| In biological systems formed by living cells, the small populations of some reactant species can result in dynamical behavior which cannot be captured by the traditional deterministic approaches. In that case, a more accurate simulation can be obtained by using the machinery of Markov process theory, specifically the Stochastic Simulation Algorithm (SSA). For realistic biochemical systems, the simulation with SSA carries an extremely high computational cost. Due to the central role that stochastic simulation plays in many fields such as system biology, ecology and materials, the development of high performance discrete stochastic simulation methods for chemically reacting systems has become an active research area in computational biology.;This thesis addresses the computational demands of the SSA in two different, but complementary ways: faster algorithms and efficient use of high performance computing architectures. We explored the speed potential of high-performance SSA simulation first on clusters of workstations, and then on a General Purpose Graphics Processing Unit, which yielded a tremendous performance improvement that was demonstrated on a number of applications including biochemical simulations and a fish schooling model. Finally, we briefly introduce the stochastic simulation toolkit STOCHKIT, in which we played a leading role on the team, which aims to make the latest simulation technology accessible to systems biologists. |
