Matrix Computation And Its Application In Simulation Of Hemodynamics Based Cloud Computing Platform

With the development of science and technology in the biomedical field, it is possible to obtain large-scale coronary vasculature for scientific research. The mathematical model simulation is an effective way to study hemodynamic properties in the coronary system. The key of the simulation is solving linear equation set. When faced with a large amount of data to analyze, use and calculation in simulation, due to the limitation of computing resources, the simulation cannot be completed in usually.The above problems can be solved with the help of the parallel computing technology and cloud computing technology. This study aims to use the related technology of parallel and cloud computing in hemodynamic simulation. The main contents of this paper are as follows:1. Based on the research and analysis of different scales of the real vascular networks and artificial vasculature, the hemodynamic models were constructed, and appropriate flow or blood pressure boundary conditions were prescribed.2. Studied the Hadoop cloud computing framework and the distributed computing framework Map Reduce. Used the frameworks in the preprocessing of the raw vasculature data. The preprocessor includes the functions of eliminating invalid blood vessels, adding other simulation data, eliminating and extracting daughter vessels through different conditions based on different Map Reduce programs.3. Solved the matrix equation set by using Intel MKL and Open MPI. The main process of simulation is the matrix distribution using the way of block circulation distribution, the parallel solving library using Sca LAPACK, and the data collection using NFS.In conclusion, this study implemented the data preprocessing and the hemodynamic simulation by building a cloud computing platform to overcome the limit of physical resource. The results of the hemodynamic simulation accord to the experimental data in literatures.