Construction of Process-Differentiable Representations for Parametric Families of Distributions

摘要

Sensitivity analysis for stochastic systems plays an important role in the statistical study of such systems, and in the optimization of stochastic processes. This paper considers questions related to the efficient computation of Monte Carlo estimates for the derivatives corresponding to such systems. These derivatives play a crucial part in the development of numerical schemes for accomplishing the statistical analysis and optimization mentioned above. Keywords: Computations; Distribution functions; Simulation.