Importance sampling for stochastic simulations
Management Science
Probability (2nd ed.)
Review: Stochastic approaches for modelling in vivo reactions
Computational Biology and Chemistry
Selecting Simulation Algorithm Portfolios by Genetic Algorithms
PADS '10 Proceedings of the 2010 IEEE Workshop on Principles of Advanced and Distributed Simulation
| Hi-index | 0.00 |
Stochastic models of biological networks are well-established in computational systems biology. However, models are abstractions and parameters may be inaccurate or perturbed. An important topic is thus the sensitivity of analysis results to parameter perturbations. This investigates how seriously potential parameter perturbations affect the analysis results and to which parameters the results are most sensitive. In this paper, a stochastic simulation algorithm is presented that yields results for multiple perturbed models from a single simulation experiment and that is thus able to perform comparisons of results for various parameter sets without explicitly simulating each of these separately. The algorithm essentially makes use of likelihood ratios in a similar fashion as in the Importance Sampling technique for variance reduction. With a suitable adaptation to the context of perturbed model parameters it yields substantial runtime savings compared to multiple separate simulations of the perturbed models without any loss in statistical accuracy.