Initialization Algorithms for Coupled Dynamic Systems

University essay from Lunds universitet/Matematik LTH

Abstract: In this master thesis the consistent initialization problem is studied and three different algorithms were developed regarding the subject area - a graph algorithm used for solving the initialization problem, a parallel algorithm to enable parallel computations when solving the initialization problem and lastly a genetic algorithm used as a preprocessing stage for parallelization. The thesis is based on the Python package PyFMI, a high-level package developed by Modelon AB for working with models compliant with the FMI standard. The algorithms were tested on test cases consisting of several synthetic examples as well as in a simulation of a real industrial physical model. The analysis based on these test cases showed that the graph algorithm outperformed previously algorithms in terms of optimization, a speedup was achieved when using the parallel algorithm and the genetic algorithm was able to further increase the speedup factor.

  AT THIS PAGE YOU CAN DOWNLOAD THE WHOLE ESSAY. (follow the link to the next page)