Monte Carlo parameterization in the VirtualLeaf framework

26th IUPAP Conference on Computational Physics

物理学;计算机科学

Dzhurakhalov, A.^1,2 ; De Vos, D.^1,2 ; Beemster, G.^2 ; Broeckhove, J.^1

Department of Mathematics and Computer Science, University of Antwerp, Belgium^1

Department of Biology, University of Antwerp, Belgium^2

关键词: Energy evaluation;    Energy minimization;    Energy minimization methods;    Metropolis Monte Carlo;    Multi variables;    Robust criteria;    Simulation framework;    Threshold-value;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/640/1/012012/pdf
DOI  :  10.1088/1742-6596/640/1/012012

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

The recently developed simulation framework VirtualLeaf uses Metropolis Monte Carlo dynamics for studying plant tissue morphogenesis. Minimizing the energy of the tissue is done by an energy evaluation-only method. We developed a more robust criterion for the energy minimization method for the multivariable and complex systems where the use of a gradient norm is impossible. The proposed criterion is based on checking energy changes in a sliding window of successive energy steps against a threshold value. The advantages of the sliding window criterion are discussed and results obtained by this method are presented. The impact of the choice threshold value on the energy minimization has been studied.