the reliability of these models is inevitably impacted by rough modeling assumptions. A strong in-tegration of patient-specific data into numerical modeling is therefore needed in order to improve the accuracy of the predictions through the calibration of important physiological parameters. The Bayesian statistical framework to inverse problems is a powerful approach that relies on posterior sampling techniques, such as Markov chain Monte Carlo algorithms. The generation of samples re-quires many evaluations of the cardiovascular parameter-to-observable model. In practice, the use of a full cardiovascular numerical model is prohibitively expensive and a computational strategy based on approximations of the system response, or surrogate models, is needed to perform the data as-similation. As the support of the parameters distribution typically concentrates on a small fraction of the initial prior distribution, a worthy improvement consists in gradually adapting the surrogate model to minimize the approximation error for parameter values corresponding to high posterior den-sity. We introduce a novel numerical pathway to construct a series of polynomial surrogate models, by regression, using samples drawn from a sequence of distributions likely to converge to the posterior distribution. The approach yields substantial gains in efficiency and accuracy over direct prior-based surrogate models, as demonstrated via application to pulse wave velocities identification in a human lower limb arterial network.

GEOMETRY OPTIMIZATION OF AIR-ASSISTED SWIRL NOZZLE BASED ON SURROGATE MODELS AND COMPUTATIONAL FLUID DYNAMICS

2019 ◽  
Vol 29 (7) ◽  
pp. 605-628
Author(s):  
Zongli Yi ◽  
Li Hou ◽  
Qi Zhang ◽  
Yousheng Wang ◽  
Yunxia You
Keyword(s):  
Fluid Dynamics ◽  
Computational Fluid Dynamics ◽  
Geometry Optimization ◽  
Surrogate Models

Applied Process Simulation-Driven Oil and Gas Separation Plant Optimization using Surrogate Modeling and Evolutionary Algorithms

2019 ◽  
Author(s):  
Anders Andreasen
Keyword(s):  
Evolutionary Algorithms ◽  
Gas Separation ◽  
Oil And Gas ◽  
Optimal Solution ◽  
Computer Experiment ◽  
Latin Hypercube Sampling ◽  
Surrogate Models ◽  
Inlet Temperature ◽  
Separation Plant ◽  
Separation Stage

In this article the optimization of a realistic oil and gas separation plant has been studied. Two different fluids are investigated and compared in terms of the optimization potential. Using Design of Computer Experiment (DACE) via Latin Hypercube Sampling (LHS) and rigorous process simulations, surrogate models using Kriging have been established for selected model responses. The surrogate models are used in combination with a variety of different evolutionary algorithms for optimizing the operating profit, mainly by maximizing the recoverable oil production. A total of 10 variables representing pressure and temperature various key places in the separation plant are optimized to maximize the operational profit. The optimization is bounded in the variables and a constraint function is included to ensure that the optimal solution allows export of oil with an RVP < 12 psia. The main finding is that, while a high pressure is preferred in the first separation stage, apparently a single optimal setting for the pressure in downstream separators does not appear to exist. In the second stage separator apparently two different, yet equally optimal, settings are revealed. In the third and final separation stage a correlation between the separator pressure and the applied inlet temperature exists, where different combinations of pressure and temperature yields equally optimal results.<br>

Sign in / Sign up

Export Citation Format

Share Document