Uniformization techniques for stochastic simulation of chemical reaction networks

Area of Science:

• Computational chemistry
• Stochastic modeling
• Chemical kinetics

Background:

• Uniformization is beneficial for time-inhomogeneous chemical reaction networks with extrinsic noise.
• Understanding uniformization through sample paths is crucial for developing new algorithms.

Purpose of the Study:

• To present an efficient pathwise stochastic simulation algorithm for time-homogeneous chemical reaction networks.
• To introduce a novel variance reduction method for moment estimators in chemical reaction networks.

Main Methods:

• Developed a pathwise stochastic simulation algorithm based on uniformization.
• Combined stratification and uniformization for variance reduction.

Main Results:

• The new algorithm has a computational complexity equivalent to Gillespie's direct method.
• The proposed method expands the applicability of uniformization for complex problems.
• A novel variance reduction technique for moment estimation was successfully developed.

Conclusions:

• The uniformization method offers a computationally attractive approach for chemical reaction network simulations.
• This work enhances the utility of uniformization in computational chemistry and stochastic modeling.