Differential Equations Seminar
Dr. Hye-Won Kang, UMBC
Monday, November 2, 2015 · 11 AM - 12 PM
Title: Reduction for stochastic reaction networks with multi-scale conservation
Abstract:
Many chemical reaction networks are known to have multi-scale property: reaction rates vary and several time scales exist in which chemical species evolve. A large amount of research have been done on constructing multi-scale methods for stochastic chemical reaction networks to approximate temporal changes in the abundance of chemical species and to reduce network complexity. Moreover, stochastic simulation algorithms with various approximation strategies have been developed using multi-scale property to increase simulation efficiency.
Among different approaches, the multi-scale approximation method developed by Ball et al. (2006) and extended by Kang and Kurtz (2013) will be introduced. When the chemical reaction network conserves linear combinations of some chemical species with quantities in different scales, the multi-scale approximation method suggested may induce some errors. In this talk, a modified multi-scale approximation method for stochastic chemical reaction networks with conservation will be investigated and will be applied to an example in enzyme kinetics and to one generating oscillations.
This is joint work with J. Kim and G.A. Rempala.
Abstract:
Many chemical reaction networks are known to have multi-scale property: reaction rates vary and several time scales exist in which chemical species evolve. A large amount of research have been done on constructing multi-scale methods for stochastic chemical reaction networks to approximate temporal changes in the abundance of chemical species and to reduce network complexity. Moreover, stochastic simulation algorithms with various approximation strategies have been developed using multi-scale property to increase simulation efficiency.
Among different approaches, the multi-scale approximation method developed by Ball et al. (2006) and extended by Kang and Kurtz (2013) will be introduced. When the chemical reaction network conserves linear combinations of some chemical species with quantities in different scales, the multi-scale approximation method suggested may induce some errors. In this talk, a modified multi-scale approximation method for stochastic chemical reaction networks with conservation will be investigated and will be applied to an example in enzyme kinetics and to one generating oscillations.
This is joint work with J. Kim and G.A. Rempala.